MQA is Vaporware

[mansr]

Just now, manisandher said:

Yep, I see what you mean. Still, it seems strange that the intermodulation distortion in the analogue circuitry sits exactly at 56 kHz, 66 kHz and 76 kHz.

Not strange at all. Intermodulation products appear at the sums and differences of integer multiples of the input frequencies, in this case 10 kHz and 86 kHz. 86 - 10 = 76, 86 - 2 × 10 = 66, and so on.

[The Computer Audiophile]

10 minutes ago, mansr said:

Not strange at all. Intermodulation products appear at the sums and differences of integer multiples of the input frequencies, in this case 10 kHz and 86 kHz. 86 - 10 = 76, 86 - 2 × 10 = 66, and so on.

Just as my wife and I were discussing over breakfast this morning (NOT).

[Ralf11]

8 minutes ago, The Computer Audiophile said:

Just as my wife and I were discussing over breakfast this morning (NOT).

 

 

Does this mean you are finally nerded out?

 

Or did you skip the Lutefisk for some reason?

[botrytis]

1 minute ago, Ralf11 said:

 

 

Does this mean you are finally nerded out?

 

Or did you skip the Lutefisk for some reason?

 

Please - it is walleye cheeks......

[Ishmael Slapowitz]

1 hour ago, John_Atkinson said:

 

If you say so.I always find it amusing when someone trues to "prove" that some else's preference is wrong. 🙂

 

John Atkinson

Technical Editor, Stereophile

You mean, what Michael Fremer has been doing for 30 years>?😎

[psjug]

22 hours ago, John_Atkinson said:

 

Actually, it is the fact that listeners seem to tend to prefer the leaky DAC reconstruction filters with most kinds of music. Certainly on my commercially released recordings, which people seem to like the sound of, I use the Ayre QA-9 set to "Listen," which has a slow-rolloff antialiasing filter, as the master A/D converter. The possibility for there being some low-level aliasing energy seems to be offset against the better time-domain performance. See the discussion of this trade-off at https://www.stereophile.com/reference/104law/index.html

 

Regarding the behavior of the upsampling performed by the renderer, I assume this will depend on which MQA filter is being used for the upsampling. Perhaps mansr could tell use which one is specified in his file's embedded MQA data? If a filter isn't specified in his data, then the spectral contamination will depend on which upsampling filter the renderer selects as its default.

 

John Atkinson

Technical Editor, Stereophile

 

Here we are talking about preferences in filters that cut at about 48KHz.  I really have a hard time believing  it matters whether these filters are linear phase, minimum phase, intermediate, as long as the filters are steep.  In the case of the MQA renderer, the weak filter gives the ultasonic images that could have an effect downstream.  So this would seem to be the issue, and I don't see how this could be considered a good thing.  If downstream equipment is affected, who can predict how this will present itself?

 

@Archimago discusses this in depth here: http://archimago.blogspot.com/2017/06/measurements-audioquest-dragonfly-black_24.html

 

He writes:

 I think it's more likely that all those who hear "huge differences" between MQA and non-MQA Studio Masters are hearing the effect of the distortion than some kind of temporal domain improvement. Presumably these people like what the ultrasonic imaging is doing in their systems!

 

[crenca]

12 minutes ago, psjug said:

Here we are talking about preferences in filters that cut at about 48KHz.  I really have a hard time believing  it matters whether these filters are linear phase, minimum phase, intermediate, as long as the filters are steep. 

 

Not to disagree with the rest of your comments, but there is a relatively small but discernible difference between filters (min phase vs. linear) in the iFi products (I have the iDAC2 micro & Nano), or the selections in HQPlayer for example.  Min phase "blurs" HF content, or to put a positive spin on it "relaxes" it.  The upper harmonics of brass & strings are particularly affected so I don't normally listen to min phase.   That said, I don't normally listen to tipped up chains, so as a HP listener for example I don't need to "relax" an Sennheiser HD800 because I don't own them

[mansr]

1 minute ago, crenca said:

Not to disagree with the rest of your comments, but there is a relatively small but discernible difference between filters (min phase vs. linear) in the iFi products (I have the iDAC2 micro & Nano), or the selections in HQPlayer for example.  Min phase "blurs" HF content, or to put a positive spin on it "relaxes" it.

