My Questions and Answers (MQA): An Interview with Andreas Koch

[mcgillroy]

10 hours ago, audiventory said:

 

Does you mean MQA here as full recording-playback system or file format only?

 

There is no MQA recording system. None of the standard studio-software made by Avid, Apple, Steinberg or Abelton supports MQA and it's doubtful they ever will.

 

The only professional tool available is a Mytek ADC/DAC which is intended for mastering studios.

 

Mastering is not recording, its adjusting the levels and compression of an already recorded piece of music. Case in point is that not even the mastering is done on the MQA file itself but levels are adjusted on a PCM-source which then is rendered to a lossy MQA target.

 

Music professionals are generally very skeptical about MQA - spare me the Bob Ludwig, George Massenburg talk - Ludwig is well know to support whatever new shiny toy is being touted as the next holy grail. Massenburg is one of the old boys who knows where big money is.

 

So in essence MQA is not a recording, not even a mastering format, it's a lossy post-mastering format targeted at consumers. The original PCM has better fidelity and actually represents what the mastering engineer intended.

 

Your point is?!

[jhwalker]

11 minutes ago, mcgillroy said:

The original PCM has better fidelity and actually represents what the mastering engineer intended.

 

Your point is?!

 

The point is some engineers and producers have now been quoted as saying the MQA version is *better* than the original PCM in that it sounds more accurate to what was actually heard in the studio.

 

I can only go by what I've read - but if current recording / mastering practice is subject to improvement (certainly it must be), why can't a new process (even if technically "lossy") be that improvement?

[Jud]

24 minutes ago, mcgillroy said:

Music professionals are generally very skeptical about MQA - spare me the Bob Ludwig, George Massenburg talk - Ludwig is well know to support whatever new shiny toy is being touted as the next holy grail. Massenburg is one of the old boys who knows where big money is.

 

So in essence MQA is not a recording, not even a mastering format, it's a lossy post-mastering format targeted at consumers. The original PCM has better fidelity and actually represents what the mastering engineer intended.

 

Your point is?!

 

Look, I prefer the original hi-res to the MQA versions myself.  But it seems to me too many people confuse being persuasive for continual brow-beating of everyone else with their personal take on things. We know your views well at this point.  Please permit others to have theirs, and let's just have a good conversation rather than an argument that someone has to win.

[mcgillroy]

39 minutes ago, jhwalker said:

 

The point is some engineers and producers have now been quoted as saying the MQA version is *better* than the original PCM in that it sounds more accurate to what was actually heard in the studio.

 

Source?

 

Beyond MQA marketing contexts if have not seen any such claims. AFAIK no independent, level matched comparisons of MQA vs PCM have been conducted, neither in consumer nor in professional contexts.

 

Again - nobody is arguing about a sound difference perhaps being present. The question is if the licensing & DRM baggage MQA comes with needs to be part of the equation.

 

[astrotoy]

Does this mean that, since this is a post processing system, that a consumer can buy the rights to use MQA to compress their own already ripped 192/24 files?  I would love to be able to have an 8 or 10 TB hard drive with 30 to 40 TB of 192/24 files compressed by MQA to take with me when I travel. With a laptop and a $300 Meridian Explorer 2, I could have a collection equivalent to 10,000 record albums on one external HD (the number of albums I have ripped at 192/24) which can be unfolded to play using my headphones or inserting into our daughter's system when I visit the grandkids.  Larry

[labjr]

On 3/17/2017 at 4:59 PM, Ralf11 said:

MQA is a lossy format which uses "deblurring" as a gimmick to say it's better. However, it's actually worse in some ways. It's a brilliantly marketed scheme. An attempt to corner the industry. DRM is the only thing the record industry sees. They didn't suddenly start caring about sound quality.   

