R1200CL Posted December 15, 2020 Share Posted December 15, 2020 1 hour ago, Mayfair said: Exactly. And with all the confusion in labeling, I think I am going to hold off buying downloads until I'm sure I know that when I am considering purchasing music that is not labeled as MQA, I can rely that I'm not being sold stealth-MQA. Yea, we need Provenance😂 https://www.discogs.com/forum/thread/732807 Dark side of the moon has presently 1021 versions. IT Freak 1 Link to comment
R1200CL Posted December 15, 2020 Share Posted December 15, 2020 @UkPhil How can one get in touch with Bob ? It seems that’s from a forum Link to comment
Popular Post FredericV Posted December 15, 2020 Popular Post Share Posted December 15, 2020 4 hours ago, R1200CL said: Does anyone understand exactly what’s been explained here. I’m totally confused. (I can’t understand it’s technically possible to have a 24 bit rate inside a 16 bit pcm transmission stream). http://bobtalks.co.uk/blog/mqaplayback/mqa-cd-examples/ When the transmission is restricted to 16 bits, as for example in an MQA CD, the decoder will maintain the output resolution at 24b, firstly to preserve the encoder’s enhanced quantisation, but also because there is more information in a decoded MQA CD than can be represented in a normal rectangular 44.4 kHz/16b PCM channel. I noticed my decoder is outputting a 24 bit file from a 16 bit MQA encoded file (from one of those albums which Tidal replaced with 16/44.1 MQA). But that does not make it a true 24 bit file, as those batch encoded files started out as 16/44.1, now batch processed to add the MQA sauce to it ... BTW: several mp3 decoders are also outputting more than 16 bits - in the early days of mp3 such decoders would mask the mp3's shortcomings a little bit better (but not completely) - also upsampling mp3 combined with a 24 bit decoder was a trick to make it sound better as the lossy errors would be distributed over more samples ... it looks like MQA uses the same little trick. Not just increasing bit depth, but upsampling can also be used to obfuscate "bad" DSP. 2 techniques which MQA also uses. In the early days of our own research, we found that foobar's soundtouch plugin had audible errors which were very obvious with certain types of music. So to reduce these errors, we just upsampled to x4 and then run the soundtouch plugin, and to get files back in a format which the car's premium sound system required, downsample back to the original redbook rate ... That idea never went into a product (we don't use foobar), but having foobar around to batch process flac files into wav's for the car saved on a USB stick, was useful, and it would get rid of all audible artefacts. This was long before MQA ... It looks like MQA is also using similar trickery as MP3 decoders (including dithering). Don't explain this in the Veter Peth group, as clamining MQA is like MP3 will get you banned - but now it is more clear why MQA uses 24 bit output for redbook sourced files, and upsampling - to distribute the error across more samples, to obfuscate their lossy codec. http://mp3decoders.mp3-tech.org/24bit.html You can actually replace mp3 by mqa and the same would apply: My mp3s are made from 16-bit CDs - why use a 24-bit decoder? As you probably know, when you encode a CD to mp3 format, you don't store an exact copy of the original signal. When an mp3 is decoded, you don't get those original 16-bits back, but an approximation that should sound similar. When the decoder puts together all the elements held in the mp3 file, the arithmetical result can be very accurate in numerical terms, even if it's not exactly what was on the original CD. If you round it to 16-bits, you add a small amount of extra distortion to this reconstructed signal, getting even further away from what was on the original CD. If you round it to 24-bits, you're still adding distortion, but it's 256 times quieter than that added by rounding to 16-bits. MikeyFresh, Kyhl and lucretius 3 Designer of the 432 EVO music server and Linux specialist Discoverer of the independent open source sox based mqa playback method with optional one cycle postringing. Link to comment
lucretius Posted December 15, 2020 Share Posted December 15, 2020 27 minutes ago, FredericV said: I noticed my decoder is outputting a 24 bit file from a 16 bit MQA encoded file (from one of those albums which Tidal replaced with 16/44.1 MQA). But that does not make it a true 24 bit file, as those batch encoded files started out as 16/44.1, now batch processed to add the MQA sauce to it ... If that was a 16 bit MQA file put into a 24 bit flac, then that should leave the bottom 8 bits all as zero. Right? IT Freak 1 mQa is dead! Link to comment
