MQA Deep Dive - I published music on tidal to test MQA

[Mountain Goat]

No - from the information capacity point of view - MQA is significantly worse, as at higher sampling rates (ie after its core unfold) MQA can represent only a small fraction of higher-frequency (‘ultrasonic‘) spectrum. While a higher-sampling-rate PCM does it fully, naturally. And even the feasibility of this ‘small fraction of spectrum’ is questioned by signal-processing algorithm experts. (Pointing out that while MQA’s ‘triangular’ spectrum is a nice visual illustration in frequency domain, performing such PSM decoding in time domain is not feasible in the near-real-time of playback.) But those are ‘opinions’, meanwhile - if exists - the encoding/decoding algorithm is kept secret, and its performance test results are not offered…

At the same time, MQA’s position is that this - the diminished ability of MQA encoding to carry ultrasonic information - is irrelevant and even by-design, as ‘only a small amount of music information exists at ultrasonic frequencies - due to significantly reduced level and dynamic-range of real music there.’ (This might be true - but it makes MQA’s algorithm (again, if it actually works) at best equal to PCM.)

If PCM and MQA are equal in audible sound quality, which has the worst processor load for penalty for decoding? With and without MQA hardware? I always listened to Tidal with the Windows App, does that give the full "unfolding"? Or do you have to invest in MQA hardware to get the best out of MQA?

I know these are elementary questions, but I'm more confused as this goes on...

 

[awdeeoh]

If MQA decoder is free I'd rather have the MQA file, plain FLAC is just a waste of space!

 

[MDT]

You see where the real battleground is? We need to get the masters cleaned up and have them released without loudness compression.

This is a battle I can certainly get behind. Getting access to the original master recording would be ideal.

But wait, wasn't this supposed to offer that :

 

[mSpot]

You think you got harmed? No, *I* just got harmed to the tune of nearly $20!!! Read on.

I looked up the album you bought on Tidal. It is available and it is indeed in MQA. I went to capture its spectrum as it played and was very surprised to see the brick wall above 22 kHz! So I got curious as to what the original 96 kHz version looked like. Fortunately one of my favorite high-res stores, Prostudiomasters had it:

View attachment 132505

Notice how they sell both the MQA and non-MQA versions. Cough up the money and buy the album. The non-MQA is indeed, non-MQA:

View attachment 132507

This is what the spectrum of the third track looks like (similar to tracks 1 and 2):

View attachment 132506

Son of gun! The actual (live) recording is at 44.1 kHz but is somehow mixed into a high-res envelop. We know this because some noise has been added there including that one peak at nearly 48 kHz.

The most shocking part is who mastered this: the famous Bob Ludwig!!! From the credits: "Mastered by Robert C. Ludwig "

Un.... believable. This is like buying a $1000 DAC from a famous designer and finding an Apple dongle in there!

You see where the real battleground is? We need to get the masters cleaned up and have them released without loudness compression. That is the battle we should put our energy toward. Not these little squabbles over MQA.

I am going to go on hot dogs and water dinners for the next few days to make up for the $20 loss on this....

You were ripped off. The 24/96 download is only $15 at Presto Classical:
https://www.prestomusic.com/classical/products/8654725--terry-riley-sun-rings

The record label is Nonesuch, and it is interesting that their website does not offer this album in 24/96:
https://store.nonesuch.com/catalogsearch/result/index/?q=terry+riley+sun+rings
However, Nonesuch does offer other hi-res downloads for sale: https://store.nonesuch.com/music/hd.html

 

[THW]

3. THE FRIGGIN SLANDER. HOW THE HELL IS THERE SO MUCH ANGER HERE TOWARDS AMIR FOR BEING SNARKY AND NOT NEARLY THE SAME LEVEL OF VITRIOLE FOR OP's PETTINESS AND SELFISHNESS??

like most people i am a smoothbrain when it comes to this sort of stuff, but the slander being posted about amir in other forums reaaaally rubs me the wrong way and makes me think OP isn’t arguing in good faith.

 

[krabapple]

Thanks for the prompt responses. So, if the null test conclusively proves that the MQA version is not identical to the original audio recording in the PCM format, I just do not understand how they can claim that all of the musical content is retained.

It rather depends on whether you can hear the missing content, doesn't it?

What is being called 'perceptually lossless' is , it seems to me, another way of saying 'transparent'. It's what lossy codecs strive for. The two files are different sizes, but they sound the same.

However, as far as I can glean, no actual DBT data have been published by Meridian or anyone else to verify this for MQA.

Along those lines I naturally wonder if Amir could use his 'tricks' (not really tricks, just very not-normal listening regimes) and training to detect differences between MQA encoded and PCM source. For example, for 16 bits vs 24, a 'trick' is to select an extremely quiet part of the music, and listen to it at a very high playback level. A level that would be deafening for most of the track. To detect 320kbps (as well as, I presume, very high VBR) mp3 encoding by a high quality encoder , a trick is to use a 'killer' musical selection, difficult to encode, and/or to zero in on a micro-instant that 'tells' and play that over and over until you learn to hear the difference.