 

[The Computer Audiophile]

2 hours ago, astrotoy said:

Does this mean that, since this is a post processing system, that a consumer can buy the rights to use MQA to compress their own already ripped 192/24 files?  I would love to be able to have an 8 or 10 TB hard drive with 30 to 40 TB of 192/24 files compressed by MQA to take with me when I travel. With a laptop and a $300 Meridian Explorer 2, I could have a collection equivalent to 10,000 record albums on one external HD (the number of albums I have ripped at 192/24) which can be unfolded to play using my headphones or inserting into our daughter's system when I visit the grandkids.  Larry

Presumably yes, some day. I don't think we are even close to this yet. 

[audiventory]

5 hours ago, mcgillroy said:

 

There is no MQA recording system. None of the standard studio-software made by Avid, Apple, Steinberg or Abelton supports MQA and it's doubtful they ever will.

 

The only professional tool available is a Mytek ADC/DAC which is intended for mastering studios.

 

Mastering is not recording, its adjusting the levels and compression of an already recorded piece of music. Case in point is that not even the mastering is done on the MQA file itself but levels are adjusted on a PCM-source which then is rendered to a lossy MQA target.

 

Music professionals are generally very skeptical about MQA - spare me the Bob Ludwig, George Massenburg talk - Ludwig is well know to support whatever new shiny toy is being touted as the next holy grail. Massenburg is one of the old boys who knows where big money is.

 

So in essence MQA is not a recording, not even a mastering format, it's a lossy post-mastering format targeted at consumers. The original PCM has better fidelity and actually represents what the mastering engineer intended.

 

Your point is?!

 

I consider this patent as concept of full system

https://patentscope.wipo.int/search/en/detail.jsf?docId=WO2014108677&recNum=1&maxRec=&office=&prevFilter=&sortOption=&queryString=&tab=PCTDescription

 

Also check this ADC (not DAC!) https://mytekdigital.com/brooklyn-adc/#specifications

 

Though I don't know what is "MQA Kernel" exactly.

 

As studio internal format, I consider and recommend float point only.

[mcgillroy]

2 hours ago, audiventory said:

 

I consider this patent as concept of full system

https://patentscope.wipo.int/search/en/detail.jsf?docId=WO2014108677&recNum=1&maxRec=&office=&prevFilter=&sortOption=&queryString=&tab=PCTDescription

 

Also check this ADC (not DAC!) https://mytekdigital.com/brooklyn-adc/#specifications

 

Though I don't know what is "MQA Kernel" exactly.

 

As studio internal format, I consider and recommend float point only.

 

Please reread my post. I specifically mentioned the Mytek ADC.

 

For a good discussion of the MQA patents please consult this analysis:

 

https://www.xivero.com/blog/hypothesis-paper-to-support-a-deeper-technical-analysis-of-mqa-by-mqa-limited/

 

In a recording context you want to avoid a lossy format like MQA as any further processing (EQs, effects, merging tracks etc) in the mixing-phase will exacerbate rounding errors, noise and clipping introduced by the lossy codecs.

 

That's why float representations you mention are used internally in all major audio software packages.

 

On top of that introducing MQA into recording gear and software would necessitate that all vendors become license takers. Which is what MQA wants and what some people - including me - consider as a rent-seeking land grab.

 

Companies like Avid or Apple - who happen to produce the biggest and most important audio-software packages - will laugh MQA out of the room. Both because of their outrageous claims with regards to the audio-signal-processing involved and with regards to the licensing conditions.

 

Apple paying MQA a fee for every iPhone, every iPad, every pair of Airpods, every Logic & Garageband package they sell?!

 

Dream on.

 

Just wait until Apple introduces it's high-rez-streaming offering and accompanying codec including their own DRM.

 

Then things will get really interesting.

 

[Jud]

1 hour ago, mcgillroy said:

 

Please reread my post. I specifically mentioned the Mytek ADC.

 

For a good discussion of the MQA patents please consult this analysis:

 

https://www.xivero.com/blog/hypothesis-paper-to-support-a-deeper-technical-analysis-of-mqa-by-mqa-limited/

 

In a recording context you want to avoid a lossy format like MQA as any further processing (EQs, effects, merging tracks etc) in the mixing-phase will exacerbate rounding errors, noise and clipping introduced by the lossy codecs.

 

That's why float representations you mention are used internally in all major audio software packages.

