0 00:00:00,000 --> 00:00:30,000 Dear viewer, these subtitles were generated by a machine via the service Trint and therefore are (very) buggy. If you are capable, please help us to create good quality subtitles: https://c3subtitles.de/talk/947 Thanks! 1 00:00:14,480 --> 00:00:16,789 Give a very warm welcome to Christopher 2 00:00:16,790 --> 00:00:18,190 Engelman and Anton. 3 00:00:26,190 --> 00:00:28,649 Yes, thank you. 4 00:00:28,650 --> 00:00:30,209 This is a bit intimidating. 5 00:00:30,210 --> 00:00:32,429 Large space. 6 00:00:32,430 --> 00:00:33,430 Hello, everybody. 7 00:00:34,320 --> 00:00:36,449 The topic of this talk is a new audio 8 00:00:36,450 --> 00:00:38,189 format called MQ, a 9 00:00:39,690 --> 00:00:42,269 master quality, authenticated 10 00:00:42,270 --> 00:00:44,489 and URLs. As the title of our talk 11 00:00:44,490 --> 00:00:46,919 suggests, it's more than just 12 00:00:46,920 --> 00:00:48,119 an audio format. 13 00:00:51,030 --> 00:00:53,069 And in order to explain how we arrived at 14 00:00:53,070 --> 00:00:55,590 the title, a clever stealth DRM Trojan, 15 00:00:56,670 --> 00:00:59,189 we have a lot of ground to cover and our 16 00:00:59,190 --> 00:01:02,179 talk is structured as you see here. 17 00:01:02,180 --> 00:01:03,719 So I'm going to talk a little bit about 18 00:01:03,720 --> 00:01:06,989 audio files and audio files. 19 00:01:06,990 --> 00:01:09,419 Then Antoniou is going to examine 20 00:01:09,420 --> 00:01:10,420 some of the claims 21 00:01:11,700 --> 00:01:13,559 that are being made about master quality 22 00:01:13,560 --> 00:01:14,560 authenticated. 23 00:01:15,870 --> 00:01:17,939 Then we will move on to 24 24 00:01:17,940 --> 00:01:19,739 bits, got some room to spare to have a 25 00:01:19,740 --> 00:01:21,899 look where they actually hide 26 00:01:21,900 --> 00:01:23,309 the idea. 27 00:01:23,310 --> 00:01:25,529 And finally, we're going to make some 28 00:01:25,530 --> 00:01:27,809 remarks where we think Amcu 29 00:01:27,810 --> 00:01:29,010 is actually quite clever. 30 00:01:32,070 --> 00:01:33,870 So wait. 31 00:01:35,010 --> 00:01:37,289 So it's good to be back here. 32 00:01:37,290 --> 00:01:39,119 Some of you may know me already as I've 33 00:01:39,120 --> 00:01:41,219 spoken at the conferences 34 00:01:41,220 --> 00:01:43,529 a couple of times before and mostly 35 00:01:43,530 --> 00:01:45,869 on policy related 36 00:01:45,870 --> 00:01:48,059 issues, on ID, 37 00:01:48,060 --> 00:01:50,519 digital identification and authentication 38 00:01:50,520 --> 00:01:52,079 related issues. 39 00:01:52,080 --> 00:01:54,329 This time I have to make 40 00:01:54,330 --> 00:01:57,119 a confession. I am an audio file 41 00:01:57,120 --> 00:01:59,879 and I believe that this cable 42 00:01:59,880 --> 00:02:02,519 sounds different than this cable. 43 00:02:02,520 --> 00:02:05,099 So I should be a millionaire. 44 00:02:05,100 --> 00:02:07,499 As some of you may know, the James Randi 45 00:02:07,500 --> 00:02:09,749 Educational Foundation had 46 00:02:09,750 --> 00:02:12,089 offered a million dollars to anybody 47 00:02:12,090 --> 00:02:14,339 who was able to identify the difference 48 00:02:14,340 --> 00:02:16,469 between two audio interconnects like 49 00:02:16,470 --> 00:02:19,229 these ones in a controlled 50 00:02:19,230 --> 00:02:20,230 testing. 51 00:02:22,410 --> 00:02:25,229 Here is a 10 year old 52 00:02:25,230 --> 00:02:26,909 report on this challenge. 53 00:02:26,910 --> 00:02:29,669 Of course, nobody managed 54 00:02:29,670 --> 00:02:31,899 to find these differences and win the 55 00:02:31,900 --> 00:02:32,900 million dollar. 56 00:02:35,200 --> 00:02:36,609 And I'm pretty sure I would fail at the 57 00:02:36,610 --> 00:02:39,789 test, too, but in my own experience, 58 00:02:39,790 --> 00:02:41,829 if I put one of these cables in the 59 00:02:41,830 --> 00:02:43,989 system for a week or two and then the 60 00:02:43,990 --> 00:02:44,990 other cable 61 00:02:46,420 --> 00:02:48,189 for another week, there will be some 62 00:02:48,190 --> 00:02:49,569 preference that emerges. 63 00:02:51,190 --> 00:02:52,780 And I cannot explain that. 64 00:02:54,160 --> 00:02:56,379 But I guess if you asked me to 65 00:02:56,380 --> 00:02:58,779 be a mattress, the subtle 66 00:02:58,780 --> 00:03:01,150 differences in text and softness 67 00:03:02,500 --> 00:03:04,779 won't be noticed by me by just laying 68 00:03:04,780 --> 00:03:05,679 down on them. 69 00:03:05,680 --> 00:03:07,419 But if I get to sleep on each of those 70 00:03:07,420 --> 00:03:09,339 mattresses for a week or so, I might 71 00:03:09,340 --> 00:03:12,519 develop some sort of preference. 72 00:03:12,520 --> 00:03:14,289 I make this point here and I'm sure 73 00:03:14,290 --> 00:03:16,539 Anton, who is more on the objective 74 00:03:16,540 --> 00:03:17,799 side of things when it comes to 75 00:03:17,800 --> 00:03:20,049 audiophile, will be skeptical about 76 00:03:20,050 --> 00:03:21,099 it. 77 00:03:21,100 --> 00:03:23,259 But what I want to illustrate is 78 00:03:23,260 --> 00:03:24,789 that we are talking about a market 79 00:03:25,870 --> 00:03:28,029 that operates and exploits the limits 80 00:03:28,030 --> 00:03:30,579 of percept ability and 81 00:03:30,580 --> 00:03:32,859 where environmental and habitual factors 82 00:03:32,860 --> 00:03:34,959 play a role in what 83 00:03:34,960 --> 00:03:36,540 you sense and what you notice. 84 00:03:37,690 --> 00:03:39,279 And I believe that certain markets have 85 00:03:39,280 --> 00:03:42,039 developed around those perceptible 86 00:03:42,040 --> 00:03:44,139 limits and exploits something like 87 00:03:44,140 --> 00:03:46,269 longtail phenomena in the distribution 88 00:03:46,270 --> 00:03:48,909 of percept ability among humans. 89 00:03:48,910 --> 00:03:50,769 In other words, in a saturated consumer 90 00:03:50,770 --> 00:03:53,739 electronics electronics market, 91 00:03:53,740 --> 00:03:56,259 there are a certain class of devices 92 00:03:56,260 --> 00:03:58,359 which offer just noticeable 93 00:03:58,360 --> 00:04:00,279 differences, and these can be 94 00:04:00,280 --> 00:04:02,469 commercially attractive and command 95 00:04:02,470 --> 00:04:04,299 premium prices. 96 00:04:04,300 --> 00:04:05,799 To illustrate this point further. 97 00:04:05,800 --> 00:04:07,839 Let me ask you, who are you? 98 00:04:07,840 --> 00:04:09,099 Here is a gamer. 99 00:04:11,180 --> 00:04:12,949 Quite a few. 100 00:04:12,950 --> 00:04:15,169 OK, so how many of you have a 240 101 00:04:15,170 --> 00:04:16,170 Hertz monitor? 102 00:04:18,230 --> 00:04:19,629 Oh, two, OK, 103 00:04:20,839 --> 00:04:23,059 how many of you have 104 00:04:23,060 --> 00:04:25,549 high rate mouse, 105 00:04:25,550 --> 00:04:27,409 quite a few more. 106 00:04:27,410 --> 00:04:28,410 Um, 107 00:04:29,750 --> 00:04:31,549 do you believe this makes a sorry this 108 00:04:31,550 --> 00:04:33,109 makes a difference in your gaming 109 00:04:33,110 --> 00:04:34,110 experience, 110 00:04:35,520 --> 00:04:37,819 OK? These are all Objectivists here, 111 00:04:37,820 --> 00:04:38,839 skeptical people. 112 00:04:40,270 --> 00:04:41,270 OK, 113 00:04:43,010 --> 00:04:45,289 values like 240 Hertz refresh rate 114 00:04:45,290 --> 00:04:47,269 are astonishingly high. 115 00:04:47,270 --> 00:04:49,519 And if you tell somebody 20 years ago 116 00:04:49,520 --> 00:04:51,499 that people are able to discern 60 years 117 00:04:51,500 --> 00:04:53,569 refresh rates from 100 hertz 118 00:04:53,570 --> 00:04:55,999 refresh rate or 250 hertz, 119 00:04:56,000 --> 00:04:57,679 they'd have told you that the latency of 120 00:04:57,680 --> 00:04:59,809 the human visual system is far 121 00:04:59,810 --> 00:05:02,329 exceeded at this rate. 122 00:05:02,330 --> 00:05:04,609 And everything everything about 60 hertz 123 00:05:04,610 --> 00:05:06,139 is moot. 124 00:05:06,140 --> 00:05:08,719 Yet Apple now sells 120 Hertz iPad 125 00:05:08,720 --> 00:05:10,339 and the difference is perceptible for 126 00:05:10,340 --> 00:05:12,619 many. It turns out that peripheral 127 00:05:12,620 --> 00:05:14,839 factors like head movement, haptic 128 00:05:14,840 --> 00:05:16,909 interaction with the screen and so 129 00:05:16,910 --> 00:05:19,109 forth modify 130 00:05:19,110 --> 00:05:22,639 certain thresholds of visibility 131 00:05:22,640 --> 00:05:24,469 in the VR context is created big 132 00:05:24,470 --> 00:05:26,089 headaches. And anybody who follows John 133 00:05:26,090 --> 00:05:28,459 Comics' Twitter Twitter feed will know 134 00:05:28,460 --> 00:05:30,919 that he continuously is surprised 135 00:05:30,920 --> 00:05:32,989 that some of his subjects are very 136 00:05:32,990 --> 00:05:35,269 sensitive towards visual artifacts 137 00:05:35,270 --> 00:05:37,129 that normal users don't notice. 138 00:05:38,670 --> 00:05:40,769 So it's very easy to make fun 139 00:05:40,770 --> 00:05:43,079 about all your files into two thousand 140 00:05:43,080 --> 00:05:44,789 dollar Ethernet cables, does this one of 141 00:05:44,790 --> 00:05:46,789 these and noticed a small error there 142 00:05:47,880 --> 00:05:49,709 that's probably there because packet's 143 00:05:49,710 --> 00:05:51,629 otherwise would get confused which way to 144 00:05:51,630 --> 00:05:52,630 go. 145 00:05:53,640 --> 00:05:56,039 So these kinds of gadgets and there are 146 00:05:56,040 --> 00:05:58,709 many, many examples of those, 147 00:05:58,710 --> 00:06:00,929 of course, give ample ground 148 00:06:00,930 --> 00:06:02,759 to make fun of audio files. 149 00:06:02,760 --> 00:06:04,259 But that's not what we want, what we want 150 00:06:04,260 --> 00:06:06,059 to do here today. 151 00:06:06,060 --> 00:06:07,889 Instead, we want to look at master 152 00:06:07,890 --> 00:06:10,019 quality authenticated this new 153 00:06:10,020 --> 00:06:12,119 format as a development in the audio 154 00:06:12,120 --> 00:06:14,249 file world that deals with these 155 00:06:14,250 --> 00:06:16,079 aforementioned elements of perceptible 156 00:06:16,080 --> 00:06:18,179 otti in a clever way, exploits 157 00:06:18,180 --> 00:06:20,639 them to create and legitimize 158 00:06:20,640 --> 00:06:22,559 a new distribution channel for audio 159 00:06:22,560 --> 00:06:23,560 content. 160 00:06:24,780 --> 00:06:26,219 And we furthermore believe that these 161 00:06:26,220 --> 00:06:28,679 developments have possible repercussions 162 00:06:28,680 --> 00:06:30,539 outside of the audio of our world and 163 00:06:30,540 --> 00:06:31,829 might affect music consumers and 164 00:06:31,830 --> 00:06:33,420 producers. And that is you 165 00:06:35,220 --> 00:06:37,799 and labels and the way 166 00:06:37,800 --> 00:06:40,589 music is distributed in profound ways. 167 00:06:40,590 --> 00:06:41,730 So again, 168 00:06:43,140 --> 00:06:44,909 NPR stands for Master Quality 169 00:06:44,910 --> 00:06:46,919 Authenticated and Promises Studio Sound 170 00:06:46,920 --> 00:06:49,019 at home. This is a website or this 171 00:06:49,020 --> 00:06:51,599 is some page on the website. 172 00:06:51,600 --> 00:06:53,879 Join the Revolution and Recorded Music. 173 00:06:54,960 --> 00:06:57,029 And there's a lot of a 174 00:06:57,030 --> 00:06:59,429 lot to unpack into the statement 175 00:06:59,430 --> 00:07:01,869 master and quality and authenticate 176 00:07:01,870 --> 00:07:02,870 it. 177 00:07:05,370 --> 00:07:06,599 To give you a little bit an idea of why 178 00:07:06,600 --> 00:07:07,799 I'm talking about this. 179 00:07:07,800 --> 00:07:09,869 As I already said, I'm an 180 00:07:09,870 --> 00:07:11,610 audio file, but I also 181 00:07:12,660 --> 00:07:15,119 sometimes spend time on the other side 182 00:07:15,120 --> 00:07:17,100 of music, namely in music production. 183 00:07:18,540 --> 00:07:19,829 I used to mention metal band or 184 00:07:19,830 --> 00:07:21,299 combination metal band called President 185 00:07:21,300 --> 00:07:22,199 Evil. 186 00:07:22,200 --> 00:07:24,389 We call to style stoner trash. 187 00:07:24,390 --> 00:07:25,409 This is a great record. 