The iFi DACs don't have a minimum phase option, unless you're using firmware 5.3C in which case they have only the MQA filters. The switch is labelled minimum phase, but it's actually a slow roll-off linear phase.

[psjug]

10 minutes ago, crenca said:

 

Not to disagree with the rest of your comments, but there is a relatively small but discernible difference between filters (min phase vs. linear) in the iFi products (I have the iDAC2 micro & Nano), or the selections in HQPlayer for example.  Min phase "blurs" HF content, or to put a positive spin on it "relaxes" it.  The upper harmonics of brass & strings are particularly affected so I don't normally listen to min phase.   That said, I don't normally listen to tipped up chains, so as a HP listener for example I don't need to "relax" an Sennheiser HD800 because I don't own them

Are you talking about redbook, though?  Possibly then the  linear phase vs minimum phase filter difference would be heard.  But a filter that cuts ~48KHz?

[crenca]

55 minutes ago, mansr said:

The iFi DACs don't have a minimum phase option, unless you're using firmware 5.3C in which case they have only the MQA filters. The switch is labelled minimum phase, but it's actually a slow roll-off linear phase.

 

Is this the case going back to early versions?  I see the last time I downloaded a driver was 7/2016...

 

edit:  Wait, that's just the previous driver (5.2)...

[crenca]

48 minutes ago, psjug said:

Are you talking about redbook, though?  Possibly then the  linear phase vs minimum phase filter difference would be heard.  But a filter that cuts ~48KHz?

 

Point taken.  Yes, I have a lot more redbook than hires locally and even when I stream Qobuz I usually choose redbook...

[mansr]

1 hour ago, crenca said:

Is this the case going back to early versions?  I see the last time I downloaded a driver was 7/2016...

 

edit:  Wait, that's just the previous driver (5.2)...

All versions up to and including 5.2 use only the built-in filters in the DSD1793 chip. It has two options, sharp or slow roll-off, both linear phase. Firmware 5.30 adds MQA rendering capability (and disables some other features) without affecting non-MQA playback. 5.3C replaces the built-in filters for non-MQA with their "GTO" filter which is actually one of the MQA variants.

[FredericV]

3 hours ago, psjug said:

He writes:

 I think it's more likely that all those who hear "huge differences" between MQA and non-MQA Studio Masters are hearing the effect of the distortion than some kind of temporal domain improvement. Presumably these people like what the ultrasonic imaging is doing in their systems!

 

This reminds me of this Parov Stelar track, which has some kind of aliasing distortion in the opening of the track. I really like this track as I like this effect as opener, but I would not apply such effect to all music The flac version on tidal gives the illusion of openness in the highs. So I can imagine why a more subtle variant of the same brings the illusion of openness.

You could also just upsample with minimum phase and one cycle of postringing to get this effect, but it just does not work well with all music. With EDM it's a nice gimmick. See the second link in my signature for the sox recipe.
 

 

Youtube's lossy audio compression makes the effect weak and replaces it by a lot of noise, so you should really hear this in FLAC or WAV or any lossless variant of the track. It's on Tidal:

https://tidal.com/browse/album/55284281

Track 10

[Al M.]

On 7/4/2019 at 7:12 PM, Ran said:

 

Is there a stock reply I can use to complain about this? Did anyone send a compliant to a label before?

 

I sent a complaint that I would never buy an MQA CD, with technical reasons mentioned and linked to,  to Decca:

 

https://shop.decca.com/help/

 

(click on "Any other questions")

 

Haven't heard back from them, but it might be useful if they would hear from a couple of people. They probably live in a bubble and don't even realize that there is criticism.

[mansr]

1 minute ago, Al M. said:

I sent a complaint that I would never buy an MQA CD, with technical reasons mentioned and linked to,  to Decca:

 

https://shop.decca.com/help/

 

(click on "Any other questions")

 

Haven't heard back from them, but it might be useful if they would hear from a couple of people. They probably live in a bubble and don't even realize that there is criticism.