 

On top of that introducing MQA into recording gear and software would necessitate that all vendors become license takers. Which is what MQA wants and what some people - including me - consider as a rent-seeking land grab.

 

Companies like Avid or Apple - who happen to produce the biggest and most important audio-software packages - will laugh MQA out of the room. Both because of their outrageous claims with regards to the audio-signal-processing involved and with regards to the licensing conditions.

 

Apple paying MQA a fee for every iPhone, every iPad, every pair of Airpods, every Logic & Garageband package they sell?!

 

Dream on.

 

Just wait until Apple introduces it's high-rez-streaming offering and accompanying codec including their own DRM.

 

Then things will get really interesting.

 

 

There are pretty extensive discussions of these points in various MQA threads on the site.  You may want to search for some of mansr's comments in particular, which provide good technical analysis.

[Jud]

4 hours ago, audiventory said:

Though I don't know what is "MQA Kernel" exactly.

 

Their triangular filter kernel, I am assuming.

[mcgillroy]

2 hours ago, Jud said:

 

There are pretty extensive discussions of these points in various MQA threads on the site.  You may want to search for some of mansr's comments in particular, which provide good technical analysis.

 

Yep Thx - been following Mansr's analysis. Great stuff and it's a sign of the times that you have to come to a forum like this to actually find such information instead of the audio "press" doing their job.

 

One really begins to wonder what keeps a well resourced mag like Stereophile to get some independent quotes on the technical merits and legal/economic implications of MQA.

 

Kudos to Computer Audiophile for providing a place for such discussions. I am aware of the difficulties involved but can only stress the importance of making this possible.

[Jud]

13 minutes ago, mcgillroy said:

 

Yep Thx - been following Mansr's analysis. Great stuff and it's a sign of the times that you have to come to a forum like this to actually find such information instead of the audio "press" doing their job.

 

One really begins to wonder what keeps a well resourced mag like Stereophile to get some independent quotes on the technical merits and legal/economic implications of MQA.

 

Kudos to Computer Audiophile for providing a place for such discussions. I am aware of the difficulties involved but can only stress the importance of making this possible.

 

One ironic aspect of this is that mansr used Stereophile's measurements/graphs for (IIRC) the Explorer DAC in at least one or two of his comments.  In the magazine, the DAC's measurements were, if I remember right, described as "superb," whereas mansr's evaluation was not so, shall we way, ebullient.

 

I haven't read AS or SPhile for a couple of decades now, except for an occasional glance when at a bookstore.  (It was one of those occasions when I read an article by Mike Fremer of all people that got me into computer audio.)  I occasionally look at Audiostream.  I think the notion of an objective/subjective divide, and of these as somehow being in a death battle with each other rather than informing each other at best or taking a "live and let live" attitude at worst, has really worked out to the disadvantage of both the publications pushing a hard subjectivist stance and those pushing the opposite.

[mansr]

1 minute ago, Jud said:

One ironic aspect of this is that mansr used Stereophile's measurements/graphs for (IIRC) the Explorer DAC in at least one or two of his comments.  In the magazine, the DAC's measurements were, if I remember right, described as "superb," whereas mansr's evaluation was not so, shall we way, ebullient.

 

That was Miska, although I agree with his assessment.

[Jud]

Just now, mansr said:

 

That was Miska, although I agree with his assessment.

 

Thanks for the correction.

[crenca]

59 minutes ago, mcgillroy said:

 

"Yep Thx - been following Mansr's analysis. Great stuff and it's a sign of the times that you have to come to a forum like this to actually find such information instead of the audio "press" doing their job.

 

One really begins to wonder what keeps a well resourced mag like Stereophile to get some independent quotes on the technical merits and legal/economic implications of MQA."

 

I have wondered out loud about this repeatedly.  So far, no answer has really satisfied.  I *think* it is because they are not actually as "well resourced" as they appear and because they really have no real depth (personal wise) when it comes to all things digital and legal and PCM/DSD/MQA is at the end of the day just software...

