MQA discussion

[awsmone]

To me the critical issue is whether hires versions of files currently downloadable will be replaced by MQA only versions on the major sites

The three major record labels have an interest in MQA and Warner have converted 50 % of their own files

This could then makes current hires a minority format and MQA dominant

This would be a move away from current trend of offering hires files as open access

MQA is a closed system

How MQA works and whether we believe it sounds good , is of secondary importance to moving away from open access to closed system of file downloads

For people who are happy with 16/44.1 this is less of an issue, and may have no impact at all

It is different potentially from other format wars as whether u liked pcm or dsd u could just purchase it and decide for yourself

I suppose to the average consumer, it may make life easy, as like blu ray, when u only have one format , that's alll u can buy

MQA has the potential to mean in future all digital files come in one format for downloading, and it's a closed system.

I have read Mark and Archimago's work

I find their analysis intriguing, but think digital file download hegemony more of a concern, than what MQA are doing with the files processing whether they are right or wrong

As to the issue of MQA not showing up for a debate, when you own the files, really you don't need to debate the people who don't, now do you ?

 

[adyc]

Very simple. MQA is lossy and high res is not. Why does one want MP3 version of high res music? I don't care whether MQA exists or not as long as the labels will continue to release high res file and not exclusively MQA.

But I don't think it is as simple as that. It seems MQA wants to take control the whole recording and mastering chain. I have serious issues on this.

 

[DaveyF]

But I don't think it is as simple as that. It seems MQA wants to take control the whole recording and mastering chain. I have serious issues on this.

That is the exact reason why I am not so enamored of the process as I was before. The SQ increase over the standard file is very impressive. However, is the price to be paid going to be too great...???

 

[adyc]

That is the exact reason why I am not so enamored of the process as I was before. The SQ increase over the standard file is very impressive. However, is the price to be paid going to be too great...???

I don't doubt your positive experience. But it is a demo. If one needs to do a demo, one will make sure their product is superior during the demo. I think we need more independent evidence. However, technical wise, I don't see how lossy is superior to lossless.

 

[awsmone]

That is the exact reason why I am not so enamored of the process as I was before. The SQ increase over the standard file is very impressive. However, is the price to be paid going to be too great...???

Exactly...

 

[PeterA]

Very simple. MQA is lossy and high res is not. Why does one want MP3 version of high res music? I don't care whether MQA exists or not as long as the labels will continue to release high res file and not exclusively MQA.

But I don't think it is as simple as that. It seems MQA wants to take control the whole recording and mastering chain. I have serious issues on this.

I have a question which may sound naive: If this is indeed the case, will it impede the advancement of digital sound quality? Or is it only about distribution and advancements in SQ will still come in the form of better hardware implementation, as is the case with vinyl playback? We have seen improvements with RBCD on the implementation side which I have heard with the dCS Rossini for example.

 

[rbbert]

I think it will impede the progress of digital audio. I suggested this over a year ago and was roundly criticised by still one, but I still believe this to be true, even more so as time goes on. There is as yet no evidence one way or the other, of course.

 

[Ken Newton]

I have a question which may sound naive: If this is indeed the case, will it impede the advancement of digital sound quality? Or is it only about distribution and advancements in SQ will still come in the form of better hardware implementation, as is the case with vinyl playback? We have seen improvements with RBCD on the implementation side which I have heard with the dCS Rossini for example.

From what I can glean from reading MQA related papers, the primary performance objective of MQA is the optimization of the audio channel for the time-domain, not for the minimization of quantization noise levels. Which is to say, not maximizing sample bit depth. This is where the idea of MQA being lossy stems.

Their own research apparently has led Meridian to conclude that lack of time-domain optimization/prioritization has been a major factor in disappointing sounding digital, both with RBCD and, to a lesser extent, High-Rez. Please see my expanded comment on MQA's performance strategy of optimizing/prioritizing the time-time.

http://www.whatsbestforum.com/showt...on-MQA-is-it-all-it-seems&p=468479#post468479

 

[microstrip]

Very simple. MQA is lossy and high res is not. Why does one want MP3 version of high res music? I don't care whether MQA exists or not as long as the labels will continue to release high res file and not exclusively MQA.

But I don't think it is as simple as that. It seems MQA wants to take control the whole recording and mastering chain. I have serious issues on this.

Perhaps I am wrong, but I got the idea that MQA is just a digital distribution format - mastering will be carried in the original files that will stay property of the recording companies.

 

[adyc]

From what I can glean from reading MQA related papers, the primary performance objective of MQA is the optimization of the audio channel for the time-domain, not for the minimization of quantization noise levels. Which is to say, not maximizing sample bit depth. This is where the idea of MQA being lossy stems.

Their own research apparently has led Meridian to conclude that lack of time-domain optimization/prioritization has been a major factor in disappointing sounding digital, both with RBCD and, to a lesser extent, High-Rez. Please see my expanded comment on MQA's performance strategy of optimizing/prioritizing the time-time.

http://www.whatsbestforum.com/showt...on-MQA-is-it-all-it-seems&p=468479#post468479

I have no problems of time domain optimisation whatever. But does time domain optimisation have to be lossy?