Popular Post FredericV Posted December 15, 2020 Popular Post Share Posted December 15, 2020 1 minute ago, lucretius said: If that was a 16 bit MQA file put into a 24 bit flac, then that should leave the bottom 8 bits all as zero. Right? This is not the MQA distribution file, but the decoded version of the file. It confirms that BS is using the same trickery as 24 bit MP3 decoders, which need to output more than 16 bits from a 16 bit source, as the lossy process introduces small errors. 5 hours ago, R1200CL said: When the transmission is restricted to 16 bits, as for example in an MQA CD, the decoder will maintain the output resolution at 24b, firstly to preserve the encoder’s enhanced quantisation, but also because there is more information in a decoded MQA CD than can be represented in a normal rectangular 44.4 kHz/16b PCM channel. Let's deblur this a little bit: As you probably know, when we encode a CD to MQA format, we don't store an exact copy of the original signal. When an MQA is decoded, you don't get those original 16-bits back, but an approximation that should sound similar. When the decoder puts together all the elements held in the MQA file, the arithmetical result can be very accurate in numerical terms, even if it's not exactly what was on the original CD. If we round it to 16-bits, we add a small amount of extra distortion to this reconstructed signal, getting even further away from what was on the original CD. If we round it to 24-bits, we're still adding distortion, but it's 256 times quieter than that added by rounding to 16-bits. We call this enhanced quantisation ;) Kyhl and lucretius 1 1 Designer of the 432 EVO music server and Linux specialist Discoverer of the independent open source sox based mqa playback method with optional one cycle postringing. Link to comment
GUTB Posted December 15, 2020 Share Posted December 15, 2020 Okay so comparing the Phil Collins album, just switching between Master and HiFi I was unable to reliably detect any difference. The Deep Purple live, Made in Japan - again, switching between Master and HiFi, the MQA version appears to have a touch more definition, but really just a touch more -- it's like the difference between a CD an a 96 kHz file. I'll listen to the others. Maybe I'll break out the HD600 for this. Currently using the TH900 (Lawton modded for balanced cable) which is a very resolving unit but its V shaped sound signature might be hiding differences. Link to comment
R1200CL Posted December 15, 2020 Share Posted December 15, 2020 1 hour ago, FredericV said: I noticed my decoder is outputting a 24 bit file from a 16 bit MQA encoded file (from one of those albums which Tidal replaced with 16/44.1 MQA). You have the same issue as me. Even Roon support won’t comment on this. Why do they avoid the issue ? What is even more surprising is Lumin firmware guy (he is here to) clame decoding is correct. It can’t be. It’s an error. Actually an display error. Which Bob says doesn’t matter if I understand his foggy explanations. There is no doubt. Transport rate is 16/44.1. Original encoded file is 16 bit. Have never been upsampled. It’s a MQA Authentication at 44.1. Roon has done a good job to show us what we get. Under the cover they always tells us what the transport rate and container is. After they display the original encoded file, which is verified by the authentication flag in SW. (Not so with tidal). But the Core Decoder has an error. I assume we must blame MQA Ltd. Not Roon. I’m quite sure the real file of cause is 16/44.1 (minus MQA bits or whatever has been done by the decoding machine. The error also applies to all 48kHz/24 that display MQA 48 kHz. Link to comment
Popular Post The Computer Audiophile Posted December 15, 2020 Popular Post Share Posted December 15, 2020 Ah, MQA. Clear as mud. This isn’t rocket science. Simple explanations could be given for everything. However, that would be in the consumer’s best interest, which collides with MQA Ltd’s and the label’s best interests. UkPhil, Teresa, MikeyFresh and 1 other 2 1 1 Founder of Audiophile Style | My Audio Systems Link to comment
Popular Post GUTB Posted December 15, 2020 Popular Post Share Posted December 15, 2020 I don't think it's an error, I believe MQA will always unfold to 88.2 kHz regardless of the original sample rate. This isn't a simple compression mechanism, it's trying to rebuild a picture of the original's timing cues along with the original information up to 50 kHz (hence the 88.2 kHz rate) and this unfold is likely generated the same across the board with the final unfold applying the remaining timing corrections. This is what I gathered from the various interviews. KeenObserver, lucretius, kumakuma and 3 others 6 Link to comment
MikeyFresh Posted December 15, 2020 Share Posted December 15, 2020 21 minutes ago, GUTB said: This isn't a simple compression mechanism, it's trying to rebuild a picture of the original's timing cues In other words, "there's more to MQA than you think" or something to that effect. Another parroting of Bob-speak. 21 minutes ago, GUTB said: the final unfold applying the remaining timing corrections. Would that be a second "unfold" (read: upsampling with lousy filters), or are you referring to some mythical third unfold, in which MQA certified hardware can literally birth a new world, causing a veritable revolution in audio. 21 minutes ago, GUTB said: This is what I gathered from the various interviews. Who was interviewed, Bob Stuart? We're back to pure marketing-speak, again. A broken record. What, did they toss you a free Dragonfly Cobalt or something? lucretius 1 Boycott HDtracks Boycott Lenbrook Boycott Warner Music Group Link to comment