Quote

 

 

 

[Jud]

7 minutes ago, crenca said:

 

I have wondered out loud about this repeatedly.  So far, no answer has really satisfied.  I *think* it is because they are not actually as "well resourced" as they appear and because they really have no real depth (personal wise) when it comes to all things digital and legal and PCM/DSD/MQA is at the end of the day just software...

 

 

As noted, Stereophile's measurements were good enough for Miska to use here in the MQA discussion.  So it isn't a lack of ability to research technical merits, it's the use they choose to make of their findings.  I think they believe (probably correctly) that their audience is primarily folks who think of themselves as "subjectivists," and therefore, IMO to the detriment of their editorial content and audience, they choose to de-emphasize technical measurements and technical explanations in favor of subjective reviews that are so uniformly favorable as to be useless to me.

[audiventory]

4 hours ago, mcgillroy said:

 

Please reread my post. I specifically mentioned the Mytek ADC.

 

For a good discussion of the MQA patents please consult this analysis:

 

https://www.xivero.com/blog/hypothesis-paper-to-support-a-deeper-technical-analysis-of-mqa-by-mqa-limited/

 

In a recording context you want to avoid a lossy format like MQA as any further processing (EQs, effects, merging tracks etc) in the mixing-phase will exacerbate rounding errors, noise and clipping introduced by the lossy codecs.

 

That's why float representations you mention are used internally in all major audio software packages.

 

1) I re-read your post. Why you mentioned Mytek ADC?

Sorry, I can't understand how MQA in ADC bound with recording. Could you explain techniclal details?

 

2) For learning patents in audio, I prefer learn patents, not links to some articles about the patents ;-)

 

Also by the link (that was discussed some time ago) no information that I'm interesting currently:

1. What is term "MQA Kernel"?

2. How it bound with ADC?

 

[audiventory]

4 hours ago, Jud said:

 

Their triangular filter kernel, I am assuming.

 

Hi Jud.

 

I suppose, you are right. But, I suspect, in MQA ADC may be some pre-correction/filter, that must cancelled/properly restored in compatible DAC, as example.

[Jud]

14 minutes ago, audiventory said:

1. What is term "MQA Kernel"?

2. How it bound with ADC?

 

Yuri, does this help at all (from an MQA patent)?:

Quote

 

We consider that this question may be relevant to the reproduction of audio, in that many natural environmental sounds such as twigs snapping and it is by no means clear that a Fourier representation is appropriate for this type of signal. The linear B-spline kernel shown in figure 1 1 is the simplest polynomial kernel that will enable unambiguous reconstruction of the location and amplitude of a Dirac pulse. We have given the name "infra-red coding" to a downsampling specification based these ideas.

 

In downsampling, we start with a signal that is already sampled but the conceptual model is that this is a continuous time signal, in which the original samples are presented a sequence of Dirac pulses. The continuous time signal is convolved with a kernel and resampled at the rate of the downsampled signal. Referring to figure 1 1 , the resampling instants are the integers 0, 1 , 2, 3 etc while the original signal is presented, on a finer grid. Assuming that the original samples and resampling instants are aligned, then the continuous time convolution with the linear B-spline followed by resampling is equivalent to a discrete-time convolution with the following sequences prior to decimation:

(1 , 2, 1 ) / 4 for decimation by 2

(1 , 2, 3, 2, 1 ) / 9 for decimation by 3

(1 , 2, 3, 4, 3, 2, 1 ) / 16 for decimation by 4

(1 , 2, 3, 4, 5, 6, 7, 8, 7, 6, 5, 4, 3, 2, 1) / 64 for decimation by 8.

 

These sequences are merely samplings at the original sampling rate of the B- spline kernel. Since the kernel has a temporal extent of two sample periods at the downsampled rate, in all cases the downsampling filter will have a temporal extent not exceeding two sample periods at the downsampled rate.

Thus for decimation by 2 the downsampling filter would have z-transform (¼ +½ z~1+1/4 z~2). We have found that very satisfactory results can be obtained using this filter for downsampling in combination with the same filter, suitably scaled in amplitude, for upsampling, with also a suitable flattener, which can be placed after upsampling, or merged with the upsampler. For downsampling from 176.4kHz to 88.2kHz the combined downsampling and upsampling droop of 2.25dB @ 20kHz can be reduced to 0.12dB using a short flattener such as:

 

2.1451346747- 1.436491673 I z 1 + 0.2913569984z"2 at 176.4kHz.