 

[Ken Newton]

I have no problems of time domain optimisation whatever. But does time domain optimisation have to be lossy?

No, it doesn't. High-Rez is perfectly capable of being optimized for the time-domain, and at lower quantization noise than MQA. The problem is there is no industry standard/requirement for time-domain optimization. I suspect this is because of the widely held belief among audio engineers that the sampling theorem provides for the perfect capture and reproduction of a signal, which, in fact, it does, so long as the sample rate is at least twice the frequency of any signal component. However, in systems where some signal components could violate the Nyquist frequency (which is located at half the system's sample rate) then band-limiting is required to prevent aliasing.

The closer the signal band comes to the Nyquist frequency, the sharper must be the band-limiting filters to prevent aliasing. Eventually necessitating employment of brickwall filters. The sharper the filters, the more distorted becomes any time-domain information carried by the signal. This is the situation presented to the audio signal by RBCD; a Nyquist frequency so closely located to the signal band that brickwall filters must be employed. High-Rez (including SACD) features a Nyquist frequency much further removed from the signal band, yet is typically still implemented utilizing brickwall filters, yet it doesn't have to be. I believe that SACD is the one commercial exception which prioritizes the time-domain.

Even should High-Res be optimized for the time-domain at recording, it doesn't support MQA's holistic strategy of time-domain optimization of the system chain from recording all the way through playback.

 

[MQA Truth]

No, it doesn't. High-Rez is perfectly capable of being optimized for the time-domain, and at lower quantization noise than MQA. The problem is there is no industry standard/requirement for time-domain optimization. I suspect this is because of the widely held belief among audio engineers that the sampling theorem provides for the perfect capture and reproduction of a signal, which, in fact, it does, so long as the sample rate is at least twice the frequency of any signal component. However, in systems where some signal components could violate the Nyquist frequency (which is located at half the system's sample rate) then band-limiting is required to prevent aliasing.

The closer the signal band comes to the Nyquist frequency, the sharper must be the band-limiting filters to prevent aliasing. Eventually necessitating employment of brickwall filters. The sharper the filters, the more distorted becomes any time-domain information carried by the signal. This is the situation presented to the audio signal by RBCD; a Nyquist frequency so closely located to the signal band that brickwall filters must be employed. High-Rez (including SACD) features a Nyquist frequency much further removed from the signal band, yet is typically still implemented utilizing brickwall filters, yet it doesn't have to be. I believe that SACD is the one commercial exception which prioritizes the time-domain.

As part of it's comprehensive strategy of prioritizing the time-domain, MQA employs an holistic strategy, optimizing the system chain from recording all the way through playback.

Would you like to know what a NON problem time domain is?
(BTW, when MQA first was shilled, we were told it corrected issues with the ORIGINAL
analog to digital conversion, and that all ADC's had issues with this, this is a farce, because except for
purist 2 channel recordings, virtually all modern recordings have multiple ADC stages.)

Here is John Atkinson reviewing the Ayre ADC:
https://www.stereophile.com/content/ayre-acoustics-qa-9-usb-ad-converter

"I tried a variety of sample rates with these LP rips: 44.1kHz was very good, but didn't capture the essence of the original LPs' sounds; 96kHz was better; but there was no doubt that with a 192kHz sample rate I could not distinguish between the LP and the digital rip. And believe me, I tried. I A/B'd the two versions until blood came out of my ears and I was heartily sick of this music I hadn't heard for, in some cases, decades."

So he could not distinguish between the LP and the digital capture. Let's repeat..he could not distinguish.

No along comes MQA to cure a problem with the very best ADCs we never knew we had. Thanks MQA!

 

[Al M.]

No, it doesn't. High-Rez is perfectly capable of being optimized for the time-domain, and at lower quantization noise than MQA. The problem is there is no industry standard/requirement for time-domain optimization. I suspect this is because of the widely held belief among audio engineers that the sampling theorem provides for the perfect capture and reproduction of a signal, which, in fact, it does, so long as the sample rate is at least twice the frequency of any signal component. However, in systems where some signal components could violate the Nyquist frequency (which is located at half the system's sample rate) then band-limiting is required to prevent aliasing.

The closer the signal band comes to the Nyquist frequency, the sharper must be the band-limiting filters to prevent aliasing. Eventually necessitating employment of brickwall filters. The sharper the filters, the more distorted becomes any time-domain information carried by the signal. This is the situation presented to the audio signal by RBCD; a Nyquist frequency so closely located to the signal band that brickwall filters must be employed. High-Rez (including SACD) features a Nyquist frequency much further removed from the signal band, yet is typically still implemented utilizing brickwall filters, yet it doesn't have to be. I believe that SACD is the one commercial exception which prioritizes the time-domain.

Even should High-Res be optimized for the time-domain at recording, it doesn't support MQA's holistic strategy of time-domain optimization of the system chain from recording all the way through playback.

