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

[abdo123]

Has there been properly done research by others ?

yes and people couldn't tell the difference between MQA and WAV.

https://www.aes.org/e-lib/browse.cfm?elib=19396

 

[KSTR]

Didn't we have simlilar AES papers back when MP3 was the hot stuff?

 

[abdo123]

Didn't we have simlilar AES papers back when MP3 was the hot stuff?

from the 'papers' i have read on AES, they're borderline a predatory journal.

the papers' transcripts follow no real academic structure, their language is intentionally difficult to understand, and their abstracts have no real meaningful information and is there to look 'sciency'

At least the name of the journal represents that, they're a bunch of engineers who wanted to do 'science'.

He asked if anyone did any research and there are two papers on AES about it, I'm a biologist so i can't really critique the audio science in the paper.

 

[GoldenOne]

Any Journal is at the end of the day the product of those that contribute to it.
The AES absolutely, unquestionably has a wealth of incredible information. But a Journal contributed to by and primarily made up of people and firms that are trying to sell products is inherently going to be quite different to a Mathematical Journal for example.

 

[Lunafag]

yes and people couldn't tell the difference between MQA and WAV.

With the bitrate of cd quality flac it damn better be audibly transparent lol, only a complete failure of a codec wouldn't.

 

[mansr]

At least the name of the journal represents that, they're a bunch of engineers who wanted to do 'science'.

That's insulting to engineers and misses the point. It seems to me that the AES (now) serves mainly to prop up the business interests of its members and as a club for collective self-admiration.

 

[PO3c]

yes and people couldn't tell the difference between MQA and WAV.

https://www.aes.org/e-lib/browse.cfm?elib=19396

That interpretation might not be in favor of MQA given one need products with special sauce playing back 24/96 MQA content.

And now with what seem to be how Tidal no longer allow MQA content to passthough non licened transports yet another important selling point of MQA are crumbling.

Borrowed form https://community.auralic.com/t/from-us-auralic-vs-drm/7328/16

 

[abdo123]

That's insulting to engineers and misses the point. It seems to me that the AES (now) serves mainly to prop up the business interests of its members and as a club for collective self-admiration.

you know what's more insulting? reading an abstract and it having neither first order or second order results.

 

[mansr]

you know what's more insulting? reading an abstract and it having neither first order or second order results.

I agree with your characterisation of AES "papers." It's your disdain for engineers I take offence with. Have you ever read any IEEE papers? One of those Es stands for engineering too. Now engineering is definitely distinct from science, and should be. That doesn't mean engineers are somehow lesser than "real" scientists and deserving of scorn. Save that for the social so-called sciences.

 

[ebslo]

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?

The big joke about "perceptually lossless" in the context of MQA is that the entire purpose of MQA is to preserve content that is perceptually irrelevant. In perceptual terms it is at best equivalent to a down-sampler or null codec.

Now I need a shower after having typed "perceptually lossless". Make that two showers, because I did it again.

 

[BlackTalon]

Nothing to be offended by in my view. Scientists are theory and the engineers actually figure out how to get things done based on that science. Without engineers there would be very little practical application for scientific research. We are the 'get stuff done' people.

 

[Jimbob54]

I agree that it is very important to try and avoid looking at things in terms of binary oppositions, however tempting it is. We are all guilty of it sometimes (I know I am), but generally speaking, things rarely tend to fall in "pure evil" vs "pure good" category in the real world.

Personally, my issue was never with MQA, but with Tidal for using MQA tracks instead of original PCM FLACs for their lossless HiFi tier. Even then, it was a question of principle rather than me being able to actually hear any real difference to be honest.

I totally understand @amirm's fascination with MQA's approach to encoding. As a consumer and an audio-enthusiast, I am also interested in seeing the audio industry move beyond PCM at some point in time, but it will have to be in terms of both recording and reproduction if we are to see a significant leap in fidelity. You cannot conjure things up in reproduction that were never captured in the first place during the recording process. Who knows, perhaps some lessons learned from MQA's approach will be incorporated in whatever comes after PCM.

Even though I have fired off fairly severe criticism at both Tidal and MQA for their marketing shenanigans in the last few weeks, I can concede that it may be unfair to oversimplify things and label MQA as pure marketing DRM scheme with no benefit to the science of audio recording and reproduction whatsoever.

The intention of MQA was always end to end - I just dont think it has/ had any traction at the recording end, leaving it as the "thing" it is now.

EDIT- I dont know if there would have been any benefit to the end listener even if a recording was end to end MQA over the MQA encoding of existing PCM.

 

[mansr]

The big joke about "perceptually lossless" in the context of MQA is that the entire purpose of MQA is to preserve content that is perceptually irrelevant. In perceptual terms it is at best equivalent to a down-sampler or null codec.

You appear to be confused. What you describe is the excuse for MQA. It's purpose is to enrich Bob Stuart.

 

[Jimbob54]

I agree with your characterisation of AES "papers." It's your disdain for engineers I take offence with. Have you ever read any IEEE papers? One of those Es stands for engineering too. Now engineering is definitely distinct from science, and should be. That doesn't mean engineers are somehow lesser than "real" scientists and deserving of scorn. Save that for the social so-called sciences.

