MQA discussion

[DaveC]

I got to ask Keith Johnson of Reference Recordings about MQA in person yesterday during the Avalon demo near RMAF.

I wish I knew more about digital music so I could ask intelligent and relevant questions, but I don't. In any case what I got out of it is that MQA is an apodizing filter that corrects for brickwall filtering used in the mastering process. At higher bitrates the brickwall filter isn't used, and other gentler filters with less side effects are used instead. So MQA won't be an improvement in this case while it may indeed improve other digital recordings that used brickwall filters.

Keith is not planning on installing MQA hardware in Spectral DACs right now, although he seemed excited about the possibility of MQA being able to contain different masterings that are intended for speakers, headphones or multi-channel listening.

edit:

Keith also mentioned the effect of the MQA filters is to shift the soundstage forward which some like but is just an artifact and not accurate.

Also, here's some info on Meridian's apodizing filter:

https://mrapodizer.wordpress.com/2011/08/16/technical-analysis-of-the-meridian-apodizing-filter/

 

[Al M.]

The so-called authentication in MQA is not doable in mass production, since you would have to actually listen to everything before it can be approved. It also does not appear to be done -- one person reported on digital clipping of files on Tidal that apparently had not been listened to and thus could not have been 'authenticated' (the files were later taken down).

All this is discussed here (lower half of the page)

https://www.computeraudiophile.com/forums/topic/30381-mqa-is-vaporware/?page=184

and here:

https://www.computeraudiophile.com/forums/topic/37226-mqa-rmaf-live/

(bottom of first page and continuing on second page)

It appears that in practice the whole authentication scheme is just a scam, a promise that is not and cannot be kept -- at least when it comes to mass output of MQA conversion. The theoretical promise of MQA is a different matter.

 

[awsmone]

Hi Al

Thanks for the links

Very depressing ....the claims of some form of authenticity by MQA seem to be blown

When I read the speed with which Warner were converting to MQA I found it hard to believe they were checking with anyone external to the process ...that more than 3000 titles, to get feedback on

If they were, then maybe they could publish some random letters of approval from artists to prove their claims, not just high profile reissues ....

 

[ack]

I got to ask Keith Johnson of Reference Recordings about MQA in person yesterday during the Avalon demo near RMAF.

I wish I knew more about digital music so I could ask intelligent and relevant questions, but I don't. In any case what I got out of it is that MQA is an apodizing filter that corrects for brickwall filtering used in the mastering process. At higher bitrates the brickwall filter isn't used, and other gentler filters with less side effects are used instead. So MQA won't be an improvement in this case while it may indeed improve other digital recordings that used brickwall filters.

Keith is not planning on installing MQA hardware in Spectral DACs right now, although he seemed excited about the possibility of MQA being able to contain different masterings that are intended for speakers, headphones or multi-channel listening.

edit:

Keith also mentioned the effect of the MQA filters is to shift the soundstage forward which some like but is just an artifact and not accurate.

Also, here's some info on Meridian's apodizing filter:

https://mrapodizer.wordpress.com/2011/08/16/technical-analysis-of-the-meridian-apodizing-filter/

I did notice that during my MQA demo everything sounded quite a bit closer to me, which was pleasing

 

[DaveC]

I did notice that during my MQA demo everything sounded quite a bit closer to me, which was pleasing

It's a result of artifacts caused by the apodizing filter which afaik are the result of aliasing noise making it through the filter. Keith did not like this but said this is why some like MQA, I suspect like any artifact it would get tiring long term.

 

[Al M.]

It's a result of artifacts caused by the apodizing filter which afaik are the result of aliasing noise making it through the filter. Keith did not like this but said this is why some like MQA, I suspect like any artifact it would get tiring long term.

Interesting. Before I had external power supplies for my tube amps, the sound was much more forward, which could be pleasing. After the insertion of the external power supplies, the sound became in many cases more recessed (putting the musicians more into a performance space), because low-level spatial information came through that had previously been suppressed by the electronic noise from the internal power supplies, and that was now cleaned up. It could be that the aliasing noise from MQA functions in a similar way to electronic noise from gear, suppressing low-level spatial information and thus shifting the soundstage more forward.

This effect could also explain why in some A/B demos people preferred MQA. If the system was good at portraying spatial depth, and thus there was a marked difference in spatial portrayal between non-MQA and MQA, then a more forward presentation of MQA could have made it sound 'clearer' and certainly, louder. And a presentation of greater loudness -- perceived or actual -- will always win.

 

[awsmone]

It's a result of artifacts caused by the apodizing filter which afaik are the result of aliasing noise making it through the filter. Keith did not like this but said this is why some like MQA, I suspect like any artifact it would get tiring long term.

That is correct the filter is leaky to aliasing allowing 32 dB through in the 0-7khz band

I am unclear if this has been shown to be audible, and am interested in these comments

 

[ack]

Right, I did point this out from Al's article; here it is again:

MQA has been described as having a “leaky filter.” This should be no surprise; in MQA patent #WO2015189533 A1, “Digital Encapsulation of Audio Signals,” they state: “We accept that aliasing may take place and are proposing to balance aliasing against ‘time-smear’ of transients

 

[rbbert]

John Siau of Benchmark presents an interesting take on MQA in his Manufacturer's response in the latest Stereophile (the Benchmark DAC3 was reviewed). Somewhat complementary to KOJ's paraphrased comments above, he says the DAC3 sounds similar to MQA because he uses wide-bandwidth analog stages to preserve accurate bass response essentially to DC. He says this maintains accurate time response by making sure the bass isn't delayed compared to the higher frequencies, whereas MQA does that instead by delaying the higher frequencies, combined with a minmum-phase (apodizing) filter to lessen the effects of brickwall filtering, although Siau calls that a "Band-Aid", presumably because using higher sampling rates and avoiding the need for brickwall filtering is a better solution.