188 00:07:25,410 --> 00:07:27,389 You just shouldn't look at the waveforms 189 00:07:27,390 --> 00:07:29,099 because it clips all the way through and 190 00:07:29,100 --> 00:07:31,079 has a d.r of three for those people who 191 00:07:31,080 --> 00:07:32,080 know what that means. 192 00:07:33,030 --> 00:07:36,089 We toured with ministry and so forth, 193 00:07:36,090 --> 00:07:38,269 and on the other side of things, I am 194 00:07:38,270 --> 00:07:40,769 also interested in 195 00:07:40,770 --> 00:07:42,899 acoustic music. This is me at the studio 196 00:07:42,900 --> 00:07:44,549 in Berlin where among others, David 197 00:07:44,550 --> 00:07:46,289 Bowie, Depeche Mode and U2 have recorded 198 00:07:46,290 --> 00:07:47,849 erbium albums. 199 00:07:47,850 --> 00:07:50,099 And we recorded this album by a bit 200 00:07:50,100 --> 00:07:52,439 like Spika, which is piano and clarinet, 201 00:07:52,440 --> 00:07:54,479 music and cello music and a wonderful 202 00:07:54,480 --> 00:07:55,439 record. 203 00:07:55,440 --> 00:07:57,419 So coming back to M.K., you will 204 00:07:57,420 --> 00:07:59,549 understand that as an audio file and as 205 00:07:59,550 --> 00:08:01,019 somebody who 206 00:08:02,310 --> 00:08:04,319 likes to be in studios, I'm somewhat 207 00:08:04,320 --> 00:08:05,639 familiar with music production, 208 00:08:05,640 --> 00:08:08,069 distribution and their marketing claims 209 00:08:08,070 --> 00:08:10,229 of studio sound and take 210 00:08:10,230 --> 00:08:12,869 me into the studio and so forth. 211 00:08:12,870 --> 00:08:14,700 Of course, raise an eyebrow with me. 212 00:08:15,930 --> 00:08:18,089 And furthermore, somebody tells me or 213 00:08:18,090 --> 00:08:20,819 promises that they want to authenticate 214 00:08:20,820 --> 00:08:23,189 a digital file and not only raise 215 00:08:23,190 --> 00:08:25,349 my eyebrows, my professional 216 00:08:25,350 --> 00:08:26,579 curiosity is peaked. 217 00:08:29,190 --> 00:08:31,319 I have given talks here about the 218 00:08:31,320 --> 00:08:33,599 history of authentication media 219 00:08:33,600 --> 00:08:36,269 and I publish on these issues, 220 00:08:36,270 --> 00:08:37,859 on the questions of media, of 221 00:08:37,860 --> 00:08:39,029 authentication. 222 00:08:39,030 --> 00:08:41,459 So all things authenticating 223 00:08:41,460 --> 00:08:43,759 of interest to me and making a promise 224 00:08:43,760 --> 00:08:44,940 is something like that. 225 00:08:45,960 --> 00:08:49,409 So what they say is it sound 226 00:08:49,410 --> 00:08:51,839 like what the engineers heard and 227 00:08:51,840 --> 00:08:53,669 they are going to authenticate that on 228 00:08:53,670 --> 00:08:55,019 your playback device. 229 00:08:56,280 --> 00:08:57,989 Everybody with basic knowledge of 230 00:08:57,990 --> 00:08:59,789 intellectual property enforcement and 231 00:08:59,790 --> 00:09:02,099 digital rights management by now should 232 00:09:02,100 --> 00:09:04,379 be a little bit skeptical about 233 00:09:04,380 --> 00:09:05,380 their claims. 234 00:09:06,780 --> 00:09:07,950 As we're going to show, 235 00:09:09,000 --> 00:09:09,989 this is rightly so. 236 00:09:09,990 --> 00:09:11,399 But the devil's in the details. 237 00:09:11,400 --> 00:09:13,589 And Makua offers an interesting 238 00:09:13,590 --> 00:09:15,659 and clever design of how to bring 239 00:09:15,660 --> 00:09:18,209 digital rights management back into music 240 00:09:18,210 --> 00:09:19,769 and the music distribution context. 241 00:09:19,770 --> 00:09:21,869 As some of you may know, the 242 00:09:21,870 --> 00:09:23,999 music music industry has tried to 243 00:09:24,000 --> 00:09:26,159 do that several times and 244 00:09:26,160 --> 00:09:28,289 always failed to 245 00:09:28,290 --> 00:09:30,029 implement and sell digital rights 246 00:09:30,030 --> 00:09:32,369 management to consumers 247 00:09:32,370 --> 00:09:34,589 and most distribution 248 00:09:34,590 --> 00:09:36,239 channels these days. 249 00:09:36,240 --> 00:09:38,330 And most files come without DRM. 250 00:09:39,660 --> 00:09:41,729 So this again is 251 00:09:41,730 --> 00:09:43,829 the website of Abakua and you can 252 00:09:43,830 --> 00:09:46,019 see how they explain 253 00:09:46,020 --> 00:09:48,089 authentication and they have this 254 00:09:48,090 --> 00:09:49,440 interesting term provenance. 255 00:09:50,670 --> 00:09:52,529 And there's the important thing is here, 256 00:09:52,530 --> 00:09:54,869 the blue and green light that will light, 257 00:09:54,870 --> 00:09:57,659 that will go on on your 258 00:09:57,660 --> 00:09:58,709 device. 259 00:09:58,710 --> 00:10:01,079 Um, if a file 260 00:10:01,080 --> 00:10:02,729 is played on this device and 261 00:10:02,730 --> 00:10:04,859 authenticated and they promise 262 00:10:04,860 --> 00:10:06,869 sound like in the studio. 263 00:10:06,870 --> 00:10:07,870 Right. 264 00:10:09,510 --> 00:10:11,549 So in order to understand 265 00:10:13,260 --> 00:10:15,719 what they mean by quality 266 00:10:15,720 --> 00:10:18,839 in the acronym, it's important 267 00:10:18,840 --> 00:10:20,249 to go back a little bit in history and 268 00:10:20,250 --> 00:10:22,079 give you a background about digital 269 00:10:22,080 --> 00:10:23,519 audio. 270 00:10:23,520 --> 00:10:25,439 So what is quality? 271 00:10:25,440 --> 00:10:27,779 In the 1980s, digital 272 00:10:27,780 --> 00:10:29,069 audio was introduced. 273 00:10:29,070 --> 00:10:31,139 That means for the introduced to the 274 00:10:31,140 --> 00:10:32,699 consumer. 275 00:10:32,700 --> 00:10:34,979 Here you see the CD as 276 00:10:34,980 --> 00:10:35,969 advertised by Sony. 277 00:10:35,970 --> 00:10:37,639 I think this ad is from 1984. 278 00:10:37,640 --> 00:10:39,739 For the underlying 279 00:10:39,740 --> 00:10:42,230 format of the city is Redbook audio, 280 00:10:43,700 --> 00:10:45,979 where you have a signal that sampled at 281 00:10:45,980 --> 00:10:48,679 16 bit, what, 282 00:10:48,680 --> 00:10:50,330 44 K 283 00:10:51,770 --> 00:10:54,199 44 a thousand times a second 284 00:10:54,200 --> 00:10:56,299 and with 16 bit, we have 285 00:10:56,300 --> 00:10:58,579 a dynamic range of 65000 values 286 00:10:58,580 --> 00:11:00,379 to represent the difference between the 287 00:11:00,380 --> 00:11:02,779 most silent and loudest signal 288 00:11:02,780 --> 00:11:04,309 in the format. 289 00:11:04,310 --> 00:11:06,439 And with 44 K, we can 290 00:11:06,440 --> 00:11:08,599 represent frequencies up to 291 00:11:08,600 --> 00:11:09,839 22 kilohertz. 292 00:11:10,850 --> 00:11:13,099 This is about 10 percent higher 293 00:11:13,100 --> 00:11:14,209 than what humans was. 294 00:11:14,210 --> 00:11:16,069 Very good and that usually means very 295 00:11:16,070 --> 00:11:18,319 young ears can hear. 296 00:11:18,320 --> 00:11:20,419 Most music today is 297 00:11:20,420 --> 00:11:22,999 recorded and distributed in this 298 00:11:23,000 --> 00:11:24,000 format. 299 00:11:25,280 --> 00:11:27,949 The Audible Will and 44 one k, 300 00:11:27,950 --> 00:11:29,119 the video broadcasting 301 00:11:30,440 --> 00:11:32,689 world standardized and forty eight K and 302 00:11:32,690 --> 00:11:35,449 those of you who have surround systems, 303 00:11:35,450 --> 00:11:37,309 DVD and Blu ray based will be familiar 304 00:11:37,310 --> 00:11:38,869 with the forty eight K sampling rate. 305 00:11:40,310 --> 00:11:42,859 So in the 1990s 306 00:11:42,860 --> 00:11:44,059 we need to see another step. 307 00:11:44,060 --> 00:11:46,339 We see music production becoming fully 308 00:11:46,340 --> 00:11:48,499 digital audio workstations. 309 00:11:48,500 --> 00:11:50,719 The first 20 24 converters 310 00:11:50,720 --> 00:11:51,949 arrive in the studios. 311 00:11:53,890 --> 00:11:55,789 We see higher sampling rates. 312 00:11:55,790 --> 00:11:57,319 So this is one of these converters from 313 00:11:57,320 --> 00:11:59,619 the from the 90's and 314 00:11:59,620 --> 00:12:01,729 the 24 indicates that 315 00:12:01,730 --> 00:12:02,870 it can do 24 bits. 316 00:12:05,480 --> 00:12:07,609 And there was also an attempt 317 00:12:07,610 --> 00:12:10,069 to sell these higher 318 00:12:10,070 --> 00:12:11,809 resolutions to the consumer in the late 319 00:12:11,810 --> 00:12:14,179 90s with a little bit different format 320 00:12:14,180 --> 00:12:16,399 as the SSD introduced 321 00:12:16,400 --> 00:12:18,469 by Sony in 99. 322 00:12:18,470 --> 00:12:20,779 It's a different encoding scheme called 323 00:12:20,780 --> 00:12:22,939 TSD offers 324 00:12:22,940 --> 00:12:23,940 120 325 00:12:25,160 --> 00:12:27,229 120 decibels 326 00:12:27,230 --> 00:12:29,389 of dynamic range and in its 327 00:12:29,390 --> 00:12:31,579 initial form went 328 00:12:31,580 --> 00:12:33,979 up to a 30 kilohertz way above human 329 00:12:33,980 --> 00:12:35,089 hearing range. 330 00:12:35,090 --> 00:12:37,159 But as many of you know, 331 00:12:37,160 --> 00:12:38,989 this failed in the market. 332 00:12:38,990 --> 00:12:40,519 And it is, of course, an interesting 333 00:12:40,520 --> 00:12:42,619 backstory to this, too, because the 334 00:12:42,620 --> 00:12:44,719 SSD was introduced in the very moment 335 00:12:44,720 --> 00:12:47,089 the patents for the CD 336 00:12:47,090 --> 00:12:49,159 expired, and Sony tried to 337 00:12:49,160 --> 00:12:51,169 put a new format into the market and have 338 00:12:51,170 --> 00:12:53,629 a new patent licensing regime 339 00:12:53,630 --> 00:12:55,429 to generate revenue. 340 00:12:55,430 --> 00:12:57,559 So why did the CCD and the supposedly 341 00:12:57,560 --> 00:12:59,689 higher quality fail? 342 00:12:59,690 --> 00:13:01,249 Not only because people couldn't really 343 00:13:01,250 --> 00:13:02,250 hear it. 344 00:13:04,070 --> 00:13:07,159 On the turn of the 2000s, we had empathy 345 00:13:07,160 --> 00:13:10,159 and empathy pretty much 346 00:13:10,160 --> 00:13:11,059 on the market. 347 00:13:11,060 --> 00:13:13,129 And Steve Jobs 348 00:13:14,390 --> 00:13:16,939 put it best when he said convenience 349 00:13:16,940 --> 00:13:17,940 beats quality. 350 00:13:19,250 --> 00:13:21,439 And as you know, he turned around Apple 351 00:13:21,440 --> 00:13:24,619 basically on with the iPod, 352 00:13:24,620 --> 00:13:26,449 which could hold in its initial form, I 353 00:13:26,450 --> 00:13:27,679 don't know, five thousand songs or 354 00:13:27,680 --> 00:13:28,680 something. 355 00:13:30,230 --> 00:13:32,299 And Steve Jobs knew what he 356 00:13:32,300 --> 00:13:34,069 was talking about. 357 00:13:34,070 --> 00:13:35,839 He was actually an audio file with a 358 00:13:35,840 --> 00:13:37,999 system worth north of fifty K in his 359 00:13:38,000 --> 00:13:39,970 home and supposedly only listen to vinyl. 360 00:13:42,890 --> 00:13:45,529 This together was Napster basically 361 00:13:45,530 --> 00:13:47,689 killed the music market and 362 00:13:47,690 --> 00:13:49,579 especially the market for high resolution 363 00:13:49,580 --> 00:13:51,709 audio, as these formats are called it 364 00:13:51,710 --> 00:13:53,419 exceed the Redbox standard 365 00:13:54,830 --> 00:13:56,899 audio files. To come back to our story 366 00:13:56,900 --> 00:13:59,239 line here, of course, 367 00:13:59,240 --> 00:14:01,639 Fort Holy Wars against MP three 368 00:14:01,640 --> 00:14:04,369 and any form of lossy compression. 369 00:14:04,370 --> 00:14:06,529 They deemed the format is inferior and 370 00:14:06,530 --> 00:14:08,539 argued that it destroys a generations 371 00:14:08,540 --> 00:14:10,759 hearing and even Insight's brain damage. 372 00:14:10,760 --> 00:14:12,979 And I was actually surprised that 373 00:14:12,980 --> 00:14:15,109 Douglas Rushkoff made this argument to 374 00:14:15,110 --> 00:14:16,549 me. Three is bad for you. 375 00:14:18,560 --> 00:14:20,629 And he asked down there, will the next 376 00:14:20,630 --> 00:14:22,789 generation of music, music listeners 377 00:14:22,790 --> 00:14:24,949 be able to hear reality the same way 378 00:14:24,950 --> 00:14:27,019 we did? So the idea is that 379 00:14:27,020 --> 00:14:29,149 we somehow lose a 380 00:14:29,150 --> 00:14:31,759 contact to reality if we 381 00:14:31,760 --> 00:14:33,440 are exposed to lousy formats. 