I wonder how MQA would fare in such tests that employ very non normal listening....and would it still qualify as 'perceptually lossless' if it failed them?

 

[KSTR]

You see where the real battleground is? We need to get the masters cleaned up and have them released without loudness compression. That is the battle we should put our energy toward. Not these little squabbles over MQA.

Ack. If the industry were any clever they would offer different masters for different user groups. I have so many CDs which are really ruined by loudness war (and I'm quite tolerant in that regard) that once the honey-moon phase is over, I can't enjoy them anymore -- well, after 5 beers, I might ;-)
Even better yet would be pre-master stereo-stems of typical instrument groups, basically the bus signals during final mixdown. For example, in almost all recordings I find the vocals (if present) way too loud...

 

[KSTR]

If MQA decoder is free I'd rather have the MQA file, plain FLAC is just a waste of space!

A reduction factor of two is irrelevant. 10 is the goal (and the reason why MP3 was so successful initially).

 

[KSTR]

As for the impulse / frequency response discussion, there is nothing special or genuine about MQA's approach to use a specialized shallow, almost Gaussian filter, starting one octave below fs/2. They just gave it a fancy name.

 

[krabapple]

Have to express some amusement that MQA issues turn in part on how well it retains 'hi rez' (ultrasonic) content, given that Bob Stuart was the biggest cheerleader for high rez (higher sample rates, more bit depth in consumer formats) back in the pre millennium days just before DVD Audio (which Meridian championed and served with its MLP encoding) and SACD were foisted on us. Indeed the publication of his solely authored JAES paper on Coding for High Resolution Audio raised the hackles of some fairly big names (like Lipshitz and Vanderkooy) among JAES members versed in digital audio, who published something like a protest letter a few issues later.

The listening test evidence for why the consumer 'needs' high resolution audio in the home was shall we say rather thin at the time, for one thing.
(It remains so...'perceptual losslessness' notwithstanding).

And now to learn that MQA is about 'lossy high rez'...who would have guessed it would turn out this way?

 

[lamode]

MQA explains exactly what this means.

Sorry but they haven't explained anything at all.

 

[amirm]

Sorry but they haven't explained anything at all.

They can only explain it. Making you understand it is your responsibility.

 

[lamode]

MQA's approach of noticing the statistical aspects of music and encoding that is appealing to me. It is simply neat!

You could say the same about MP3. Surely a lossless codec is even neater!

 

[solderdude]

It looks like MQA is going to stay for quite a while to come. Folks will have to deal with that.
Eventually it will disappear along with other formats that have existed till now.
25 years from now, when someone mentions MQA, people will have forgotten about it already.
It will never reach 'vinyl' or even 'RtoR' status either.

I haven't read many reviews where folks complained about bad SQ.
Only about lossy vs lossless. We all know that it is lossy by now.
Also about the need for >48kHz which is debatable.
It's about digital BW and the money grabbing.

A lot of fuss.. understandable... but fortunately for me irrelevant.

 

[amirm]

You could say the same about MP3. Surely a lossless codec is even neater!

Mathematically lossless codecs are quite boring actually. The best and the worse are pretty close to each other. They don't have any kind of psychoacoustic model in them which is another reason they are boring. They also become less efficient the nosier the content even though we don't care much about the noise.

Perceptual codecs are completely different game and have a far tougher job so they are indeed much more exciting as far as technology. There is a ton more research papers on these than on lossless encoding.

 

[solderdude]

Technically more challenging and 'interesting' yes... better than plain 24/192 it is not regardless of the lossless compression.
Understandable viewpoint for someone that worked on a lossy codec in a time where BW and memory was an issue.

 

[lamode]

They can only explain it. Making you understand it is your responsibility.

They have only responded with deflections and gibberish which hardly qualifies as a genuine explanation. It's a little insulting to your forum members to claim that their lack of comprehension is the problem.

Mathematically lossless codecs are quite boring actually. The best and the worse are pretty close to each other. They don't have any kind of psychoacoustic model in them which is another reason they are boring. They also become less efficient the nosier the content even though we don't care much about the noise.

If bit-perfect is boring then give me boring. I don't understand why you would be negative about that. I also don't see anyone complaining that FLAC isn't efficient enough already.

 

[amirm]

They have only responded with deflections and gibberish which hardly qualifies as a genuine explanation. It's insulting to your forum members to claim that their lack of comprehension is the problem.

I had no problem following their answers. I then explained them more simply here. For you to keep saying you don't know what they are saying only reflects on you.

 

[lamode]

I had no problem following their answers. I then explained them more simply here. For you to keep saying you don't know what they are saying only reflects on you.

Wow, that response tells me all I need to know about you. I'll leave it there.

 

[amirm]

If bit-perfect is boring then give me boring. I don't understand why you would be negative about that. I also don't see anyone complaining that FLAC isn't efficient enough already.

You seem to have serious problem following a discussion. I explained that I like the technology underlying MQA. You asked if some other technologies are just as cool and I explained that. This has nothing to do with what you as a user and lay person like. Don't ask me stuff if you can't even follow the thread of what is being discussed. And make sure your next post is informative and not complaining in nature or you will get a reply ban.