 

[Al M.]

John Siau of Benchmark presents an interesting take on MQA in his Manufacturer's response in the latest Stereophile (the Benchmark DAC3 was reviewed). Somewhat complementary to KOJ's paraphrased comments above, he says the DAC3 sounds similar to MQA because he uses wide-bandwidth analog stages to preserve accurate bass response essentially to DC. He says this maintains accurate time response by making sure the bass isn't delayed compared to the higher frequencies, whereas MQA does that instead by delaying the higher frequencies, combined with a minmum-phase (apodizing) filter to lessen the effects of brickwall filtering, although Siau calls that a "Band-Aid", presumably because using higher sampling rates and avoiding the need for brickwall filtering is a better solution.

Schiit claims something not too dissimilar, that it's digital filter is optimized for BOTH frequency and time domain. Mike Moffat says developing the algorithm took a university math professor and solving a 'divide-by-zero' problem; it's usually claimed that optimization can only be for either of these two domains. Developing was back in the Theta Digital days, now the algorithm runs on the much more powerful Analog Devices SHARC processors which implement it more optimally.

All that, Benchmark and Schiit, is on the DAC side though. MQA claims that it can reverse, or at the least alleviate, the 'pre-ringing' on the A/D side -- after the recording was made. Seems a highly dubious claim to me. (Not that I think pre-ringing is a big issue in practice anyway.)

 

[opus112]

MQA claims that it can reverse, or at the least alleviate, the 'pre-ringing' on the A/D side -- after the recording was made. Seems a highly dubious claim to me. (Not that I think pre-ringing is a big issue in practice anyway.)

Its possible with an apodizing filter - the pre-ringing is roughly at a frequency equal to the corner frequency of the anti-aliasing filter. Which is normally 20kHz or above. So if the output of the ADC is band-limited with a filter with narrower bandwidth, the pre-ringing can be filtered out. But there will then at the very least be post-ringing at a lower frequency, assuming a minimum phase apodizing filter's used.

I am not sure that there's any evidence that pre-ringing at frequencies beyond most people's hearing is audible so I think this could just be a distraction.

 

[Al M.]

Recording engineer Brian Lucey on MQA:

https://www.gearslutz.com/board/12548751-post460.html

Brian Lucey's website:

http://magicgardenmastering.com/

 

[adyc]

Recording engineer Brian Lucey on MQA:

https://www.gearslutz.com/board/12548751-post460.html

Brian Lucey's website:

http://magicgardenmastering.com/

++++1

 

[awsmone]

Recording engineer Brian Lucey on MQA:

https://www.gearslutz.com/board/12548751-post460.html

Brian Lucey's website:

http://magicgardenmastering.com/

There are two premises here

1. That it sounds thin
2. Audiophile’s are easily fooled

 

[KeithR]

up to almost 8,000 titles:

http://www.meridianunplugged.com/ubbthreads/ubbthreads.php?ubb=showflat&Number=269510#Post269510

wow.

 

[audioguy]

I am not a technical guy but I just read an article on MQA in (I think) The Absolute Sound, and the statement is made that MQA is better than the original. How is that even possible? Better than the original?????

Are we dealing with the 2017 version of "The Emperor's New Clothes" ? ..... or maybe the quote attributed to P.T. Barnum (There is a sucker ...................).

 

[adyc]

I am not a technical guy but I just read an article on MQA in (I think) The Absolute Sound, and the statement is made that MQA is better than the original. How is that even possible? Better than the original?????

Are we dealing with the 2017 version of "The Emperor's New Clothes" ? ..... or maybe the quote attributed to P.T. Barnum (There is a sucker ...................).

Yes, it is emperor new clothes. No need to pay attention to TAS. The reviewers are never digital signal experts. I rather listen to people who have deep digital signal knowledge.

 

[still-one]

Yes, it is emperor new clothes. No need to pay attention to TAS. The reviewers are never digital signal experts. I rather listen to people who have deep digital signal knowledge.

I listen to my ears.

 

[audioguy]

I listen to my ears.

Me too. Unless this is one of those technologies where you need a bazillion dollar audio system and/or perfect hearing, I don't hear what the proponents are hearing. Listened (blind) to some music recorded standard and with the "new clothes", and I must be missing something. Which works out perfectly since I was OK with the plain good 'ol CD format. This may be one of those technologies where "squinting" is required to hear the differences and if so, I have no interest.

I know many on this forum despise blind listening (still don't understand why) but to eliminate any "expectation bias" try it with some music recorded both ways, be honest and see how well you do ............ or not!!!

 

[still-one]

Me too. Unless this is one of those technologies where you need a bazillion dollar audio system and/or perfect hearing, I don't hear what the proponents are hearing. Listened (blind) to some music recorded standard and with the "new clothes", and I must be missing something. Which works out perfectly since I was OK with the plain good 'ol CD format. This may be one of those technologies where "squinting" is required to hear the differences and if so, I have no interest.

I know many on this forum despise blind listening (still don't understand why) but to eliminate any "expectation bias" try it with some music recorded both ways, be honest and see how well you do ............ or not!!!

I am not going to get into one of WBF's infamous blind listening debates . If you don't hear differences at least you tried, that is better than most. I too have no issue with plain old CD quality and that is how I do most of my listening. I do own some hi-rez files. I own a few MQA titles. I the many comparisons I have performed I have never heard the MQA version sound worse than the CD or streamed version. There are some that the differences are negligible, while most sound much better.