382 00:14:34,970 --> 00:14:37,189 So but however, audio files 383 00:14:37,190 --> 00:14:39,409 to abandoned 384 00:14:39,410 --> 00:14:42,499 physical media and 385 00:14:42,500 --> 00:14:44,269 things like these standalone digital 386 00:14:44,270 --> 00:14:46,369 audio converters found 387 00:14:46,370 --> 00:14:48,619 a way in in their systems of 388 00:14:48,620 --> 00:14:50,779 streaming amplifiers 389 00:14:50,780 --> 00:14:53,029 that either via USB, either it 390 00:14:53,030 --> 00:14:55,189 allow you to stream music from 391 00:14:55,190 --> 00:14:58,009 local networks or from the Internet 392 00:14:58,010 --> 00:14:59,539 into your system. 393 00:14:59,540 --> 00:15:01,729 And these converters 394 00:15:01,730 --> 00:15:03,859 you can see here, it offers DSD, which 395 00:15:03,860 --> 00:15:06,139 is as the format, and 16, 24 396 00:15:06,140 --> 00:15:08,359 bits and higher sampling rates allowed 397 00:15:08,360 --> 00:15:10,309 access to high resolution, higher 398 00:15:10,310 --> 00:15:11,310 betrayed files. 399 00:15:13,230 --> 00:15:14,230 So 400 00:15:16,170 --> 00:15:18,089 this is important because we come back to 401 00:15:18,090 --> 00:15:20,579 this in the context of Makua 402 00:15:20,580 --> 00:15:22,989 and of course, you can ask about 403 00:15:22,990 --> 00:15:25,049 the benefit of these kinds of 404 00:15:25,050 --> 00:15:27,479 devices and audibility of 405 00:15:27,480 --> 00:15:29,549 resolutions higher than 16 BITTON 406 00:15:29,550 --> 00:15:30,550 44 one key. 407 00:15:31,400 --> 00:15:33,479 Um, again, I 408 00:15:33,480 --> 00:15:35,339 think there might be individuals that 409 00:15:35,340 --> 00:15:36,359 actually have differences. 410 00:15:39,450 --> 00:15:41,399 Let's jump to the current decade 411 00:15:42,530 --> 00:15:44,819 to 2010. 412 00:15:44,820 --> 00:15:46,109 I think we are actually in a golden age 413 00:15:46,110 --> 00:15:48,929 for audio files because music 414 00:15:48,930 --> 00:15:50,999 comes basically into the house more or 415 00:15:51,000 --> 00:15:53,909 less at very low prices 416 00:15:53,910 --> 00:15:55,019 via streaming. 417 00:15:55,020 --> 00:15:57,239 And furthermore, 418 00:15:57,240 --> 00:16:00,209 they can actually buy lossless files 419 00:16:00,210 --> 00:16:02,999 in resolutions up to 24 bitta 192 420 00:16:03,000 --> 00:16:05,129 kilohertz in online stores like this one 421 00:16:05,130 --> 00:16:07,439 is the French company, Cobos. 422 00:16:07,440 --> 00:16:09,809 And you can see that 423 00:16:09,810 --> 00:16:12,539 some of the contemporary current 424 00:16:12,540 --> 00:16:14,129 records are actually offered in these 425 00:16:14,130 --> 00:16:16,619 formats. So you can buy the new 426 00:16:16,620 --> 00:16:18,809 Bjork record in high 427 00:16:18,810 --> 00:16:20,969 risk audio. I actually haven't looked at 428 00:16:20,970 --> 00:16:23,399 what resolution, I guess 2106 429 00:16:23,400 --> 00:16:24,509 or something like that. 430 00:16:24,510 --> 00:16:27,029 Incubus will even stream at that 431 00:16:27,030 --> 00:16:28,030 resolution 432 00:16:29,220 --> 00:16:30,220 over the Internet. 433 00:16:31,820 --> 00:16:33,929 There's this little website where 434 00:16:33,930 --> 00:16:36,779 you can compare the the different 435 00:16:36,780 --> 00:16:38,549 offerings. You just enter some band or 436 00:16:38,550 --> 00:16:41,699 some music and it will list you 437 00:16:41,700 --> 00:16:43,919 the availability of these files 438 00:16:43,920 --> 00:16:45,210 at different resolutions. 439 00:16:46,440 --> 00:16:48,539 So, again, in essence 440 00:16:48,540 --> 00:16:50,609 to 2010, some kind of a golden decade 441 00:16:50,610 --> 00:16:52,439 for audio files. They got docs and 442 00:16:52,440 --> 00:16:54,809 streamers in their homes and could play 443 00:16:54,810 --> 00:16:56,969 the files at the same resolution 444 00:16:56,970 --> 00:16:59,489 as their recorded, mixed and mastered 445 00:16:59,490 --> 00:17:00,539 in the studios, 446 00:17:01,860 --> 00:17:03,089 in their homes. 447 00:17:03,090 --> 00:17:05,368 And it is this context, basically, that 448 00:17:05,369 --> 00:17:07,649 the master quality authenticated format 449 00:17:07,650 --> 00:17:08,940 was launched and promoted 450 00:17:09,990 --> 00:17:11,839 into 2014. 451 00:17:13,770 --> 00:17:15,659 There was a big launch event at the chart 452 00:17:15,660 --> 00:17:17,579 at this high rise in London. 453 00:17:18,660 --> 00:17:20,669 Looks a little bit like this tower in 454 00:17:20,670 --> 00:17:22,319 Middle Earth for Zahran resides 455 00:17:25,290 --> 00:17:27,299 and they claimed 456 00:17:28,410 --> 00:17:30,929 a revolutionary advance 457 00:17:30,930 --> 00:17:32,969 in audio and to deliver studio sound at 458 00:17:32,970 --> 00:17:33,970 home. 459 00:17:34,380 --> 00:17:36,089 But not only that, and that's where the 460 00:17:36,090 --> 00:17:37,230 story gets interesting. 461 00:17:38,790 --> 00:17:41,279 The inventors credited 462 00:17:41,280 --> 00:17:43,439 as Peter Craven and Bob Stewart. 463 00:17:43,440 --> 00:17:45,209 You already saw Stewart on a picture. 464 00:17:45,210 --> 00:17:47,699 I'm going, where is it when it 465 00:17:47,700 --> 00:17:50,009 comes back later, they backed 466 00:17:50,010 --> 00:17:51,209 their claims with a paper that was 467 00:17:51,210 --> 00:17:52,889 published in the Proceedings of the Audio 468 00:17:52,890 --> 00:17:54,359 Engineering Society. 469 00:17:54,360 --> 00:17:55,799 And they say that M.K. 470 00:17:55,800 --> 00:17:58,559 builds on advances in neuroscience 471 00:17:58,560 --> 00:18:01,109 and also sometimes say advances 472 00:18:01,110 --> 00:18:03,389 in assembling theory beyond Shannon 473 00:18:03,390 --> 00:18:05,459 Nykvist, which are strong claims 474 00:18:05,460 --> 00:18:07,679 to make. So if you go to the website, 475 00:18:07,680 --> 00:18:09,779 you will find this explanation 476 00:18:09,780 --> 00:18:11,519 in the Q How does it work? 477 00:18:11,520 --> 00:18:13,619 And it says, Makua is based entirely on 478 00:18:13,620 --> 00:18:14,639 science. 479 00:18:14,640 --> 00:18:16,499 Specifically, it's based on new findings 480 00:18:16,500 --> 00:18:18,029 in neuroscience that have told us that 481 00:18:18,030 --> 00:18:19,619 the resolution of timing information is 482 00:18:19,620 --> 00:18:21,629 critical to our hearing and our brain 483 00:18:21,630 --> 00:18:22,739 into action. 484 00:18:22,740 --> 00:18:24,989 So if you look in this paper 485 00:18:24,990 --> 00:18:27,419 published in the proceedings 486 00:18:27,420 --> 00:18:29,819 and look specifically at the neuroscience 487 00:18:29,820 --> 00:18:31,649 work site, it there is I think it's seven 488 00:18:31,650 --> 00:18:32,759 or something like that. 489 00:18:32,760 --> 00:18:34,379 There's only one paper that actually 490 00:18:34,380 --> 00:18:36,139 deals with humans. 491 00:18:36,140 --> 00:18:38,549 The other papers test 492 00:18:38,550 --> 00:18:41,429 hearing with Mongolian Gabriele's 493 00:18:41,430 --> 00:18:42,759 and barnacles. 494 00:18:42,760 --> 00:18:43,709 Right. 495 00:18:43,710 --> 00:18:46,089 And it's probably 496 00:18:46,090 --> 00:18:48,359 another mistake to believe 497 00:18:48,360 --> 00:18:51,059 that these animals have a very different 498 00:18:51,060 --> 00:18:53,609 auditory system than we have and 499 00:18:53,610 --> 00:18:56,369 that the hearing exceeds 500 00:18:56,370 --> 00:18:58,559 our hearing when it comes to high 501 00:18:58,560 --> 00:18:59,560 frequencies. 502 00:19:01,290 --> 00:19:04,019 So building on these neuroscientific 503 00:19:04,020 --> 00:19:07,169 advances, they claim to correct 504 00:19:07,170 --> 00:19:09,179 for the time domain errors in digital 505 00:19:09,180 --> 00:19:11,759 audio chain, namely errors introduced 506 00:19:11,760 --> 00:19:13,469 in analog digital converters. 507 00:19:13,470 --> 00:19:15,659 So when you record an analog signal 508 00:19:15,660 --> 00:19:17,999 in the studio and convert it to digital 509 00:19:20,160 --> 00:19:21,160 and 510 00:19:22,290 --> 00:19:24,299 they say that this would create temporal 511 00:19:24,300 --> 00:19:26,969 blurr, this is what they here say. 512 00:19:26,970 --> 00:19:29,159 There's a problem with digital, 513 00:19:29,160 --> 00:19:30,959 with digital. It's called blurring. 514 00:19:30,960 --> 00:19:32,400 It's also from the website. 515 00:19:34,920 --> 00:19:36,899 And it's temporal because the is 516 00:19:36,900 --> 00:19:38,879 basically the main marketing mantra of 517 00:19:38,880 --> 00:19:40,429 Makua. 518 00:19:40,430 --> 00:19:42,839 And the quality argument, the argument 519 00:19:42,840 --> 00:19:43,890 that it sounds better 520 00:19:45,180 --> 00:19:47,669 rests to a large part 521 00:19:47,670 --> 00:19:49,229 on this claim, not only to bring the 522 00:19:49,230 --> 00:19:51,389 studio sound to the home, but by 523 00:19:51,390 --> 00:19:53,699 reducing what they call 524 00:19:53,700 --> 00:19:54,700 temporal blur 525 00:19:55,920 --> 00:19:58,319 the format, the format so they claim 526 00:19:58,320 --> 00:20:00,359 contains mechanisms to cope and correct 527 00:20:00,360 --> 00:20:02,699 these temporal inefficiencies present 528 00:20:02,700 --> 00:20:04,859 in all digital systems where an analog 529 00:20:04,860 --> 00:20:07,229 digital analog conversion takes place 530 00:20:07,230 --> 00:20:09,719 and remove their temporal blur. 531 00:20:09,720 --> 00:20:11,849 And reviewers and the audio file press 532 00:20:11,850 --> 00:20:14,009 called. Innovations revolutionary 533 00:20:14,010 --> 00:20:16,589 and claimed them not only to be audible, 534 00:20:16,590 --> 00:20:18,869 but to lift music to a whole new level. 535 00:20:18,870 --> 00:20:20,519 This is from an audio file. 536 00:20:22,060 --> 00:20:24,129 Public online publication, 537 00:20:24,130 --> 00:20:26,409 The Sound, they say, let the revolution 538 00:20:26,410 --> 00:20:27,410 begin. 539 00:20:28,330 --> 00:20:30,339 And here they're more or less compared. 540 00:20:30,340 --> 00:20:31,629 Bob Stewart with Einstein. 541 00:20:31,630 --> 00:20:32,680 If you read this here, 542 00:20:33,770 --> 00:20:35,429 this captures a lot, Kevin, up there. 543 00:20:35,430 --> 00:20:36,819 There's nothing new to be discovered in 544 00:20:36,820 --> 00:20:38,499 physics. Now, all that remains is one 545 00:20:38,500 --> 00:20:39,669 more precise measurement. 546 00:20:39,670 --> 00:20:41,709 Lord Kelvin, five years before Einstein's 547 00:20:41,710 --> 00:20:43,959 paper on relativity. 548 00:20:43,960 --> 00:20:45,879 And if you look deep in the text and I'm 549 00:20:45,880 --> 00:20:48,009 sorry for this, I know you can 550 00:20:48,010 --> 00:20:49,010 read that now, 551 00:20:50,080 --> 00:20:52,209 but they basically bring forward Thomas 552 00:20:52,210 --> 00:20:54,279 Kuhn revolution, the structure of 553 00:20:54,280 --> 00:20:56,979 scientific revolutions and talk about a 554 00:20:56,980 --> 00:20:59,079 paradigm shift in digital audio and 555 00:20:59,080 --> 00:21:00,819 even come up with the Copernican model 556 00:21:00,820 --> 00:21:02,709 supplanting plutonomy worldview and so 557 00:21:02,710 --> 00:21:05,049 forth. Right. So big things are happening 558 00:21:05,050 --> 00:21:06,399 already right now. You probably haven't 559 00:21:06,400 --> 00:21:08,890 heard yet, but we're entering 560 00:21:10,090 --> 00:21:12,279 a new paradigm and we are currently and 561 00:21:12,280 --> 00:21:14,139 the fact that we are standing here is 562 00:21:14,140 --> 00:21:15,669 actually an indicator of the of the 563 00:21:15,670 --> 00:21:17,319 scientific crisis because we belong to 564 00:21:17,320 --> 00:21:19,569 the old world view and think 565 00:21:19,570 --> 00:21:20,589 probably not. 566 00:21:20,590 --> 00:21:21,549 Right. 