They'd probably have to receive about 10 million complaints before noticing. The entertainment industry is remarkably adept at ignoring what is blatantly obvious to anyone else.

[Hugo9000]

Quote

Unnecessary Inventions solve real world problems that do not exist by creating products that no one is asking for. All of the products that we design are not for sale including some of our most well known inventions such as the Crocs Gloves, Cuisine Curtain, Pizza Fanny Pack, Avocado on a Stick, AirSticks, Infinity Saucelet, and many more.* What should we create next?!

 

*and many more?  Could they have been consultants on the MQA project?    

 

 

My apologies if this fine youtube channel (they also have a subreddit) has been linked here before.

[esldude]

On 8/8/2019 at 10:41 AM, botrytis said:

 

Imaging is a brain phenomenon, not a file issue.

 

I am not aware of how one can measure imaging in a file since your ears and brain determine that.

In case you don't know, aliasing is when ultrasonics beyond the ADC range show up below Nyquist for the sample rate in use.

As an example say you have a 48 khz sample rate, which should roll off at 24 khz.  A really strong 34 khz signal would alias and show up at 24khz -(34khz-24khz).   Or 14 khz in the recorded file if the input filter isn't steep enough.  

EDIT math mistake pointed out by mansr. 

 

Imaging is when audible frequencies are imaged above the Nyquist frequency into the ultrasonic range.  Using the example above a strong 10 khz signal in a 48 khz file would image at 24khz+14khz or 38 khz in the output if the output filtering isn't steep enough. 

 

 

 Quite often people discuss aliasing of DACs which is a misnomer.  Aliasing occurs in the sampling of the analog signal.  Imaging occurs upon playback of a DAC. 

 

 

[mansr]

3 minutes ago, esldude said:

In case you don't know, aliasing is when ultrasonics beyond the ADC range show up below Nyquist for the sample rate in use.

As an example say you have a 48 khz sample rate, which should roll off at 24 khz.  A really strong 34 khz signal would alias and show up as a signal at 34 khz-24khz or 10 khz in the recorded file if the input filter isn't steep enough.  

I think your maths may be a little off there.

[Arpiben]

On 8/8/2019 at 6:47 PM, mansr said:

There are a few things happening here. First we have the 10 kHz tone that's actually in the file. Then we have the 86 kHz image created by the MQA filter. At 20 kHz and 30 kHz we have harmonics (distortion) of the 10 kHz tone. The spikes at 66 kHz, 76 kHz, and 96 kHz are the result of intermodulation between the 10 kHz and 86 kHz frequencies. No idea what's going on at 94 kHz. Anyone care to speculate?

 

Probable FFT artefact from the distribution of quantization error along the frequency bins in coherent sampling of sine waves.

Have no idea of  window size and its number of cycles used for FFTs shown ,etc...

Anyhow surprised by its level.

 

 

 

[Al M.]

17 hours ago, mansr said:

They'd probably have to receive about 10 million complaints before noticing. The entertainment industry is remarkably adept at ignoring what is blatantly obvious to anyone else.

Perhaps. But if you don't even try, you will never change things. What is time better spent, writing to Decca or writing another post here commiserating about a technical aspect of MQA in a thread that few in the industry follow? I would suggest, once you're done with the former, you can merrily continue with the latter, to your heart's content.

[astrotoy]

Having not participated in this thread, when it hit 16000 posts, I thought it would be a good time.  I turned around, and the post count was 17000.  First, a disclaimer.  I have never heard an MQA file, nor do I think I have anything that will play one (although I do use Roon but don't subscribe to anything with it).  It appears that there is a goal to this thread for the great majority posting on it.  That is to end MQA.  However, it appears to me that this thread is an extremely ineffective way to accomplish it.  First, there appears to be no one from any of the customers of MQA (I mean the companies that pay MQA for licensing their processes) that participate in the thread or pay any obvious attention to it.  Second, it doesn't appear that MQA itself participates or pays attention to it.  Third, that for these parties hi-fi is a business and they want to make money in that business, and see MQA as a way to do that.  