wklie Posted December 15, 2020 Share Posted December 15, 2020 2 hours ago, R1200CL said: You have the same issue as me. Even Roon support won’t comment on this. Why do they avoid the issue ? Officially, Roon support did answer you here: https://community.roonlabs.com/t/roon-core-decode-16-bit-mqa-to-what/131491/9?u=wklie "The MQA Core Decoder always outputs either 24bit/88.2kHz or 24bit/96kHz, regardless of the original bit-rate or sample rate of the content. -John" I disagree this to be a case of "Roon support won’t comment on this". lucretius 1 Peter Lie LUMIN Firmware Lead Link to comment
Popular Post kumakuma Posted December 15, 2020 Popular Post Share Posted December 15, 2020 1 hour ago, GUTB said: I don't think it's an error, I believe MQA will always unfold to 88.2 kHz regardless of the original sample rate. This isn't a simple compression mechanism, it's trying to rebuild a picture of the original's timing cues along with the original information up to 50 kHz (hence the 88.2 kHz rate) and this unfold is likely generated the same across the board with the final unfold applying the remaining timing corrections. This is what I gathered from the various interviews. firedog, MikeyFresh, lucretius and 1 other 4 Sometimes it's like someone took a knife, baby Edgy and dull and cut a six inch valley Through the middle of my skull Link to comment
GUTB Posted December 15, 2020 Share Posted December 15, 2020 1 hour ago, kumakuma said: I appreciate the reaction image, but if you have a better understanding of the process please share. I understand that MQA Ltd. doesn't want to tell the public how exactly it works because it would be stolen. So we have no choice to go by what Stuart says in interviews and other industry people who have an inside track, for example Roon. It's been repeated several times that time domain correction was the driving motivation of MQA. The business aspect of it, providing a compression technique to make hi-res content delivery more economical is there to bring the streaming services onboard. The authentication aspect is for labels. These are things which MQA, a for-profit company, will try to market to consumers. How they chose to market it isn't relevant to me, I only care about the sonic benefits. Don't you care about the sonic benefits? HumanMedia 1 Link to comment
Popular Post IT Freak Posted December 15, 2020 Popular Post Share Posted December 15, 2020 I fully agree : the 16b / 44.1k Warner MQA flacs suck. The original redbook flacs were the best ! lucretius and MikeyFresh 2 Link to comment
UkPhil Posted December 15, 2020 Share Posted December 15, 2020 7 hours ago, R1200CL said: @UkPhil How can one get in touch with Bob ? It seems that’s from a forum Bob Stuart is only contactable via MQA Ltd or sometimes on the MQA Facebook site where he will reply to users comments Link to comment
UkPhil Posted December 15, 2020 Share Posted December 15, 2020 6 hours ago, lucretius said: If that was a 16 bit MQA file put into a 24 bit flac, then that should leave the bottom 8 bits all as zero. Right? I believe that's what Bob Stuart said when it was highlighted on the MQA Facebook site, it was put down to an error in the batch processing of these Red Book PCM files as some were dumped into a 24 bit container so I assume just padded with zeros with no difference in its "sonic" output IT Freak 1 Link to comment
UkPhil Posted December 15, 2020 Share Posted December 15, 2020 5 hours ago, GUTB said: I don't think it's an error, I believe MQA will always unfold to 88.2 kHz regardless of the original sample rate. This isn't a simple compression mechanism, it's trying to rebuild a picture of the original's timing cues along with the original information up to 50 kHz (hence the 88.2 kHz rate) and this unfold is likely generated the same across the board with the final unfold applying the remaining timing corrections. This is what I gathered from the various interviews. I am not sure, this is true their are a lot of these conversions giving this info, Not sure in 1979 The Scorpions were ahead of the digital game this would have been an analogue recording with about 11/12 bits of DR why output it to 88.2 ??.....such a mess In fact it's practically up sampling for no reason Link to comment
DuckToller Posted December 15, 2020 Share Posted December 15, 2020 3 hours ago, GUTB said: How they chose to market it isn't relevant to me, I only care about the sonic benefits. Don't you care about the sonic benefits? That may work on a individual scale with selected "white gloved" recordings due to expectation bias and perhaps available "craftmanship", who knows? As a mass processing with hundred thousends of tracks from different provenance with different masterings - it's a recipe for a sonic benefits desaster, imho. And eventually your preferences aren't too relevant for the majority and any scientific approach on the subject ? MikeyFresh 1 Link to comment