567 00:21:21,550 --> 00:21:22,880 So, um, 568 00:21:25,030 --> 00:21:27,189 there is, of course, a big debate 569 00:21:27,190 --> 00:21:29,049 happening all over the world around 570 00:21:29,050 --> 00:21:30,549 claims like this. 571 00:21:30,550 --> 00:21:32,829 And much of this debate 572 00:21:32,830 --> 00:21:35,199 is happening on a website called Computer 573 00:21:35,200 --> 00:21:36,369 Audio File. 574 00:21:36,370 --> 00:21:38,739 And there is 575 00:21:38,740 --> 00:21:40,869 specifically this thread, which was 576 00:21:40,870 --> 00:21:43,419 opened on January two this year, 577 00:21:43,420 --> 00:21:45,519 and it has amassed 250 K views 578 00:21:45,520 --> 00:21:47,739 by now. It's really interesting because 579 00:21:47,740 --> 00:21:49,839 there's a couple of of really reputable 580 00:21:49,840 --> 00:21:51,430 people publishing 581 00:21:52,570 --> 00:21:54,699 on this on this forum in this 582 00:21:54,700 --> 00:21:55,700 thread. 583 00:21:57,760 --> 00:21:59,739 For many outspoken critics, one of them 584 00:21:59,740 --> 00:22:02,349 is mastering engineer Brian Lucey, who 585 00:22:02,350 --> 00:22:04,449 has six Grammys and mastered, among 586 00:22:04,450 --> 00:22:06,099 others, Depeche Mode and Merriman's and 587 00:22:06,100 --> 00:22:07,059 so forth. 588 00:22:07,060 --> 00:22:09,219 And he basically said that 589 00:22:09,220 --> 00:22:11,169 the available MQ airfoils that you can 590 00:22:11,170 --> 00:22:14,049 buy now a stream via the tidal service 591 00:22:14,050 --> 00:22:16,089 were not created in mastering studios, 592 00:22:17,140 --> 00:22:19,359 but encoded in an automatic way 593 00:22:19,360 --> 00:22:21,519 somewhere probably on the cloud. 594 00:22:21,520 --> 00:22:23,709 So the claim of Makua that, 595 00:22:23,710 --> 00:22:25,809 you know, you get the sound that the 596 00:22:25,810 --> 00:22:27,609 artist and the mastering engineer signed 597 00:22:27,610 --> 00:22:30,009 off on is 598 00:22:30,010 --> 00:22:32,229 definitely false for the files 599 00:22:32,230 --> 00:22:33,639 available now. 600 00:22:33,640 --> 00:22:35,829 And we must presume that it remain 601 00:22:35,830 --> 00:22:37,390 remains false because 602 00:22:38,620 --> 00:22:40,869 there is no commercial 603 00:22:40,870 --> 00:22:42,939 convertors available that include 604 00:22:42,940 --> 00:22:44,169 MKR encoding. 605 00:22:44,170 --> 00:22:46,299 So nobody running a studio can actually 606 00:22:46,300 --> 00:22:48,549 use this format right 607 00:22:48,550 --> 00:22:50,359 now. You have to send the files often and 608 00:22:50,360 --> 00:22:52,479 encoding house where they get 609 00:22:52,480 --> 00:22:53,680 encoded and signed. 610 00:22:55,620 --> 00:22:56,620 So 611 00:22:58,560 --> 00:23:00,749 part of this controversy is a couple of 612 00:23:00,750 --> 00:23:03,229 people online who did deeper 613 00:23:03,230 --> 00:23:05,219 research, one of them is Archipelago, a 614 00:23:05,220 --> 00:23:07,529 blogger, an anonymous 615 00:23:07,530 --> 00:23:09,659 blogger who actually set 616 00:23:09,660 --> 00:23:10,980 up an Internet blind test, 617 00:23:12,110 --> 00:23:13,559 quote, According to our standards, high 618 00:23:13,560 --> 00:23:15,059 virus files. If you're interested in 619 00:23:15,060 --> 00:23:16,439 that, please look it up. 620 00:23:16,440 --> 00:23:17,979 It's actually pretty complicated what he 621 00:23:17,980 --> 00:23:20,010 does, because since 622 00:23:21,120 --> 00:23:22,680 the encoders are not available, 623 00:23:23,970 --> 00:23:26,459 we cannot properly encode music 624 00:23:26,460 --> 00:23:29,189 Abakua. And in the past four years, 625 00:23:29,190 --> 00:23:31,109 the company has not allowed anybody to 626 00:23:31,110 --> 00:23:33,209 basically sent them files in 627 00:23:33,210 --> 00:23:35,159 order to create a proper comparison 628 00:23:35,160 --> 00:23:37,419 between Mkhitaryan 629 00:23:37,420 --> 00:23:39,489 codified sound files that in PCM 630 00:23:39,490 --> 00:23:41,279 or any other format that you would have. 631 00:23:41,280 --> 00:23:43,349 So I can invented a pretty 632 00:23:43,350 --> 00:23:45,749 interesting way to work around a 633 00:23:45,750 --> 00:23:46,829 limitation. 634 00:23:46,830 --> 00:23:48,359 I can't go further into that. 635 00:23:48,360 --> 00:23:50,429 If you want to take this test, 636 00:23:50,430 --> 00:23:51,430 look at this website. 637 00:23:52,980 --> 00:23:55,059 Another Hecke that had a look and 638 00:23:55,060 --> 00:23:57,419 is Mansor God, who 639 00:23:57,420 --> 00:24:00,089 is who's working on the Sox 640 00:24:00,090 --> 00:24:03,209 resampling package. 641 00:24:03,210 --> 00:24:05,369 And he 642 00:24:05,370 --> 00:24:07,919 has a Gitlow site up 643 00:24:07,920 --> 00:24:10,589 where you find a filter coefficients of 644 00:24:10,590 --> 00:24:12,869 a in tools to actually 645 00:24:12,870 --> 00:24:14,949 look into control bits at the 646 00:24:14,950 --> 00:24:17,259 street, at the control 647 00:24:17,260 --> 00:24:18,360 stream of a. 648 00:24:19,800 --> 00:24:22,169 We will also list later 649 00:24:22,170 --> 00:24:23,170 in the. 650 00:24:26,200 --> 00:24:28,269 Presentation once we put it up, so 651 00:24:28,270 --> 00:24:31,539 to wrap up my part so far in 2017, 652 00:24:31,540 --> 00:24:33,609 24 bit files as 653 00:24:33,610 --> 00:24:34,989 they are created in the studio, worked on 654 00:24:34,990 --> 00:24:38,199 Scutti studio, available for audio files 655 00:24:38,200 --> 00:24:39,909 so they can actually listen to what's 656 00:24:39,910 --> 00:24:41,349 happening in the studio in a way. 657 00:24:41,350 --> 00:24:43,569 Of course, there's other, you know, 658 00:24:43,570 --> 00:24:45,699 like the listening room and the playback 659 00:24:45,700 --> 00:24:47,949 change and so forth. Our other parameter 660 00:24:47,950 --> 00:24:50,769 affecting that, what they hear. 661 00:24:50,770 --> 00:24:52,570 But does that actually matter? 662 00:24:54,370 --> 00:24:56,439 Because most music is all 663 00:24:56,440 --> 00:24:58,539 of you will know gets consumed 664 00:24:58,540 --> 00:25:00,759 in lossy formats via Spotify 665 00:25:00,760 --> 00:25:02,679 or any other streaming service and 666 00:25:02,680 --> 00:25:05,049 listened on speakers like this, 667 00:25:05,050 --> 00:25:06,549 mono Bluetooth speakers. 668 00:25:07,840 --> 00:25:10,329 So in this context now, empty 669 00:25:10,330 --> 00:25:12,879 promises, lossless studio sound at home 670 00:25:12,880 --> 00:25:14,529 that is authenticated. 671 00:25:14,530 --> 00:25:16,659 And they also 672 00:25:16,660 --> 00:25:19,209 promise a sound improvement via fixing 673 00:25:19,210 --> 00:25:21,279 the temporal blur of supposedly 674 00:25:21,280 --> 00:25:24,459 happening in the adult stages. 675 00:25:24,460 --> 00:25:26,769 So Anton, who actually 676 00:25:26,770 --> 00:25:28,239 works on these matters, will now 677 00:25:28,240 --> 00:25:29,889 investigate these technical claims sort 678 00:25:29,890 --> 00:25:30,890 of further. 679 00:25:39,870 --> 00:25:42,509 OK, you go to 680 00:25:42,510 --> 00:25:45,509 your high 681 00:25:45,510 --> 00:25:47,579 school as Christopher introduced 682 00:25:47,580 --> 00:25:49,979 himself, I do the same, 683 00:25:49,980 --> 00:25:52,259 my background in digital 684 00:25:52,260 --> 00:25:54,599 audio stuff, it's 685 00:25:54,600 --> 00:25:56,549 changing slides by itself. 686 00:25:56,550 --> 00:25:58,250 Is there a way to to press? 687 00:26:00,310 --> 00:26:02,889 All right, so my background in 688 00:26:02,890 --> 00:26:05,649 digital audio started in nineteen ninety 689 00:26:05,650 --> 00:26:08,109 three when I bought this CD player 690 00:26:11,590 --> 00:26:13,719 together with an amplifier and two Hakoah 691 00:26:13,720 --> 00:26:14,720 speakers. 692 00:26:15,430 --> 00:26:17,739 And looking back at that time, it was 693 00:26:17,740 --> 00:26:19,329 the only point in my life where I 694 00:26:19,330 --> 00:26:21,489 actually carefully listened to 695 00:26:21,490 --> 00:26:23,559 audio quality as 696 00:26:23,560 --> 00:26:25,899 a listener and I repeatedly 697 00:26:25,900 --> 00:26:28,299 listened to my favorite records, 698 00:26:28,300 --> 00:26:30,399 which were mostly dense music at that 699 00:26:30,400 --> 00:26:32,559 time. Again and again, 700 00:26:32,560 --> 00:26:35,079 the player did so well and 701 00:26:35,080 --> 00:26:36,689 still makes the job today. 702 00:26:38,980 --> 00:26:41,499 Later on in 2000, 703 00:26:41,500 --> 00:26:44,319 I started studying at the Teacher 704 00:26:44,320 --> 00:26:47,169 Ilmenite close to lapsation, 705 00:26:47,170 --> 00:26:49,689 and I heard the lectures of Karlheinz 706 00:26:49,690 --> 00:26:52,299 Brandenburg, who led 707 00:26:52,300 --> 00:26:54,519 in the 90s group that developed 708 00:26:54,520 --> 00:26:57,279 the military algorithm. 709 00:26:57,280 --> 00:27:00,129 And with that invention, I later 710 00:27:00,130 --> 00:27:02,259 in the decade, in the 711 00:27:02,260 --> 00:27:05,469 last decade, I became a very convenient 712 00:27:05,470 --> 00:27:08,979 listener. I downloaded continuously 713 00:27:08,980 --> 00:27:11,169 and then in the current 714 00:27:11,170 --> 00:27:14,439 decade I am streaming. 715 00:27:14,440 --> 00:27:18,219 So nevertheless, 716 00:27:18,220 --> 00:27:20,499 or of course, Anku 717 00:27:20,500 --> 00:27:21,500 wants to sell 718 00:27:22,570 --> 00:27:25,629 a system again to me and I'm inclined 719 00:27:25,630 --> 00:27:28,239 to do to to do a purchase 720 00:27:28,240 --> 00:27:30,400 soon, but not 721 00:27:31,450 --> 00:27:33,719 the one that comes with 722 00:27:33,720 --> 00:27:36,069 me and why 723 00:27:36,070 --> 00:27:38,169 that is like that. 724 00:27:38,170 --> 00:27:40,739 I'm going to show up now. 725 00:27:43,600 --> 00:27:45,990 So apart from the DRM, I, 726 00:27:47,230 --> 00:27:49,359 I do not appreciate the 727 00:27:49,360 --> 00:27:51,459 audio quality quality of 728 00:27:51,460 --> 00:27:52,460 of. 729 00:27:53,680 --> 00:27:55,869 And before we have a quick look. 730 00:27:57,320 --> 00:27:59,150 At the audio technology, 731 00:28:00,560 --> 00:28:02,599 let me state the following, there's not 732 00:28:02,600 --> 00:28:04,789 much information out there 733 00:28:04,790 --> 00:28:06,919 from McCooey on the algorithm 734 00:28:06,920 --> 00:28:09,349 itself, there are Cristoff 735 00:28:09,350 --> 00:28:11,479 mentioned it already. There are mainly 736 00:28:11,480 --> 00:28:13,730 two papers, 737 00:28:14,930 --> 00:28:18,559 convention papers of the A-S. 738 00:28:18,560 --> 00:28:20,659 There are patents which, of 739 00:28:20,660 --> 00:28:22,729 course, highlight the crucial 740 00:28:22,730 --> 00:28:25,309 parts and increase 741 00:28:25,310 --> 00:28:27,019 their freedom to operate. 742 00:28:27,020 --> 00:28:29,929 So it's it's vague, 743 00:28:29,930 --> 00:28:31,819 but there's luckily a lot of information 744 00:28:31,820 --> 00:28:33,500 from engineers that 745 00:28:34,640 --> 00:28:36,799 that looked into 746 00:28:36,800 --> 00:28:39,589 the details and 747 00:28:39,590 --> 00:28:41,300 did a reverse engineering. 748 00:28:42,470 --> 00:28:44,869 So MQ is an encoder 749 00:28:44,870 --> 00:28:48,019 decoder system nested in a container 750 00:28:48,020 --> 00:28:50,269 such as FLAC, free lossless audio 751 00:28:50,270 --> 00:28:52,659 coding or audio 752 00:28:52,660 --> 00:28:54,919 recording of Apple Lossless or Decoding 753 00:28:54,920 --> 00:28:57,439 ALAC or the Microsoft 754 00:28:57,440 --> 00:28:58,440 WAF. 755 00:28:59,620 --> 00:29:00,620 Container. 756 00:29:03,090 --> 00:29:05,879 And thereby can be read 757 00:29:05,880 --> 00:29:08,189 and decoded by standard 758 00:29:08,190 --> 00:29:11,279 audio players 759 00:29:11,280 --> 00:29:12,480 in hard and software. 760 00:29:14,480 --> 00:29:16,949 Of course, MTV wants to replace 761 00:29:16,950 --> 00:29:19,019 the entire audio chain, 762 00:29:19,020 --> 00:29:21,119 the entire audio chain of ad 763 00:29:21,120 --> 00:29:23,939 converters and de converters 764 00:29:23,940 --> 00:29:26,759 and wants to get 765 00:29:26,760 --> 00:29:28,679 royalties also from the streaming 766 00:29:28,680 --> 00:29:29,680 services. 