 

So, to end MQA one should be thinking about ways to directly affect the business decisions that these companies are making.  The easiest way is to wait for the market to make the decision.  It requires no work, not even a thread (so you can go about with more useful activity). However, it very well could be very slow and may not be successful. Take Bose, for example (as our fearless leader did a couple of weeks ago) or Dolby (if we follow Mr. Dyson's comments about Dolby A whose lack of quality did not lead to the demise of Dolby).  

 

Another way is to put economic pressure on the companies, such as a massive boycott of their products.  Takes much more effort, but has been effective in some cases.  For the many customers of MQA it would take a large number of such boycotts, not so easy to do.  There may be a critical customer whose loss would so affect the business of MQA that they would go bankrupt. If so, going after that company might make sense.  Another is to initiate a class action lawsuit (most effective if targeted directly at MQA, since class action suits against the many different companies which license MQA would take much more effort.)  You would need to either have an attorney (perhaps a member of this thread) who would be willing to take on the suit pro bono, or have wealthy members of the thread finance the law suit.  There may not be reasonable grounds to file the suit, but that has not stopped entities from pursuing such actions, especially if they are tenacious and have the financial backing to wear the company down.

 

A third way is the way that was successful in the last US Presidential Election.  Do a version of 'Catch and Kill'.  This would be best directed at MQA, not at its customers.  Make an offer to buy them out.  If successful, then shut down the company.  Happens all the time with a competitor buying out a rival.  Quick and efficient example of capitalism at work.   I am sure MQA has a price.  If as reported by some posters on this thread that MQA is losing money, then the offer would be lower than if they were profitable.  Given their reported annual sales (taken from earlier posts on this thread) that they have annual sales in the 7 digit range (updates and corrections are appreciated, but the exact number is not important for the argument), then an offer which I would guess might be in the $1M (USD or GBP) to $100M range could get you the company (I am guessing that the real number is in the lower part of the range).   There may be members of the thread or people who have friends who may be willing to put up the necessary funds.   If there is not an individual willing to do it, then perhaps a go fund me campaign would be an easy test to see whether there is sufficient interest. 

 

Larry (who is not interested in having the last word on this or any other thread)

 

 

 

 

 

 

[gdpr]

3 hours ago, astrotoy said:

Having not participated in this thread, when it hit 16000 posts, I thought it would be a good time.  I turned around, and the post count was 17000.  First, a disclaimer.  I have never heard an MQA file, nor do I think I have anything that will play one (although I do use Roon but don't subscribe to anything with it).  It appears that there is a goal to this thread for the great majority posting on it.  That is to end MQA.  However, it appears to me that this thread is an extremely ineffective way to accomplish it.  First, there appears to be no one from any of the customers of MQA (I mean the companies that pay MQA for licensing their processes) that participate in the thread or pay any obvious attention to it.  Second, it doesn't appear that MQA itself participates or pays attention to it.  Third, that for these parties hi-fi is a business and they want to make money in that business, and see MQA as a way to do that.  

 

So, to end MQA one should be thinking about ways to directly affect the business decisions that these companies are making.  The easiest way is to wait for the market to make the decision.  It requires no work, not even a thread (so you can go about with more useful activity). However, it very well could be very slow and may not be successful. Take Bose, for example (as our fearless leader did a couple of weeks ago) or Dolby (if we follow Mr. Dyson's comments about Dolby A whose lack of quality did not lead to the demise of Dolby).  

 

Another way is to put economic pressure on the companies, such as a massive boycott of their products.  Takes much more effort, but has been effective in some cases.  For the many customers of MQA it would take a large number of such boycotts, not so easy to do.  There may be a critical customer whose loss would so affect the business of MQA that they would go bankrupt. If so, going after that company might make sense.  Another is to initiate a class action lawsuit (most effective if targeted directly at MQA, since class action suits against the many different companies which license MQA would take much more effort.)  You would need to either have an attorney (perhaps a member of this thread) who would be willing to take on the suit pro bono, or have wealthy members of the thread finance the law suit.  There may not be reasonable grounds to file the suit, but that has not stopped entities from pursuing such actions, especially if they are tenacious and have the financial backing to wear the company down.