 

The total upsampling and downsampling response is then FIR with just 7 taps, hence a total temporal extent of six sample periods at the 176.4 sample rate or three sample periods at the downsampled rate. This is the shortest total filter response known to us that is often audibly satisfactory and maintains a flat response over 0-20kHz.

 

 

[audiventory]

1 minute ago, Jud said:

 

Yuri, does this help at all (from an MQA patent)?:

 

 

Thank you for information, Jud.

 

At Mytek's page mentioned "MQA Kernel" trademark.

 

In the patent we see mention of kernel too.

 

This patent relate to digital compression related, if I understand correctly.

 

Thus, I may suppose, if ADC contains "MQA Kernel", it provide compressed and, probably, pre-distorted/filtered signal.

[Jud]

2 minutes ago, audiventory said:

 

Thank you for information, Jud.

 

At Mytek's page mentioned "MQA Kernel" trademark.

 

In the patent we see mention of kernel too.

 

This patent relate to digital compression related, if I understand correctly.

 

Thus, I may suppose, if ADC contains "MQA Kernel", it provide compressed and, probably, pre-distorted/filtered signal.

 

See also pages 13-19 and 34-38 here: https://www.xivero.com/downloads/MQA-Technical_Analysis-Hypotheses-Paper.pdf

[JustListen]

On 03/20/2017 at 2:50 AM, jhwalker said:

 

 

it just seems there's a very aggressive anti-MQA movement underway, and I don't really recall seeing such with other new technologies - makes me wonder why.   Probably just me.

Me too.

[audiventory]

3 hours ago, Jud said:

 

See also pages 13-19 and 34-38 here: https://www.xivero.com/downloads/MQA-Technical_Analysis-Hypotheses-Paper.pdf

 

Yes. If analog low frequency filter and digital downsampler filter of ADC don't suppress aliases, it can't be removed after by any kind of known me processing.

But may be MQA suggest other decission. Need see to results of real system.

 

Wider time of impulse at output may be released with band limitation. Resolution (part of ultrasound, as example) will be lost. But it don't degrade sound. Though I don't know why need limit band for "lazy" filter. By my logic, for such case need have sample rate as high as possibly.

 

Anyway recording in MQA system demands editing in ordinary PCM, like DSD. I dont't sure what MQA is better option for it than DSD source record. Though as end-user format in broadcast applications it have advantages by size comparing FLAC, as far as I know.

 

Probably pure MQA recording-palyback system may be use in records-without-editing.

[Jud]

23 minutes ago, audiventory said:

 

Yes. If analog low frequency filter and digital downsampler filter of ADC don't suppress aliases, it can't be removed after by any kind of known me processing.

But may be MQA suggest other decission. Need see to results of real system.

 

Wider time of impulse at output may be released with band limitation. Resolution (part of ultrasound, as example) will be lost. But it don't degrade sound. Though I don't know why need limit band for "lazy" filter. By my logic, for such case need have sample rate as high as possibly.

 

Anyway recording in MQA system demands editing in ordinary PCM, like DSD. I dont't sure what MQA is better option for it than DSD source record. Though as end-user format in broadcast applications it have advantages by size comparing FLAC, as far as I know.

 

Probably pure MQA recording-palyback system may be use in records-without-editing.

 

I think the Xivero fellow is hypothesizing (of course no one knows for sure except the MQA folks) that MQA is using an apodizing filter at the recording/encoding stage to band-limit (not completely effectively), in to avoid some aliasing/ringing at the DAC end, since we know the MQA filtering at the rendering/DAC end is not effective to remove either of those.

 

I suppose we could see how effective any MQA recording/encoding process was at band-limiting by looking at the analog result on playback of a variety of MQA-encoded files.  (I'm thinking of MQA 44.1kHz files in particular, like Beyonce's "Lemonade.")