767 00:29:31,470 --> 00:29:33,569 And the two main approaches 768 00:29:33,570 --> 00:29:34,570 to 769 00:29:35,640 --> 00:29:37,889 reduce defiles size 770 00:29:37,890 --> 00:29:40,169 such that you sent a 771 00:29:40,170 --> 00:29:42,899 24 hour 20 772 00:29:42,900 --> 00:29:45,629 to a bandwidth 773 00:29:45,630 --> 00:29:48,119 file over the Internet, 774 00:29:48,120 --> 00:29:50,309 but nevertheless by unfolding is it 775 00:29:50,310 --> 00:29:53,009 you can still have the high 776 00:29:53,010 --> 00:29:54,150 resolution audio. 777 00:29:58,400 --> 00:30:01,039 So the MQ claims are 778 00:30:01,040 --> 00:30:03,199 there, it's lawless, and 779 00:30:03,200 --> 00:30:05,839 it shall improve the audio quality. 780 00:30:09,440 --> 00:30:11,569 Is scalable, the more 781 00:30:11,570 --> 00:30:13,919 you pay, the more bandwidth you get. 782 00:30:13,920 --> 00:30:15,140 So from 783 00:30:16,880 --> 00:30:19,459 the sampling rate, forty eight kilohertz 784 00:30:19,460 --> 00:30:21,619 you can unfolded four 785 00:30:21,620 --> 00:30:22,620 times 786 00:30:23,690 --> 00:30:25,969 and download prices are currently at 787 00:30:25,970 --> 00:30:28,039 2L, higher than 788 00:30:28,040 --> 00:30:30,169 for the original high relative high 789 00:30:30,170 --> 00:30:33,409 resolution audio, the 24 bits 790 00:30:33,410 --> 00:30:36,289 and sampled at 192 791 00:30:36,290 --> 00:30:37,609 kilohertz files. 792 00:30:39,800 --> 00:30:41,959 Important to note, the end user does not 793 00:30:41,960 --> 00:30:43,009 own the file. 794 00:30:44,960 --> 00:30:46,699 This has implications. 795 00:30:46,700 --> 00:30:48,949 For example, if you do a allow less 796 00:30:48,950 --> 00:30:51,469 loudness normalization on the file itself 797 00:30:51,470 --> 00:30:54,019 or a room compensation, you lose 798 00:30:54,020 --> 00:30:56,159 the you lose 799 00:30:56,160 --> 00:30:57,479 the MQ a 800 00:30:59,060 --> 00:31:01,189 license, so to say, and it's 801 00:31:01,190 --> 00:31:03,289 reduced to CD quality 802 00:31:03,290 --> 00:31:04,290 or below. 803 00:31:07,280 --> 00:31:10,029 So one of those claims in the 804 00:31:10,030 --> 00:31:12,229 U.S. paper, which is written at 805 00:31:12,230 --> 00:31:15,169 the bottom, is we advocate 806 00:31:15,170 --> 00:31:17,329 lossless compression, lossless processing 807 00:31:17,330 --> 00:31:19,489 and hierarchical updown 808 00:31:19,490 --> 00:31:21,469 sampling. We highlighted the quality and 809 00:31:21,470 --> 00:31:23,629 efficiency gains possible if the 810 00:31:23,630 --> 00:31:25,799 encoder and Dakoda are mutually 811 00:31:25,800 --> 00:31:28,309 aware and each matched to their analog 812 00:31:28,310 --> 00:31:29,310 converters. 813 00:31:30,840 --> 00:31:33,629 So this is a bold claim 814 00:31:33,630 --> 00:31:35,849 and not based on evidence, 815 00:31:35,850 --> 00:31:38,159 so we look into that 816 00:31:38,160 --> 00:31:40,949 lossless compression in audio means 817 00:31:40,950 --> 00:31:43,549 a redundant redundancy removal. 818 00:31:44,690 --> 00:31:47,119 And the possibility to reconstruct 819 00:31:47,120 --> 00:31:49,279 the original recording bit by 820 00:31:49,280 --> 00:31:51,409 bit, including the noise, 821 00:31:51,410 --> 00:31:53,029 whether it had been and the noise belongs 822 00:31:53,030 --> 00:31:54,030 to the music or not, 823 00:31:55,190 --> 00:31:58,609 lossy compression removes information, 824 00:31:58,610 --> 00:32:00,109 be it audible or not. 825 00:32:02,030 --> 00:32:04,639 So it's a lot less 826 00:32:04,640 --> 00:32:06,829 Shannons. Entropy allows to us allow 827 00:32:06,830 --> 00:32:09,919 us to assess the maximum redundancy 828 00:32:09,920 --> 00:32:12,469 in an audio file that can be removed 829 00:32:12,470 --> 00:32:14,210 without information loss. 830 00:32:15,380 --> 00:32:18,019 So on a data set of music, 831 00:32:18,020 --> 00:32:20,119 Stefan Hotto 832 00:32:20,120 --> 00:32:22,879 a link is given to his paper. 833 00:32:22,880 --> 00:32:25,249 Not a link, but the paper itself is given 834 00:32:25,250 --> 00:32:26,900 in the bottom of this slide. 835 00:32:28,430 --> 00:32:29,430 Did a 836 00:32:30,560 --> 00:32:33,109 entropy optimization and FLAC 837 00:32:33,110 --> 00:32:35,689 compression and reached for 838 00:32:35,690 --> 00:32:37,759 a certain high resolution 839 00:32:37,760 --> 00:32:38,760 data rate 840 00:32:40,040 --> 00:32:42,319 50 percent of 841 00:32:42,320 --> 00:32:43,369 a file size. 842 00:32:46,520 --> 00:32:49,219 So this is quite a lot, and 843 00:32:49,220 --> 00:32:51,319 nevertheless, eight, however, has 844 00:32:51,320 --> 00:32:54,439 a data reduction of 57 845 00:32:54,440 --> 00:32:55,440 percent. 846 00:33:00,790 --> 00:33:03,819 And it's it's 847 00:33:03,820 --> 00:33:06,309 obvious that this reduction, 848 00:33:06,310 --> 00:33:08,619 lowering the depth and changing amplitude 849 00:33:08,620 --> 00:33:12,219 and phase response is a loss to process. 850 00:33:12,220 --> 00:33:14,349 This is also highlighted in the 851 00:33:14,350 --> 00:33:17,559 paper that I give here by Stefan 852 00:33:17,560 --> 00:33:18,560 Otto. 853 00:33:22,120 --> 00:33:24,609 So them Slide 854 00:33:24,610 --> 00:33:27,049 16 exclaim, MQ 855 00:33:27,050 --> 00:33:29,199 approach of recoding results is 856 00:33:29,200 --> 00:33:31,599 superior in superior sound 857 00:33:31,600 --> 00:33:33,789 and significant lower data rate when 858 00:33:33,790 --> 00:33:36,069 compared to unstructured 859 00:33:36,070 --> 00:33:38,649 encoding and playback, and has been 860 00:33:38,650 --> 00:33:41,049 enthusiastically supported in listening 861 00:33:41,050 --> 00:33:43,299 trials with a number of recording 862 00:33:43,300 --> 00:33:45,219 and mastering engineers, artists and 863 00:33:45,220 --> 00:33:46,220 producers. 864 00:33:47,860 --> 00:33:48,860 So this is 865 00:33:49,960 --> 00:33:52,239 a truly scientific ruling that 866 00:33:52,240 --> 00:33:54,459 may give you an impression of 867 00:33:54,460 --> 00:33:57,339 the ASX publications 868 00:33:57,340 --> 00:33:58,340 that 869 00:33:59,560 --> 00:34:01,750 that we have as the resource 870 00:34:03,460 --> 00:34:05,679 and we analyze that. 871 00:34:05,680 --> 00:34:07,809 I quickly take you 872 00:34:07,810 --> 00:34:09,669 through the sampling process. 873 00:34:09,670 --> 00:34:11,739 On the left side, you see two columns 874 00:34:12,760 --> 00:34:13,749 in the top. 875 00:34:13,750 --> 00:34:17,138 You see a drawing of the sampling 876 00:34:17,139 --> 00:34:18,759 scheme. 877 00:34:18,760 --> 00:34:20,859 Fifty minutes. Yes. 878 00:34:20,860 --> 00:34:23,079 So we start with a blank, limited 879 00:34:23,080 --> 00:34:24,939 analog time function. 880 00:34:24,940 --> 00:34:27,939 And this is then multiplied in 881 00:34:27,940 --> 00:34:28,988 the time domain 882 00:34:30,730 --> 00:34:33,069 with a series of pulses. 883 00:34:33,070 --> 00:34:35,238 Then in the search row 884 00:34:35,239 --> 00:34:37,599 search line, 885 00:34:37,600 --> 00:34:41,099 we have the multiplication. 886 00:34:41,100 --> 00:34:43,178 Uh, the right column 887 00:34:43,179 --> 00:34:45,399 shows you the frequency 888 00:34:45,400 --> 00:34:47,469 equivalent and 889 00:34:48,520 --> 00:34:51,249 the time discrete signal 890 00:34:51,250 --> 00:34:54,569 gets quantized sized with certain steps, 891 00:34:54,570 --> 00:34:57,519 steps for high 892 00:34:57,520 --> 00:35:00,609 resolution audio. It's 24 bit per sample. 893 00:35:00,610 --> 00:35:02,829 And then if you put an analog 894 00:35:02,830 --> 00:35:05,019 Lopez filter on the on the signal, 895 00:35:05,020 --> 00:35:07,689 you can reconstruct the 896 00:35:07,690 --> 00:35:10,629 original analog signal 897 00:35:10,630 --> 00:35:14,159 bit by bit and 898 00:35:14,160 --> 00:35:16,869 at 24 bit practically 899 00:35:16,870 --> 00:35:18,709 equal to the original signal. 900 00:35:21,280 --> 00:35:23,469 What a change this 901 00:35:23,470 --> 00:35:26,159 is for. For example, the 902 00:35:26,160 --> 00:35:28,929 DB spline kernel 903 00:35:28,930 --> 00:35:32,079 for a sparse sampling scheme 904 00:35:32,080 --> 00:35:34,239 and this shall improve the 905 00:35:34,240 --> 00:35:35,350 temporal resolution. 906 00:35:37,670 --> 00:35:39,979 It's the highlight 907 00:35:39,980 --> 00:35:41,749 Cristoforetti already mentioned that they 908 00:35:41,750 --> 00:35:44,659 highlight the detection of 909 00:35:44,660 --> 00:35:46,859 humans, the ability to detect 910 00:35:46,860 --> 00:35:49,969 transgenes with 10 microseconds, 911 00:35:49,970 --> 00:35:52,339 and the colonel 912 00:35:52,340 --> 00:35:53,999 also saw match to the song. 913 00:35:55,130 --> 00:35:57,439 So a second 914 00:35:57,440 --> 00:36:00,709 point is they change the 915 00:36:00,710 --> 00:36:02,839 the low the 916 00:36:02,840 --> 00:36:05,599 low pass filter for avoiding 917 00:36:05,600 --> 00:36:07,909 pre ringing and 918 00:36:07,910 --> 00:36:10,549 for improving together with the kernel 919 00:36:10,550 --> 00:36:11,629 to time resolution. 920 00:36:11,630 --> 00:36:14,329 They call it deploring and 921 00:36:14,330 --> 00:36:16,429 compensation for signal distortions 922 00:36:16,430 --> 00:36:18,539 from the analog 923 00:36:18,540 --> 00:36:20,659 digital conversion during 924 00:36:20,660 --> 00:36:21,790 the recording process. 925 00:36:25,380 --> 00:36:27,449 So is it actually 926 00:36:27,450 --> 00:36:29,699 improving sound quality, the temporal 927 00:36:29,700 --> 00:36:31,859 resolution, or 928 00:36:31,860 --> 00:36:34,049 I should start with a listening test that 929 00:36:34,050 --> 00:36:36,909 Chris already mentioned, 930 00:36:36,910 --> 00:36:39,329 there was no preference 931 00:36:39,330 --> 00:36:42,059 towards EMCDDA compared 932 00:36:42,060 --> 00:36:45,269 to high resolution audio. 933 00:36:45,270 --> 00:36:47,579 In fact, in some situations, 934 00:36:47,580 --> 00:36:49,319 standard high resolution audio was 935 00:36:49,320 --> 00:36:50,320 preferred. 936 00:36:52,580 --> 00:36:54,709 So in terms of temporal resolution, is 937 00:36:54,710 --> 00:36:56,899 it improving or 938 00:36:56,900 --> 00:36:57,900 sound quality? 939 00:37:00,140 --> 00:37:01,849 If you look at their 940 00:37:03,020 --> 00:37:05,689 claim, it shall 941 00:37:05,690 --> 00:37:07,939 allow for the detection of events 942 00:37:07,940 --> 00:37:10,819 that are as short as 10 microseconds, 943 00:37:10,820 --> 00:37:12,889 which would be a bandwidth of 944 00:37:12,890 --> 00:37:14,839 100 megahertz. 945 00:37:14,840 --> 00:37:17,299 They call it blurring from forty 946 00:37:17,300 --> 00:37:19,219 eight kilohertz sampled signal. 947 00:37:19,220 --> 00:37:20,989 This is questionable. 948 00:37:20,990 --> 00:37:23,359 So I don't think that's actually 949 00:37:23,360 --> 00:37:24,360 the case. 950 00:37:27,950 --> 00:37:30,319 The next point is, is it tuned 951 00:37:30,320 --> 00:37:32,449 to the song at first they 952 00:37:32,450 --> 00:37:34,909 call it the day they post 953 00:37:34,910 --> 00:37:36,979 to change the balance between 954 00:37:36,980 --> 00:37:38,809 time and frequency representation of 955 00:37:38,810 --> 00:37:40,199 their filters. 956 00:37:40,200 --> 00:37:43,639 Both are equivalent representations 957 00:37:43,640 --> 00:37:45,739 of a filter, but if they use an 958 00:37:45,740 --> 00:37:46,740 adaptive filter. 959 00:37:48,150 --> 00:37:50,009 We don't know actually what is what is 960 00:37:50,010 --> 00:37:51,329 tuned to the songs, so there's 961 00:37:51,330 --> 00:37:53,429 information missing in terms 962 00:37:53,430 --> 00:37:55,409 of the minimum phase photos to suppress 963 00:37:55,410 --> 00:37:56,410 pre ringing 964 00:37:58,740 --> 00:38:00,239 minimum face photos have their 965 00:38:00,240 --> 00:38:02,579 shortcomings. They change the 966 00:38:02,580 --> 00:38:05,249 frequency response, the 967 00:38:05,250 --> 00:38:07,499 the amplitude and the phase response 968 00:38:07,500 --> 00:38:09,539 that has been also shown by Stephane 969 00:38:09,540 --> 00:38:12,449 Hotto and 970 00:38:12,450 --> 00:38:13,450 in the last. 971 00:38:16,560 --> 00:38:18,149 The last point I want to make on that 972 00:38:18,150 --> 00:38:21,329 slide is the deputization 973 00:38:21,330 --> 00:38:23,879 and in the time domain 974 00:38:23,880 --> 00:38:26,309 to avoid the ringing 975 00:38:26,310 --> 00:38:28,289 of the filters, 976 00:38:29,520 --> 00:38:31,589 the minimum phase has 977 00:38:31,590 --> 00:38:34,049 ringing after the center of the 978 00:38:34,050 --> 00:38:36,209 filter and 979 00:38:36,210 --> 00:38:39,359 to improve also temporal resolution 980 00:38:39,360 --> 00:38:41,429 and to compensate for the 981 00:38:41,430 --> 00:38:42,430 signal change. 