 

A third way is the way that was successful in the last US Presidential Election.  Do a version of 'Catch and Kill'.  This would be best directed at MQA, not at its customers.  Make an offer to buy them out.  If successful, then shut down the company.  Happens all the time with a competitor buying out a rival.  Quick and efficient example of capitalism at work.   I am sure MQA has a price.  If as reported by some posters on this thread that MQA is losing money, then the offer would be lower than if they were profitable.  Given their reported annual sales (taken from earlier posts on this thread) that they have annual sales in the 7 digit range (updates and corrections are appreciated, but the exact number is not important for the argument), then an offer which I would guess might be in the $1M (USD or GBP) to $100M range could get you the company (I am guessing that the real number is in the lower part of the range).   There may be members of the thread or people who have friends who may be willing to put up the necessary funds.   If there is not an individual willing to do it, then perhaps a go fund me campaign would be an easy test to see whether there is sufficient interest. 

 

Larry (who is not interested in having the last word on this or any other thread)

 

 

 

 

 

 

Are you maybe working for MQA and looking for a way out? (just joking)

Dirk

[mansr]

4 hours ago, astrotoy said:

It appears that there is a goal to this thread for the great majority posting on it.  That is to end MQA.  However, it appears to me that this thread is an extremely ineffective way to accomplish it.

I don't think anyone is under the impression that thread in and of itself will lead to the destruction of MQA. It is, however, a useful place to exchange uncensored information about MQA. Also, they have tried to have the thread shut down, so it must having some effect.

 

4 hours ago, astrotoy said:

First, there appears to be no one from any of the customers of MQA (I mean the companies that pay MQA for licensing their processes) that participate in the thread or pay any obvious attention to it.

Their NDAs quite possibly prevent this.

 

4 hours ago, astrotoy said:

Second, it doesn't appear that MQA itself participates or pays attention to it.

They've been invited to respond to our findings. They chose table-banging instead.

 

4 hours ago, astrotoy said:

Third, that for these parties hi-fi is a business and they want to make money in that business, and see MQA as a way to do that.

They probably do. They are probably mistaken.

[John Dyson]

As someone who ONLY has a slight technical interest in the math games done by the MQA algorithm, and a dislike for anything that messes up audio quality -- here is my own impression of that scheme.   I am sometimes a bit confused -- so I'd like for both pro-MQA and anti-MQA people to correct my impression:

 

Proposed sales disclaimer for MQA 'processed' product:

Here, I am selling you a modified quality (less dangerous) digital recording than what you used to purchase, and this new, safety improved version is what you now get by default!  However, if you want to *listen* to the less protected quality closer to you have previously purchased and as a legacy & still might expect, now we help you do so with safely 'approved' adjunct technologies.   The previous dangerous, direct digital version of the recording, not having safety 'authorized' processing,  without the important special 'folding' of any kind, is no longer available to cause anyone harm.  With the more sophisticated MQA technology, you will not even have mistaken access to the mostly still dangerous, mostly corrected and carefully unfolded digital signal.   Additionally, you will now benefit from ONLY indirectly (no direct payments needed!!! yay!!!)  paying the MQA license holders for the mechnanisms that improve  the lesser quality digital signal  to a 'better' version, which is in the safety listen-only 'lockbox'.  That is, these sophisticated algorithms will return the recordings (for listening, safely in a lockbox) to a state is just about as good as the more dangerous, original digital signal.  When further, dangerous, unlicensed consumer signal processing is needed, you'll be protected by utilizing the undecoded, still safely folded material by default.

This is a breakthrough for the audiophile consumer, don't you think?

 

If my understanding is close to correct, the effect on the consumer is that it seems to give the 'very beneficial' 🙂 opportunity to pay more indirect license fees to the MQA license holders.  We need MORE research like MQA!!! 🙂

 

 

 

[4est]

On 8/5/2019 at 7:19 PM, sandyk said:

 Hi Stephen

 Are you aware of the recording(s) that he will be using ?

 It doesn't help to get accurate results when you are unfamiliar with the type of music used, which may be very different to what you normally listen to. Probably a bit like asking you to choose the hi res version of an original Abba recording before John Dyson's Dolby A decoder cleaned it up.

 Regards

Alex

Please don't ask me to listen to Abba, pretty please?