I am not a digital expert, but from what I understand, brickwall filters are not really necessary, even for RBCD. In the A to D domain, you can avoid analog brickwall fliters by using a higher sampling frequency (e.g., 96 kHz) and then downsampling in the digital domain.

For the D to A conversion, you can avoid steep brickwall filters by oversampling to high frequencies and then filtering from there with a gentle filter.

Or is this wrong?

 

[Ken Newton]

Would you like to know what a NON problem time domain is?
(BTW, when MQA first was shilled, we were told it corrected issues with the ORIGINAL
analog to digital conversion, and that all ADC's had issues with this, this is a farce, because except for
purist 2 channel recordings, virtually all modern recordings have multiple ADC stages.)

Here is John Atkinson reviewing the Ayre ADC:
https://www.stereophile.com/content/ayre-acoustics-qa-9-usb-ad-converter

"I tried a variety of sample rates with these LP rips: 44.1kHz was very good, but didn't capture the essence of the original LPs' sounds; 96kHz was better; but there was no doubt that with a 192kHz sample rate I could not distinguish between the LP and the digital rip. And believe me, I tried. I A/B'd the two versions until blood came out of my ears and I was heartily sick of this music I hadn't heard for, in some cases, decades."

So he could not distinguish between the LP and the digital capture. Let's repeat..he could not distinguish.

No along comes MQA to cure a problem with the very best ADCs we never knew we had. Thanks MQA!

I'm neither supporter nor detractor of MQA. I've simply offered my outsider's critique of what they seem to be about as far as audio performance startegy. However, I've read enough of your comments to have observed that you are a determined MQA detractor. My experience is that discussions with folks who arent open minded are a fruitless wast of effort. So, I won't. I hope you understand.

 

[Ken Newton]

I am not a digital expert, but from what I understand, brickwall filters are not really necessary, even for RBCD. In the A to D domain, you can avoid analog brickwall fliters by using a higher sampling frequency (e.g., 96 kHz) and then downsampling in the digital domain.

For the D to A conversion, you can avoid steep brickwall filters by oversampling to high frequencies and then filtering from there with a gentle filter.

Or is this wrong?

In either downsampling or upsampling, brickwall filters are employed, particularly so with a 44.1KHz sample rate (22.05kHz Nyqust frequency) system. In the digital domain, brickwall filtering is usually performed with low-pass FIR digital filters. The technical reasons for this are not a simple matter to discuss here. Suffice to say that aliasing would otherwise occur during downsampling, and the image bands need to be cut-off (filtered) for upsampling. The lowest image band of RBCD starts at only 22kHz, so, brickwall filtering is still needed if we're to cut-off 22kHz and above, while leaving 22kHz and below unmolested. The filter performance requirements are, however, much more relaxed via High-Rez.

High-Rez (96kHz and 192kHz) doesn't actually require brickwall filtering to capture the audio band without aliasing, it's just that the commercial ADC and DAC chips typically already feature such filters by default. Custom filtering could instead be employed, but that must be a planned effort by the production engineer. I have always thought that some sort of logo or graphic symbol identifying a High-Rez album as being time-domain optimized, should the production engineer have done so, would be very useful for audiophile consumers. Even should an high-rez album be time-domain optimized, the playback end of the chain (the DAC) would still represent an uncontrolled random element of sorts because there is not an industry standard for a PCM based time-domain optimization, outside of MQA.

 

[PeterA]

With regards to industry standards, is the industry moving more rapidly towards a downloading/storage standard with priority towards time domain optimization like MQA or towards a multi channel format? Where is MQA in the whole multi channel discussion? And has anyone asked what the consumers really want, or is this more or less decided by the industry and then pushed onto the consumer?

 

[Ken Newton]

With regards to industry standards, is the industry moving more rapidly towards a downloading/storage standard with priority towards time domain optimization like MQA or towards a multi channel format? Where is MQA in the whole multi channel discussion?

I'm just a hobbyist, and not an industry insider so, I couldn't say.

And has anyone asked what the consumers really want, or is this more or less decided by the industry and then pushed onto the consumer?

I suspect that the industry knows what the mass consumer wants. The basic drivers are probably the same as they are for any type of consumer product; more convenience of use, and lower cost. As for the audiophile consumer, either they don't know, or don't care, or both.

 

[KeithR]

guys, we've had hirez downloads for a decade that serve only a tiny market and are over 60% classical. the goal of MQA is to increase digital distribution and get hirez to the masses for a much broader musical experience. If its not streamable, there is no solution.

 

[mauidan]

The Politics of MQA:

http://www.theabsolutesound.com/articles/the-politics-of-mqa/

 

[KeithR]

The Politics of MQA:

http://www.theabsolutesound.com/articles/the-politics-of-mqa/

"Think about it. This is music arriving in your room via the Internet, for a fairly minimal monthly charge—an exponentially increasing body of recordings for playback with sound quality that’s, at the very least, as good as the best present high-resolution digital files. Imagine being told 20 years ago that this was the future of the high end. We all would have been astounded, energized, and maybe even joyous."

bingo. this is the potential.