982 00:38:43,410 --> 00:38:45,569 That is also a point Stefanutto 983 00:38:45,570 --> 00:38:47,939 looked at and he clearly shows 984 00:38:47,940 --> 00:38:49,589 that polarization. 985 00:38:49,590 --> 00:38:51,869 So windowing, in 986 00:38:51,870 --> 00:38:54,239 fact, is a Lopez filter and not 987 00:38:54,240 --> 00:38:55,589 improving sound quality. 988 00:38:57,820 --> 00:39:00,369 So do we need extra 989 00:39:00,370 --> 00:39:02,409 enhancement based on today's evidence 990 00:39:02,410 --> 00:39:04,509 from neuroscience and 991 00:39:04,510 --> 00:39:06,699 we heard we need a 992 00:39:06,700 --> 00:39:08,889 time resolution in a paper 993 00:39:08,890 --> 00:39:11,530 of six microseconds 994 00:39:13,540 --> 00:39:16,209 and it was referred 995 00:39:16,210 --> 00:39:18,519 to the binaural hearing. 996 00:39:18,520 --> 00:39:20,769 And in fact, we are able 997 00:39:20,770 --> 00:39:23,619 to discriminate into all time differences 998 00:39:23,620 --> 00:39:25,989 with just noticeable differences of 999 00:39:25,990 --> 00:39:27,850 10 microseconds. 1000 00:39:29,560 --> 00:39:31,779 Those differences are, 1001 00:39:31,780 --> 00:39:34,269 however, based on a central auditory 1002 00:39:34,270 --> 00:39:36,849 comparison of sound waves 1003 00:39:36,850 --> 00:39:38,919 between the ears and 1004 00:39:38,920 --> 00:39:41,709 because of the spatial nyqvist limit. 1005 00:39:41,710 --> 00:39:43,749 Only defined structure of the frequency 1006 00:39:43,750 --> 00:39:46,089 up to one point three kilohertz 1007 00:39:46,090 --> 00:39:47,589 is analyzed. 1008 00:39:47,590 --> 00:39:50,409 And if we remember, the 1009 00:39:50,410 --> 00:39:53,019 department, for example, has of course 1010 00:39:53,020 --> 00:39:54,519 years that are closer together. 1011 00:39:54,520 --> 00:39:57,129 And that's why they 1012 00:39:57,130 --> 00:39:58,780 can even detect 1013 00:40:00,430 --> 00:40:02,699 differences that are smaller then 1014 00:40:02,700 --> 00:40:04,989 than in humans. 1015 00:40:04,990 --> 00:40:07,179 So and from the media 1016 00:40:07,180 --> 00:40:09,279 production side, there's no 1017 00:40:09,280 --> 00:40:11,349 need for an enhancement of audio from 1018 00:40:11,350 --> 00:40:13,209 a scientific point, as well as from an 1019 00:40:13,210 --> 00:40:14,709 artistic point of view. 1020 00:40:14,710 --> 00:40:17,859 The artistic concept with all variables, 1021 00:40:17,860 --> 00:40:20,439 composition, performance, practice, etc., 1022 00:40:20,440 --> 00:40:22,779 is captured on multichannel, multichannel 1023 00:40:22,780 --> 00:40:24,849 files. And the transmission chain 1024 00:40:24,850 --> 00:40:27,009 has always been kept free from 1025 00:40:27,010 --> 00:40:28,659 influences. 1026 00:40:28,660 --> 00:40:30,759 Quote, The music is the message 1027 00:40:30,760 --> 00:40:33,399 and finally interpreted 1028 00:40:33,400 --> 00:40:35,319 by the interpreters, by the auditory 1029 00:40:35,320 --> 00:40:37,509 system. So we don't need a channel 1030 00:40:37,510 --> 00:40:40,929 that is preprocessing 1031 00:40:40,930 --> 00:40:43,269 what's going on in our heads eventually. 1032 00:40:45,010 --> 00:40:47,229 A general question, do we need a 1033 00:40:47,230 --> 00:40:49,359 HRR audio, high resolution audio 1034 00:40:49,360 --> 00:40:51,099 based on today's evidence from 1035 00:40:51,100 --> 00:40:54,159 neuroscience for listening to music? 1036 00:40:54,160 --> 00:40:56,589 And I have here a 1037 00:40:56,590 --> 00:40:58,719 graph from Speaker and Fasel and 1038 00:40:58,720 --> 00:41:00,129 psychoacoustics. 1039 00:41:00,130 --> 00:41:02,559 So the audible threshold in 1040 00:41:02,560 --> 00:41:04,929 quiet is this 1041 00:41:04,930 --> 00:41:07,179 solid line and it shows clearly we 1042 00:41:07,180 --> 00:41:09,669 are not able to listen 1043 00:41:09,670 --> 00:41:12,099 beyond, let's say, 20 1044 00:41:12,100 --> 00:41:13,100 kilohertz. 1045 00:41:14,380 --> 00:41:17,019 Moreover, music has a 1046 00:41:17,020 --> 00:41:19,299 amplitude spectrum 1047 00:41:19,300 --> 00:41:22,029 with one over F, so 1048 00:41:22,030 --> 00:41:24,279 beyond 20 kilohertz, 1049 00:41:24,280 --> 00:41:26,439 the sound power is very faint and 1050 00:41:26,440 --> 00:41:28,599 it's from my point 1051 00:41:28,600 --> 00:41:30,370 of view, hardly. 1052 00:41:31,660 --> 00:41:33,789 I can hardly imagine that we are 1053 00:41:33,790 --> 00:41:34,790 able to listen 1054 00:41:36,340 --> 00:41:38,169 beyond 20 kilohertz. 1055 00:41:38,170 --> 00:41:40,239 Nevertheless, in the 1056 00:41:40,240 --> 00:41:42,339 production on the production side, 1057 00:41:42,340 --> 00:41:44,919 we use filters that 1058 00:41:44,920 --> 00:41:47,889 perform interpolation and for that case 1059 00:41:47,890 --> 00:41:49,749 it's good to have high, really high 1060 00:41:49,750 --> 00:41:50,949 resolution audio. 1061 00:41:50,950 --> 00:41:53,589 Another point is conservation 1062 00:41:53,590 --> 00:41:56,649 of of recordings. 1063 00:41:56,650 --> 00:41:59,349 So if we store music 1064 00:41:59,350 --> 00:42:01,509 digitally, then of course it's 1065 00:42:01,510 --> 00:42:03,579 good to have a lot of 1066 00:42:03,580 --> 00:42:06,609 samples such that we can interpolate 1067 00:42:06,610 --> 00:42:08,679 once there are errors 1068 00:42:08,680 --> 00:42:09,879 arising. 1069 00:42:09,880 --> 00:42:11,889 So conclusion, there is no evidence based 1070 00:42:11,890 --> 00:42:13,659 research that supports audio quality 1071 00:42:13,660 --> 00:42:16,599 claims of McCooey MKC 1072 00:42:16,600 --> 00:42:18,909 losse because it reduces 1073 00:42:18,910 --> 00:42:21,159 Tibbett depths and the frequency response 1074 00:42:22,240 --> 00:42:24,339 with the degradation of audio 1075 00:42:24,340 --> 00:42:26,619 quality, which in 1076 00:42:26,620 --> 00:42:28,629 its current version might not always be 1077 00:42:28,630 --> 00:42:30,789 audible to algorithm of a 1078 00:42:30,790 --> 00:42:32,559 folding scheme, is only vaguely 1079 00:42:32,560 --> 00:42:34,839 described, and its disadvantages 1080 00:42:34,840 --> 00:42:37,239 for high risk, high resolution 1081 00:42:37,240 --> 00:42:39,249 audio are obviously disguised by 1082 00:42:39,250 --> 00:42:42,039 marketing slogans like The Blurring 1083 00:42:42,040 --> 00:42:44,349 Once and MTA infrastructure 1084 00:42:44,350 --> 00:42:46,089 is established, audio quality can be 1085 00:42:46,090 --> 00:42:48,369 easily scaled down and a description 1086 00:42:48,370 --> 00:42:51,039 of kernel optimization for each track 1087 00:42:51,040 --> 00:42:53,499 song alludes to a perceptual 1088 00:42:53,500 --> 00:42:55,959 coding scheme which might eventually 1089 00:42:55,960 --> 00:42:58,250 end up at an auditory quality. 1090 00:42:59,530 --> 00:43:02,139 One thing to remember, there are free 1091 00:43:02,140 --> 00:43:03,969 high rollers, high resolution audio 1092 00:43:03,970 --> 00:43:06,849 codecs that compress audio lawlessly 1093 00:43:06,850 --> 00:43:08,379 like FLAC. 1094 00:43:08,380 --> 00:43:10,509 Internet bandwidth is not a big concern 1095 00:43:10,510 --> 00:43:12,039 and the Western world anymore. 1096 00:43:12,040 --> 00:43:13,629 It increases. 1097 00:43:13,630 --> 00:43:15,699 Hence there's no need to 1098 00:43:15,700 --> 00:43:17,439 follow the audio coding. 1099 00:43:17,440 --> 00:43:19,119 And most of the music, as Chris have 1100 00:43:19,120 --> 00:43:21,429 already told us, is still 1101 00:43:21,430 --> 00:43:23,829 stored in CD quality. 1102 00:43:23,830 --> 00:43:26,859 If we sample 1103 00:43:26,860 --> 00:43:29,199 those, um, those that 1104 00:43:29,200 --> 00:43:31,269 information signal to 1105 00:43:31,270 --> 00:43:33,349 noise ratio decreases. 1106 00:43:33,350 --> 00:43:34,350 OK, thanks. 1107 00:43:46,770 --> 00:43:48,929 All right, so 1108 00:43:48,930 --> 00:43:49,930 things aren't on 1109 00:43:51,080 --> 00:43:52,290 cash or that. 1110 00:43:54,240 --> 00:43:56,279 Good. So the question is, where is the 1111 00:43:56,280 --> 00:43:57,280 and 1112 00:43:58,710 --> 00:44:01,139 basically Enten already has showed that 1113 00:44:01,140 --> 00:44:03,719 MQ employs a clever encoding scheme. 1114 00:44:03,720 --> 00:44:05,789 And as far as we 1115 00:44:05,790 --> 00:44:07,499 know, because this is proprietary 1116 00:44:07,500 --> 00:44:09,569 technology and we have to work 1117 00:44:09,570 --> 00:44:11,189 with what we got, and that is reverse 1118 00:44:11,190 --> 00:44:13,229 engineering by some people online and 1119 00:44:13,230 --> 00:44:15,209 some of the papers and other 1120 00:44:15,210 --> 00:44:16,619 publications. 1121 00:44:16,620 --> 00:44:18,689 That makes this is, again, from 1122 00:44:18,690 --> 00:44:21,059 the paper that Anton just mentioned 1123 00:44:21,060 --> 00:44:22,060 for Stefanutto, 1124 00:44:23,190 --> 00:44:26,429 where he drew up 1125 00:44:26,430 --> 00:44:28,199 the MQ encoding scheme. 1126 00:44:28,200 --> 00:44:29,879 And important part, I don't have a laser 1127 00:44:29,880 --> 00:44:32,009 pointer here, but is the PCM simple up 1128 00:44:32,010 --> 00:44:34,679 there, which you can see is 24 bits. 1129 00:44:34,680 --> 00:44:36,959 And as you can furthermore see, 1130 00:44:36,960 --> 00:44:39,389 it's divided into two parts 1131 00:44:39,390 --> 00:44:41,469 of baseband. It's the audio bits and 1132 00:44:41,470 --> 00:44:43,409 the serpent is the data bits. 1133 00:44:43,410 --> 00:44:45,659 So basically what Abakua does 1134 00:44:45,660 --> 00:44:48,300 is using a 24 bit sample, 1135 00:44:50,910 --> 00:44:52,889 throwing away eight bits of that, more or 1136 00:44:52,890 --> 00:44:55,289 less eight bits of that audio 1137 00:44:55,290 --> 00:44:58,739 information and very 1138 00:44:58,740 --> 00:45:00,929 other stuff basically in 1139 00:45:00,930 --> 00:45:03,809 the noise floor that a 24 bit sample 1140 00:45:03,810 --> 00:45:04,810 allows for. 1141 00:45:07,110 --> 00:45:08,110 Give me a second. 1142 00:45:10,710 --> 00:45:11,710 So 1143 00:45:13,500 --> 00:45:15,689 what if you look at the history 1144 00:45:15,690 --> 00:45:17,909 I just gave you on digital audio, this 1145 00:45:17,910 --> 00:45:19,919 isn't a way back to square one because we 1146 00:45:19,920 --> 00:45:22,019 had in 81 with the city Redbook standard 1147 00:45:22,020 --> 00:45:24,329 16 bits at 44 one kilohertz. 1148 00:45:24,330 --> 00:45:26,999 What we now would get with Makua 1149 00:45:27,000 --> 00:45:30,269 is, in their own words, right. 1150 00:45:30,270 --> 00:45:32,669 Something to the extent of 15 to 17 1151 00:45:32,670 --> 00:45:34,739 bits of dynamic 1152 00:45:34,740 --> 00:45:36,809 bandwidth sampled 1153 00:45:36,810 --> 00:45:39,029 at forty eight kilohertz, everything 1154 00:45:39,030 --> 00:45:40,530 about 48 kilohertz 1155 00:45:41,910 --> 00:45:44,129 is encoded in a lossy compression 1156 00:45:44,130 --> 00:45:46,259 scheme, as Anton just showed the sisters 1157 00:45:46,260 --> 00:45:48,839 folding what they talk about, 1158 00:45:48,840 --> 00:45:50,879 how this actually is accomplished in 1159 00:45:50,880 --> 00:45:51,869 detail, we don't know. 1160 00:45:51,870 --> 00:45:54,269 It's supposedly I 1161 00:45:54,270 --> 00:45:56,339 guess it's a pretty clever way 1162 00:45:56,340 --> 00:45:58,469 because you have to basically get 1163 00:45:58,470 --> 00:46:00,779 spectrum from 24 kilohertz to 96 1164 00:46:00,780 --> 00:46:02,969 kilohertz into just a few bits and 1165 00:46:02,970 --> 00:46:05,309 bury that in the 1166 00:46:05,310 --> 00:46:07,859 six, seven, eight bits that you get 1167 00:46:07,860 --> 00:46:09,390 with a compression. 