Popular Post mevdinc Posted December 15, 2020 Popular Post Share Posted December 15, 2020 7 hours ago, GUTB said: Okay so comparing the Phil Collins album, just switching between Master and HiFi I was unable to reliably detect any difference. The Deep Purple live, Made in Japan - again, switching between Master and HiFi, the MQA version appears to have a touch more definition, but really just a touch more -- it's like the difference between a CD an a 96 kHz file. I'll listen to the others. Maybe I'll break out the HD600 for this. Currently using the TH900 (Lawton modded for balanced cable) which is a very resolving unit but its V shaped sound signature might be hiding differences. In my comparisons I always detected that MQA sounded louder (around 3db or so), which generally makes the MQA versions sound as if they are more dynamic with more detailed highs. I suggest you level match both versions when comparing. This is the very reason why I never switched to listening to MQA versions of my favourite albums. To me, they sounded a little too harsh and forward. I guess some people call it more detailed sound. I have recently started using Qobuz and hi-res music is really excellent. HumanMedia and MikeyFresh 1 1 mevdinc.com (My autobiography) Recently sold my ATC EL 150 Actives! Link to comment
KeenObserver Posted December 15, 2020 Share Posted December 15, 2020 6 hours ago, GUTB said: I appreciate the reaction image, but if you have a better understanding of the process please share. I understand that MQA Ltd. doesn't want to tell the public how exactly it works because it would be stolen. So we have no choice to go by what Stuart says in interviews and other industry people who have an inside track, for example Roon. It's been repeated several times that time domain correction was the driving motivation of MQA. The business aspect of it, providing a compression technique to make hi-res content delivery more economical is there to bring the streaming services onboard. The authentication aspect is for labels. These are things which MQA, a for-profit company, will try to market to consumers. How they chose to market it isn't relevant to me, I only care about the sonic benefits. Don't you care about the sonic benefits? What MQA is giving you is "TOT". MQA is supplying you with contaminated brandy. It's a metaphor. Look it up. Boycott Warner Boycott Tidal Boycott Roon Boycott Lenbrook Link to comment
DuckToller Posted December 15, 2020 Share Posted December 15, 2020 18 hours ago, Norton said: looks to me that the CD layer on the hybrid SACD is MQA. See the footer on penultimate page of booklet. If I’m right, it may be confusing that they also describe this layer as “Standard CD Stereo” Attached the response from Eudora Records, for everyone who wants to know the truth: "thank you for your email and your interest. Yes, you can buy it at our website, you have the "BUY SACD" button next to the cover:) It is a hybrid SACD disc, one layer is the SACD/DSD version, and the other layer the MQA-CD. Let me know if you still have doubts!" Case closed ? Link to comment
R1200CL Posted December 15, 2020 Share Posted December 15, 2020 9 hours ago, wklie said: Officially, Roon support did answer you here: https://community.roonlabs.com/t/roon-core-decode-16-bit-mqa-to-what/131491/9?u=wklie "The MQA Core Decoder always outputs either 24bit/88.2kHz or 24bit/96kHz, regardless of the original bit-rate or sample rate of the content. -John" I disagree this to be a case of "Roon support won’t comment on this". Point is, nobody says if this is an display error, or if the file actually is 16/44.1 kHz that is mysterious converted to 24 bet depth as well as a double sample rate. I’m sure you know the truth, as you must know what your DAC receives. So it would be very helpful if you tell us. Link to comment
Dr Tone Posted December 15, 2020 Share Posted December 15, 2020 28 minutes ago, R1200CL said: Point is, nobody says if this is an display error, or if the file actually is 16/44.1 kHz that is mysterious converted to 24 bet depth as well as a double sample rate. I’m sure you know the truth, as you must know what your DAC receives. So it would be very helpful if you tell us. They have told you multiple times it is converted to 24bit 88.2. It's the way MQA's core decoder works and they have no choice to do as MQA wants. Just another place where MQA is upsampling. MikeyFresh 1 Roon Rock->Auralic Aria G2->Schiit Yggdrasil A2->McIntosh C47->McIntosh MC301 Monos->Wilson Audio Sabrinas Link to comment
R1200CL Posted December 15, 2020 Share Posted December 15, 2020 10 hours ago, GUTB said: I don't think it's an error, I believe MQA will always unfold to 88.2 kHz regardless of the original sample rate. Will it that’s true, MQA have some explanations to do. Cause there isn’t anything to unfold. Then some sort of upsampling must happen, which is also hard to believe. Link to comment
Dr Tone Posted December 15, 2020 Share Posted December 15, 2020 7 minutes ago, R1200CL said: Will it that’s true, MQA have some explanations to do. Cause there isn’t anything to unfold. Then some sort of upsampling must happen, which is also hard to believe. How is it hard to believe? Everything after the first unfold has always been upsampling to get back to the source sample rate. MikeyFresh 1 Roon Rock->Auralic Aria G2->Schiit Yggdrasil A2->McIntosh C47->McIntosh MC301 Monos->Wilson Audio Sabrinas Link to comment
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now