Economists have feelings you know . *Sad face*

 

[PierreV]

He brought a album from bandcamp,not from a major label,maybe he didn't know how to ask for the non mqa version

Obviously, you have no idea who/what Kronos Quartet is, what the track record of Nonesuch Records is (https://www.nonesuch.com) and as Amir confirmed below, who Bob Ludwig is (https://en.wikipedia.org/wiki/Bob_Ludwig).

 

[bboris77]

A bit of a tangent here - this MQA debate has inspired me to try and learn a bit more about the inherent limitations and challenges of PCM when it comes to recording and digitizing real-life audio sources. Are there any good resources that discuss this subject matter in more depth? I mean, we spend so much time arguing about the digital means of audio reproduction, but if we are putting the same old PCM-based lossy recording in, we can't really expect anything to improve down the chain.

 

[abdo123]

I agree with your characterisation of AES "papers." It's your disdain for engineers I take offence with. Have you ever read any IEEE papers? One of those Es stands for engineering too. Now engineering is definitely distinct from science, and should be. That doesn't mean engineers are somehow lesser than "real" scientists and deserving of scorn. Save that for the social so-called sciences.

i didn't mean to be scornful, but it is what it is.

The science engineers do is useless if they can't express it properly. if you want to partake in academia and publish papers then you need to be part of academia and receive proper training to do so.

the scientific community is RUTHLESS when it comes to writing, because if you fail to communicate the results of a 100k$ experiment then that's 100k$ wasted, if you leave 'space' for interpenetration then your entire research could be invalidated.

Scientists are basically the personification of masochism and sadism in one person, your life's work is stepped on by other people, and you step on other people's work (peer review). having someone tell you 'your statistical model is invalid and your results are insignificant' is the equivalent of a dominatrix stepping on your balls sans the pleasure.

I (and scientists in general) can not be an engineer, my knowledge while very deep is limited to few topics in my domain. While the knowledge of an engineer is less deep in the same topics that I focus on, their general knowledge of the domain is much more comprehensive than mine.

an experienced engineer can make a fully functioning product/project with the assistance of several technicians. I can never do that.

However, a big team of scientists can unite their efforts and become a 'Super engineer' (think NASA), but a team of engineers cannot unite and become a 'Super scientist' or even a regular scientist without training.

 

[PierreV]

A bit of a tangent here - this MQA debate has inspired me to try and learn a bit more about the inherent limitations and challenges of PCM when it comes to recording and digitizing real-life audio sources. Are there any good resources that discuss this subject matter in more depth? I mean, we spend so much time arguing about the digital means of audio reproduction, but if we are putting the same old PCM-based lossy recording in, we can't really expect anything to improve down the chain.

I suspect that, even if the whole chain suddenly became a paragon of transparency and honesty, if maniacal engineers monitored every step from the initial recording to the end-user content delivery, we would still not hear significantly "better" than what CD quality offers due to the limitations of our hearing. And that is the core reason why the audio/audiophile industry is replete with cons from fake hi-res tracks to myrtle wood cable lifters...

 

[mansr]

The science engineers do is useless if they can't express it properly.

That's true for everybody. Why do you feel the need to single out engineers? AES entries are written by businessmen pretending to be engineers. That is the real problem.

 

[Raindog123]

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?

Woke up, getting to my shift (yet to read a couple of pages-full…)

Without additional signal processing functionality, the basic PCM data stream would involve FLAC uncompressing, and that’s it (plus the associated data moving in/out memory and I/O interface (USB). To compare apples and apples - comparable sound quality - let’s assume PCM at 24/96ksps (kilo samples per second), and has audio content up to 48kHz.

The MQA stream would also requre the same FLAC processing, but coming at 24/48ksps, that step will be half the effort. However, next will be the MQA core decoding step - splitting each sample in two, buffer a sliding window of samples, and applying some ‘secret sauce‘ transformation to those. I do not thing it’s as bad as an FFT (fast Fourier) but it most definitely involves some interpolation-associated filtering, lookups, etc. So, this step will be rather complex (and ultimately performed on the higher - 24/96ksps). Resulting in reduced but apparently ‘all that needed for real music’ content at the same - up to 48kHz… So the net will be that MQA has higher processor load. [But as the data rates are rather low, I do not think that this extra load is excessively large. And it might be a good benchmarking/profiling test for the future, with MQA cooperation.]

The Tidal app on Win does not give you the full unfolding - only the first “core” (to 24/96ksps) above. So if you want ‘full MQA experience’ - per MQA’s own explanation of their operation - you will have to purchase a ‘MQA rendering’ DAC device. It applies some correcting digital filters and (sometimes) upsamples the MQA stream even further (to 192ksps)…. Now, the caveat. While you do need a dedicated MQA hardware for this step, many (including myself) believe that this MQA-specific filtering step (that MQA calls ‘second unfold’) is no different form the [interpolation and upsampling] filtering also performed by any non-MQA DAC. So, in reality there is no improvement, thus no need for MQA DAC ‘investment’ [A controlled blind - with/without (ABX) - listening test would answer this ‘many believe‘ once and forever.]