1168 00:46:10,440 --> 00:46:12,869 So in a way, MQ 1169 00:46:12,870 --> 00:46:15,269 A acknowledges 1170 00:46:15,270 --> 00:46:17,309 that the standard that Sony and Philips 1171 00:46:17,310 --> 00:46:19,409 arrived at sixteen bits, forty four 1172 00:46:19,410 --> 00:46:21,599 one K is good enough for human 1173 00:46:21,600 --> 00:46:24,069 hearing and all the information above 16 1174 00:46:24,070 --> 00:46:25,799 billion above forty four or forty eight 1175 00:46:25,800 --> 00:46:28,019 kilohertz. And it is more 1176 00:46:28,020 --> 00:46:30,119 or less useless, not 1177 00:46:30,120 --> 00:46:32,159 audible. We can do something else with 1178 00:46:32,160 --> 00:46:34,529 that. So what do you do, 1179 00:46:34,530 --> 00:46:36,049 what do you put there. 1180 00:46:36,050 --> 00:46:38,519 Um and it's a fairly complex 1181 00:46:38,520 --> 00:46:40,799 and sophisticated cryptographic scheme. 1182 00:46:40,800 --> 00:46:42,239 Oh wait. I forgot this one. 1183 00:46:42,240 --> 00:46:44,459 This is the blue light on the, on 1184 00:46:44,460 --> 00:46:46,859 the on one of these 1185 00:46:46,860 --> 00:46:49,289 MQ adex that lights up 1186 00:46:49,290 --> 00:46:51,209 when the file is authenticated 1187 00:46:52,950 --> 00:46:55,049 and authentication to remind you of that 1188 00:46:55,050 --> 00:46:56,050 it's a loaded term 1189 00:46:58,590 --> 00:46:59,849 in the digital realm. 1190 00:46:59,850 --> 00:47:01,769 It promises to verify the origin 1191 00:47:03,180 --> 00:47:05,249 of the data that so who is the 1192 00:47:05,250 --> 00:47:08,069 sender and integrity 1193 00:47:08,070 --> 00:47:10,199 of the data under the conditions of open 1194 00:47:10,200 --> 00:47:12,359 networks where data can be altered 1195 00:47:12,360 --> 00:47:14,040 at any point in the transmission. 1196 00:47:15,160 --> 00:47:17,229 And these requirements can only be met 1197 00:47:17,230 --> 00:47:18,940 by a cryptographically means, 1198 00:47:20,290 --> 00:47:22,509 as the continuous continuous reports 1199 00:47:22,510 --> 00:47:24,070 on hacked cryptos indicate, 1200 00:47:25,360 --> 00:47:27,489 MKR has to come up with a workable and 1201 00:47:27,490 --> 00:47:29,079 sufficiently secure system 1202 00:47:30,130 --> 00:47:32,559 to to to make this possible. 1203 00:47:32,560 --> 00:47:34,719 And MQ a partnered with the German data 1204 00:47:34,720 --> 00:47:36,819 security company TIMAKOVA 1205 00:47:36,820 --> 00:47:38,109 Macau to create an authentication 1206 00:47:38,110 --> 00:47:40,269 component for the format, which 1207 00:47:40,270 --> 00:47:42,579 includes a major purveyor 1208 00:47:42,580 --> 00:47:43,929 of cryptographic products and well 1209 00:47:43,930 --> 00:47:45,609 established in the banking sector in 1210 00:47:45,610 --> 00:47:47,379 Europe and also in the health care 1211 00:47:47,380 --> 00:47:47,749 sector. 1212 00:47:47,750 --> 00:47:49,689 And this is the website for DeMarco. 1213 00:47:49,690 --> 00:47:51,339 So they offer lawful interception data 1214 00:47:51,340 --> 00:47:53,289 retention services, which is nice to 1215 00:47:53,290 --> 00:47:55,119 have, and they offer these hardware 1216 00:47:55,120 --> 00:47:57,399 security modules, 1217 00:47:57,400 --> 00:47:59,479 which actually would make a 1218 00:47:59,480 --> 00:48:00,480 user's 1219 00:48:02,890 --> 00:48:05,799 automatical. Also published a white paper 1220 00:48:05,800 --> 00:48:07,809 on MQ A it's available on the website. 1221 00:48:07,810 --> 00:48:10,329 It's very sparse on details, 1222 00:48:10,330 --> 00:48:12,399 but it suggests that Emeco 1223 00:48:12,400 --> 00:48:14,709 and tales of public key infrastructure 1224 00:48:14,710 --> 00:48:16,809 and that hardware security modules like 1225 00:48:16,810 --> 00:48:18,969 this one shown here will 1226 00:48:18,970 --> 00:48:21,459 be deployed in at least the Makua 1227 00:48:21,460 --> 00:48:23,949 and coding houses and properly 1228 00:48:23,950 --> 00:48:25,479 later on in the mastering studio. 1229 00:48:25,480 --> 00:48:27,609 So the convertors in 1230 00:48:27,610 --> 00:48:29,859 mastering studios would need to 1231 00:48:29,860 --> 00:48:32,289 have this hardware security modules 1232 00:48:32,290 --> 00:48:33,789 in them or some sort 1233 00:48:35,170 --> 00:48:37,179 of hardware security modules derived from 1234 00:48:37,180 --> 00:48:39,249 from these devices and 1235 00:48:39,250 --> 00:48:41,109 automatical. Was IMNSHO nice? 1236 00:48:41,110 --> 00:48:43,509 Is it to provide hardware 1237 00:48:43,510 --> 00:48:45,549 security modules for Dummies book on 1238 00:48:45,550 --> 00:48:48,009 their site, which apparently audio 1239 00:48:48,010 --> 00:48:49,690 so far I haven't been able to find 1240 00:48:51,820 --> 00:48:54,399 and is hardly a secure module. 1241 00:48:54,400 --> 00:48:56,529 Security modules basically provide 1242 00:48:56,530 --> 00:48:59,259 the means to embed the necessary 1243 00:48:59,260 --> 00:49:01,419 keys in the encoded 1244 00:49:01,420 --> 00:49:02,509 files. 1245 00:49:02,510 --> 00:49:04,449 The exact means of this key management 1246 00:49:04,450 --> 00:49:06,459 schemes and the handling of the route of 1247 00:49:06,460 --> 00:49:09,339 trust is unknown at this point. 1248 00:49:09,340 --> 00:49:11,409 MQM might outsource this to what 1249 00:49:11,410 --> 00:49:13,569 you might call or might use 1250 00:49:13,570 --> 00:49:14,829 this crypto servers that 1251 00:49:15,880 --> 00:49:17,149 do the key management. 1252 00:49:17,150 --> 00:49:18,340 OK, I have to come to an end 1253 00:49:20,740 --> 00:49:22,630 in-house, so 1254 00:49:24,460 --> 00:49:26,079 I'd like to ask the audience here. 1255 00:49:26,080 --> 00:49:27,099 Is this DRM? 1256 00:49:31,020 --> 00:49:32,020 I don't hear anybody 1257 00:49:33,090 --> 00:49:35,249 I guess that means yes, yes, 1258 00:49:35,250 --> 00:49:37,799 OK, so I'm asking this 1259 00:49:37,800 --> 00:49:39,959 because Akua claims, of 1260 00:49:39,960 --> 00:49:42,089 course it's otherwise 1261 00:49:42,090 --> 00:49:43,619 they say it's not DIAM. 1262 00:49:43,620 --> 00:49:45,209 And here's from ARS Technica. 1263 00:49:46,290 --> 00:49:47,309 Abakua has no idea. 1264 00:49:47,310 --> 00:49:48,839 Uncommon application outside of the 1265 00:49:48,840 --> 00:49:50,609 studio. We have no opinion on this beyond 1266 00:49:50,610 --> 00:49:53,159 nothing, noting that DRM is a futile 1267 00:49:53,160 --> 00:49:54,269 exercise. 1268 00:49:54,270 --> 00:49:55,829 And this is like a large part of the 1269 00:49:55,830 --> 00:49:57,569 narrative and the online debate because 1270 00:49:57,570 --> 00:49:59,519 EMCDDA and all the people involved, 1271 00:49:59,520 --> 00:50:01,469 including the all you press, keeps 1272 00:50:01,470 --> 00:50:03,599 touting this is not DRM because 1273 00:50:03,600 --> 00:50:04,949 you can copy the files. 1274 00:50:04,950 --> 00:50:05,939 Right. 1275 00:50:05,940 --> 00:50:07,800 But copy protection is just one form of 1276 00:50:08,900 --> 00:50:10,349 the goal of DRM. 1277 00:50:12,090 --> 00:50:13,559 If you just look it up on Wikipedia, for 1278 00:50:13,560 --> 00:50:15,559 example, it is the first thing you see is 1279 00:50:15,560 --> 00:50:16,919 the rights of a user with data. 1280 00:50:16,920 --> 00:50:19,139 And this means both in a restrictive and 1281 00:50:19,140 --> 00:50:21,989 a permissive use cases. 1282 00:50:21,990 --> 00:50:24,269 So NCUA permits copying, 1283 00:50:24,270 --> 00:50:25,379 but it restricts 1284 00:50:26,580 --> 00:50:28,409 access to parts of the file, namely the 1285 00:50:28,410 --> 00:50:29,610 higher resolution parts. 1286 00:50:30,960 --> 00:50:33,839 So India and its conditional access 1287 00:50:33,840 --> 00:50:35,809 and the access only granted if you have 1288 00:50:35,810 --> 00:50:38,669 an acceptable digital audio converter, 1289 00:50:38,670 --> 00:50:41,279 which you have paid for, of course. 1290 00:50:41,280 --> 00:50:43,349 So I guess what we can say is 1291 00:50:43,350 --> 00:50:45,299 that Amcu is something like a freemium 1292 00:50:45,300 --> 00:50:46,199 model, right? 1293 00:50:46,200 --> 00:50:48,269 You get the file that you can play 1294 00:50:48,270 --> 00:50:50,549 on any kind of convertors that 1295 00:50:50,550 --> 00:50:52,799 you can vote on, it actually would 1296 00:50:52,800 --> 00:50:54,269 be playable. 1297 00:50:54,270 --> 00:50:55,859 But if you want access to the premium 1298 00:50:55,860 --> 00:50:57,899 part of that file, you have to buy a 1299 00:50:57,900 --> 00:51:00,209 device or software that allows 1300 00:51:00,210 --> 00:51:01,210 that. 1301 00:51:02,310 --> 00:51:04,209 So why do they do that? 1302 00:51:04,210 --> 00:51:06,419 Oh, one more point and then I'm going 1303 00:51:06,420 --> 00:51:08,759 to end the 1304 00:51:08,760 --> 00:51:10,620 film as it has been 1305 00:51:12,200 --> 00:51:14,549 a reverse engineered so far shows 1306 00:51:14,550 --> 00:51:15,989 that they actually have provisions in 1307 00:51:15,990 --> 00:51:18,359 place for copyright protection. 1308 00:51:18,360 --> 00:51:20,489 Um, there are provisions there 1309 00:51:20,490 --> 00:51:22,709 to encrypt 1310 00:51:22,710 --> 00:51:24,989 the files and 1311 00:51:24,990 --> 00:51:27,209 with all sorts of 20 cipha. 1312 00:51:27,210 --> 00:51:29,069 And there's also Britton's transposition 1313 00:51:29,070 --> 00:51:30,070 schemes in there. 1314 00:51:31,770 --> 00:51:33,749 Why do they do this? 1315 00:51:33,750 --> 00:51:36,899 Jimmy Eivind just said to Apple Music 1316 00:51:36,900 --> 00:51:38,879 boss that the streaming services have a 1317 00:51:38,880 --> 00:51:40,949 bad situation. There's no margins. 1318 00:51:40,950 --> 00:51:42,029 They don't make any money. 1319 00:51:42,030 --> 00:51:43,679 So he's talking about titles, Spotify and 1320 00:51:43,680 --> 00:51:45,029 so forth. 1321 00:51:45,030 --> 00:51:47,249 Um, whereas Amazon sales 1322 00:51:47,250 --> 00:51:49,649 by Prime Apple sales, telephones 1323 00:51:49,650 --> 00:51:51,239 and iPads and so forth. 1324 00:51:51,240 --> 00:51:53,099 So the problem is that you have to sell 1325 00:51:53,100 --> 00:51:55,199 something else outside of 1326 00:51:55,200 --> 00:51:56,759 your streaming service in order to make 1327 00:51:56,760 --> 00:51:58,139 money, which of course, everyone would 1328 00:51:58,140 --> 00:51:59,429 place. 1329 00:51:59,430 --> 00:52:01,799 Um, Amazon 1330 00:52:01,800 --> 00:52:04,109 is currently the biggest speaker 1331 00:52:04,110 --> 00:52:05,549 company in the world with these kinds of 1332 00:52:05,550 --> 00:52:07,159 devices, and Google is catching up. 1333 00:52:08,160 --> 00:52:10,679 So our interpretation 1334 00:52:10,680 --> 00:52:12,899 of what it's trying to do is 1335 00:52:12,900 --> 00:52:15,059 it's trying to generate a new licensing 1336 00:52:15,060 --> 00:52:17,159 stream for the music industry or 1337 00:52:17,160 --> 00:52:18,629 specifically for Makua. 1338 00:52:20,670 --> 00:52:22,739 These elements 1339 00:52:22,740 --> 00:52:24,839 of the music production, distribution and 1340 00:52:24,840 --> 00:52:27,059 consumption value chain 1341 00:52:27,060 --> 00:52:29,459 would need EMCDDA licenses, the recording 1342 00:52:29,460 --> 00:52:31,529 software and plug ins, the recording 1343 00:52:31,530 --> 00:52:34,169 Marsing Studios new convertors. 1344 00:52:34,170 --> 00:52:36,739 The artists would get their 1345 00:52:36,740 --> 00:52:38,999 back catalogs remastered and 1346 00:52:39,000 --> 00:52:40,349 would have to agree on that. 1347 00:52:40,350 --> 00:52:42,269 Streaming services would have to sign up 1348 00:52:42,270 --> 00:52:43,649 with them. 1349 00:52:43,650 --> 00:52:45,419 Physical media manufacturers if it comes 1350 00:52:45,420 --> 00:52:46,739 to that. And of course, all the hyphae 1351 00:52:46,740 --> 00:52:48,929 manufacturers and 1352 00:52:48,930 --> 00:52:51,809 customers also are looking 1353 00:52:51,810 --> 00:52:53,489 at paying MTU out. 1354 00:52:55,440 --> 00:52:57,629 OK, let's skip all this. 1355 00:52:57,630 --> 00:52:59,519 Why is that clever ways of stealth? 1356 00:52:59,520 --> 00:53:01,619 I think the encoding schemes 1357 00:53:01,620 --> 00:53:03,330 that Amcu employees 1358 00:53:04,380 --> 00:53:06,029 make very clever use of the space 1359 00:53:06,030 --> 00:53:08,519 afforded by 24 bit samples, 1360 00:53:08,520 --> 00:53:10,949 namely to stealthily 1361 00:53:10,950 --> 00:53:13,319 introduce a DRM in the music, 1362 00:53:13,320 --> 00:53:15,989 uh, um, production, 1363 00:53:15,990 --> 00:53:18,929 distribution and consumption value 1364 00:53:18,930 --> 00:53:21,149 chain. And they don't openly talk 1365 00:53:21,150 --> 00:53:22,739 about it and it makes it actually very, 1366 00:53:22,740 --> 00:53:24,989 very difficult to find these elements. 1367 00:53:24,990 --> 00:53:27,059 And if you ask what 1368 00:53:27,060 --> 00:53:29,819 to do to that in this context of the 1369 00:53:29,820 --> 00:53:31,799 conference here, it would be worthwhile 1370 00:53:31,800 --> 00:53:33,869 if people start to dig deeper on 1371 00:53:33,870 --> 00:53:35,340 what Amcu is actually doing. 1372 00:53:36,390 --> 00:53:37,499 OK, I just stop here. 1373 00:53:37,500 --> 00:53:38,500 Thank you. 1374 00:53:46,910 --> 00:53:49,079 So thank you very much for the talk 1375 00:53:49,080 --> 00:53:51,229 we have now like five minutes for 1376 00:53:51,230 --> 00:53:53,299 questioning and answering, so please line 1377 00:53:53,300 --> 00:53:54,679 up to the microphones. 1378 00:53:54,680 --> 00:53:56,269 And do we have a question from the 1379 00:53:56,270 --> 00:53:57,469 Internet? 1380 00:53:57,470 --> 00:53:58,470 No. 1381 00:54:06,550 --> 00:54:07,989 So microphone number one, 1382 00:54:09,790 --> 00:54:10,709 hello. 1383 00:54:10,710 --> 00:54:11,710 Thanks for your talk. 1384 00:54:13,720 --> 00:54:16,059 You talked about the reverse engineering 1385 00:54:16,060 --> 00:54:18,459 of the firmware, so it somehow 1386 00:54:18,460 --> 00:54:20,379 sounds that they wouldn't have these 1387 00:54:20,380 --> 00:54:22,359 complicated hardware security modules in 1388 00:54:22,360 --> 00:54:23,360 the techs. 1389 00:54:24,940 --> 00:54:27,369 Well, it's a specific 1390 00:54:27,370 --> 00:54:28,899 firmware in one product. 1391 00:54:28,900 --> 00:54:30,529 It has been reverse engineered from 1392 00:54:30,530 --> 00:54:31,530 boozehound. 1393 00:54:32,710 --> 00:54:34,809 I don't know the 1394 00:54:34,810 --> 00:54:37,329 security provisions of that, but 1395 00:54:37,330 --> 00:54:39,909 there's a lot of security modules 1396 00:54:39,910 --> 00:54:41,979 won't be at this point part of 1397 00:54:41,980 --> 00:54:43,809 the consumer products. 1398 00:54:43,810 --> 00:54:46,279 They will be part of the convertors 1399 00:54:46,280 --> 00:54:49,389 encoding house and the distribution 1400 00:54:49,390 --> 00:54:52,059 context, um, 1401 00:54:52,060 --> 00:54:54,459 how the exact, uh, crypto 1402 00:54:54,460 --> 00:54:57,069 works between DMCA 1403 00:54:57,070 --> 00:54:58,870 files and 1404 00:54:59,950 --> 00:55:02,999 the consumer docs is not yet known. 1405 00:55:03,000 --> 00:55:04,899 There's some hints and indications you 1406 00:55:04,900 --> 00:55:07,029 will find this in the resources of how 1407 00:55:07,030 --> 00:55:08,349 that works. 1408 00:55:08,350 --> 00:55:10,659 Um, but the hardware security models 1409 00:55:10,660 --> 00:55:12,249 would be employed in the in the 1410 00:55:12,250 --> 00:55:14,439 production distribution part of 1411 00:55:14,440 --> 00:55:15,910 the, um, 1412 00:55:17,170 --> 00:55:18,170 of the scheme. 1413 00:55:19,540 --> 00:55:20,949 OK, thanks. 1414 00:55:20,950 --> 00:55:22,749 I have another question from microphone, 1415 00:55:22,750 --> 00:55:24,669 phone number four. 1416 00:55:24,670 --> 00:55:26,109 Yeah. Thank you for a talk. 1417 00:55:26,110 --> 00:55:27,759 Um, I have two questions. 1418 00:55:27,760 --> 00:55:29,889 One of them is, do you think 1419 00:55:29,890 --> 00:55:32,169 that this marketing concept will 1420 00:55:32,170 --> 00:55:34,449 have any success at all? 1421 00:55:34,450 --> 00:55:36,789 And the second one is, um, 1422 00:55:36,790 --> 00:55:39,159 do you think that music 1423 00:55:39,160 --> 00:55:41,709 and coding does evolve anywhere 1424 00:55:41,710 --> 00:55:43,839 else than multi-channel and 1425 00:55:43,840 --> 00:55:45,969 coding instead of, uh, 1426 00:55:45,970 --> 00:55:48,189 increasing sample rate and sample 1427 00:55:48,190 --> 00:55:50,109 bits? Oh, thank you. 1428 00:55:50,110 --> 00:55:52,239 Good questions. Uh, I think 1429 00:55:52,240 --> 00:55:54,549 this, um, this 1430 00:55:54,550 --> 00:55:56,109 is a good example of how not to market 1431 00:55:56,110 --> 00:55:57,849 something because I cannot think of any 1432 00:55:57,850 --> 00:56:00,009 product. It has managed to create more 1433 00:56:00,010 --> 00:56:02,619 doubt with both the consumers 1434 00:56:02,620 --> 00:56:04,959 of of this product and also hardware 1435 00:56:04,960 --> 00:56:07,179 manufacturers and all kinds of other 1436 00:56:07,180 --> 00:56:08,439 people which are involved in digital 1437 00:56:08,440 --> 00:56:10,449 audio. Then MQM. 1438 00:56:10,450 --> 00:56:12,909 I think they went out, um, 1439 00:56:12,910 --> 00:56:14,979 with very bold claims, 1440 00:56:14,980 --> 00:56:17,079 try to play the audio files 1441 00:56:17,080 --> 00:56:21,009 with their easy, easy suggestibility 1442 00:56:21,010 --> 00:56:22,959 and were surprised by the competence in 1443 00:56:22,960 --> 00:56:24,219 the audience. 1444 00:56:24,220 --> 00:56:26,289 So I don't think this will 1445 00:56:26,290 --> 00:56:28,059 go anywhere in this terms. 1446 00:56:28,060 --> 00:56:30,249 Uh, what was your second question again 1447 00:56:30,250 --> 00:56:32,559 about how the encoding of music 1448 00:56:32,560 --> 00:56:34,629 evolves? If we are going 1449 00:56:34,630 --> 00:56:36,789 for more channels like five 1450 00:56:36,790 --> 00:56:39,759 channels, eight channels, maybe 1451 00:56:39,760 --> 00:56:42,069 sixteen channels at some point, 1452 00:56:42,070 --> 00:56:44,199 or do some people still try to increase 1453 00:56:44,200 --> 00:56:46,779 the bitrate and 1454 00:56:46,780 --> 00:56:47,709 frequency? 1455 00:56:47,710 --> 00:56:49,779 Yeah. So, um, it's 1456 00:56:49,780 --> 00:56:50,949 probably a question for Enten. 1457 00:56:50,950 --> 00:56:53,379 Um, I can only 1458 00:56:53,380 --> 00:56:54,619 as audio file answer. 1459 00:56:54,620 --> 00:56:56,469 There's, uh, in the audio market, 1460 00:56:56,470 --> 00:56:57,789 multichannel audio is dead. 1461 00:56:57,790 --> 00:56:58,959 It's stereo. 1462 00:56:58,960 --> 00:57:00,819 But I mean, for the consumers, this is 1463 00:57:00,820 --> 00:57:02,139 why as short as Bluetooth speaker, 1464 00:57:02,140 --> 00:57:04,239 everybody's happy with mono now. 1465 00:57:04,240 --> 00:57:06,969 Um, and it doesn't seem to be a problem. 1466 00:57:06,970 --> 00:57:09,069 Um, stereo files will have, 1467 00:57:09,070 --> 00:57:11,829 uh, audio files will have stereo set ups, 1468 00:57:11,830 --> 00:57:14,079 uh, and multichannel setups, create 1469 00:57:14,080 --> 00:57:15,729 problems with Groomes and, you know, 1470 00:57:15,730 --> 00:57:17,679 calibration all kinds of things. 1471 00:57:17,680 --> 00:57:19,809 Um, it 1472 00:57:19,810 --> 00:57:22,389 might be different in the AV sector. 1473 00:57:22,390 --> 00:57:23,390 Um, 1474 00:57:24,460 --> 00:57:26,499 in more general terms, I think audio is a 1475 00:57:26,500 --> 00:57:28,569 solve problem. Really, 24 bits in one 1476 00:57:28,570 --> 00:57:30,369 hundred ninety two kilohertz should 1477 00:57:30,370 --> 00:57:31,599 should be enough. 1478 00:57:31,600 --> 00:57:33,699 Um, the front here now is 1479 00:57:33,700 --> 00:57:36,069 room correction and face correction 1480 00:57:36,070 --> 00:57:37,689 with speakers. So that's where the real 1481 00:57:37,690 --> 00:57:39,189 problems are, where there's actually real 1482 00:57:39,190 --> 00:57:40,719 progress. Digital. 1483 00:57:40,720 --> 00:57:43,089 Um, um, 1484 00:57:43,090 --> 00:57:44,090 what do you call vison. 1485 00:57:45,960 --> 00:57:46,960 Uh, 1486 00:57:48,460 --> 00:57:49,389 crossover's. Yes. 1487 00:57:49,390 --> 00:57:50,559 Digital crossover's. These kinds of 1488 00:57:50,560 --> 00:57:52,110 things I think you 1489 00:57:54,190 --> 00:57:56,349 can add to parametric coding 1490 00:57:56,350 --> 00:57:58,809 is used in multichannel audio. 1491 00:57:58,810 --> 00:58:01,449 So you sent one mono signal 1492 00:58:01,450 --> 00:58:04,209 and then you can reproduce by parameters. 1493 00:58:04,210 --> 00:58:06,849 I have room 1494 00:58:06,850 --> 00:58:09,009 for for a multi-channel, 1495 00:58:09,010 --> 00:58:10,149 uh, setup. 1496 00:58:10,150 --> 00:58:11,619 Yeah. That's often used 1497 00:58:13,510 --> 00:58:15,909 since I think we have time for 1498 00:58:15,910 --> 00:58:17,319 another question. 1499 00:58:17,320 --> 00:58:18,320 Do we. 1500 00:58:18,760 --> 00:58:20,619 Yeah. OK, another question format for 1501 00:58:20,620 --> 00:58:21,699 number one. 1502 00:58:21,700 --> 00:58:23,769 I went to three conferences and 1503 00:58:23,770 --> 00:58:25,899 digital audio and computer music this 1504 00:58:25,900 --> 00:58:28,599 year. I didn't hear anybody mention this 1505 00:58:28,600 --> 00:58:29,559 inquiry. 1506 00:58:29,560 --> 00:58:31,269 I think the story about the 1507 00:58:33,100 --> 00:58:35,229 toxicity authentication 1508 00:58:35,230 --> 00:58:36,230 and what did you call it? 1509 00:58:37,640 --> 00:58:39,789 Can you step up? Because it's it's 1510 00:58:39,790 --> 00:58:41,949 fabulous. I wonder if they do they 1511 00:58:41,950 --> 00:58:44,049 make any claims they do like virtual 1512 00:58:44,050 --> 00:58:46,179 analog models of the original 1513 00:58:46,180 --> 00:58:48,519 converters or how do we 1514 00:58:48,520 --> 00:58:50,949 this story is just completely like 1515 00:58:50,950 --> 00:58:51,939 crazy now. 1516 00:58:51,940 --> 00:58:54,009 Yeah, that actually is being 1517 00:58:54,010 --> 00:58:56,079 mentioned in the marketing model 1518 00:58:56,080 --> 00:58:58,659 that they they call it fingerprint the 1519 00:58:58,660 --> 00:59:01,509 original analog digital converter 1520 00:59:01,510 --> 00:59:03,759 and compensate for for the 1521 00:59:03,760 --> 00:59:04,759 deficiencies. 1522 00:59:04,760 --> 00:59:07,089 Then aerostats convertor makes 1523 00:59:07,090 --> 00:59:09,309 and that's one of the big claims that 1524 00:59:09,310 --> 00:59:11,859 there is a device specific compensation 1525 00:59:11,860 --> 00:59:14,589 embedded in the file 1526 00:59:14,590 --> 00:59:17,049 being called upon upon replay and 1527 00:59:17,050 --> 00:59:19,149 hence a better quality 1528 00:59:19,150 --> 00:59:21,429 experience for the listener. 1529 00:59:21,430 --> 00:59:23,589 But all the evidence so 1530 00:59:23,590 --> 00:59:25,689 far shows that this is absolutely 1531 00:59:25,690 --> 00:59:27,669 not true. What we have is a pencil and 1532 00:59:27,670 --> 00:59:29,469 chording, just like you would have an MP 1533 00:59:29,470 --> 00:59:30,969 three or something else. 1534 00:59:30,970 --> 00:59:33,249 The encode, it apparently analyzes 1535 00:59:33,250 --> 00:59:35,769 the file that the song or the track, 1536 00:59:35,770 --> 00:59:38,019 um, in makes certain provisions on 1537 00:59:38,020 --> 00:59:40,279 how to fold the signal. 1538 00:59:40,280 --> 00:59:42,339 So how many bits get allocated in the 1539 00:59:42,340 --> 00:59:43,340 dynamic range? 1540 00:59:44,320 --> 00:59:46,449 Um, on a per track 1541 00:59:46,450 --> 00:59:47,439 basis. 1542 00:59:47,440 --> 00:59:49,719 So there is no authentication 1543 00:59:49,720 --> 00:59:52,119 or fingerprinting or any kind of modeling 1544 00:59:52,120 --> 00:59:54,279 of the analog part or 1545 00:59:54,280 --> 00:59:55,839 the part 1546 00:59:57,250 --> 00:59:58,250 in the chain. 1547 00:59:59,690 --> 01:00:01,239 So thank you very much for the 1548 01:00:01,240 --> 01:00:03,459 questioning and thanks again very much 1549 01:00:03,460 --> 01:00:05,090 for the talk about.