Comparison of PCM and MQA

I believe your experience, for me it is the oppposite though. Which DAC do you use for this test?

1 Like

Hi. Did you compare Tidal to Qobuz without knowing which was which?

My DAC isn’t MQA compatible so I’m using Roon to do the first unfold only but given that MQA is marketed as sounding better than CD even with no unfolding or upsampling, this shouldn’t matter. I find it odd that we have a different preference as to what sounds best!

I’ve preferred MQA versions on other comparisons I’ve done in the past but the more I do, the more I seem to prefer the PCM versions. If I was to generalise, I find MQA sounds a little flatter, darker and less alive than non MQA.

What I find really odd about these MQA threads is that there is very little discussion about specific tracks and how the versions sounds different - it just seems to be endless technical and political discussion about the MQA format itself. I’m just happy to listen to whatever version sounds best.

3 Likes

This was a post from Rugby a couple of days ago.

RugbyDaniel BeyerCommunity: Moderator

6d

The quality of Tidal’s content, non-MQA, went downhill last Oct. Not sure what it was, but, it is extremely noticeable. So much so, that I actively choose the Qobuz CD version over the Tidal CD version every time now, and again, nothing to do with MQA.

Make sure to check whether Tidal and Qobuz sound the same on your system with a range of identical non-MQA CDs before assuming differences are due to MQA or PCM. Especially timbral differences like bass, which give a darker sound.

Hi, listening to 1st unfolds sounds very good, but on a good system, the full end-to-end MQA is the best experience. This is related to the conjugating filters in the MQA DAC.

3 Likes

on any system, full end-to-end MQA is the best MQA chips can offer.

I can also say…

1st unfolds sound very good, but on a good system, 1st unfolds can provide a better experience by not using generic suboptimal digital and analog filters in the mqa chip and incorporating a higher quality filter, modulator and upsampling to higher DSD rates provided by hqplayer. ymmv.

Just sayin’… :wink:

Hi, I find the method using HQplayer interesting and DSD sounds good. MQA certification is not limited to the DAC chipset only though. 2nd and 3rd unfolds also sound better to my experience. Greetings

1 Like

The problem is, the “deblurring” is technically impossible. With a couple of exceptions, most TIDAL “masters” of non-recent music just seem to have boosted bass, and in the case of what originated as reasonably high-quality recordings, have a lower level of detail than the original.

In the case of classical music especially, the subtle details that express the depth of the soundstage are destroyed and the music ruined.

Quite a few MQA-flagged 44.1/16 files on TIDAL have also been analysed, showing that they are bit-perfect identical to the non-MQA versions. This is clearly dishonest behaviour on TIDAL’s part.

The fewer “white glove” treated files seem to have been run through some kind of compression, enhancing the more subtle sounds, to make them sound clearer. However, doing this wouldn’t require the whole MQA system. I’m sure there’s a plug-in somewhere that does this to music which anyone could use.

The 2L recordings are interesting, as it is physically impossible for them to have more than one unfold. This can be clearly shown by doing a spectral analysis of the tracks, which show the high level of high-frequency noise from the ADC, which is clearly and simply filtered out by whatever process MQA uses. An analysis of the unfolded MQA shows that the high frequency noise has been removed.

That being said, any “unfold” beyond the first, is only up-sampling using a very short digital filter. I have seen absolutely zero proof of any actual digital origami beyond the first. Even the early diagrams from the MQA group suggested the requirement of a 32-bit digital file for beyond the first unfold, which would make the file gigantic, the complete opposite of the original intent.

6 Likes

And yet it’s done…
Your other observations on the sound are just that, personal observation and opinion. Never, of course, the result of the technically achieved, de blurring :joy:

MQA has never actually specifically said what deblurring is or does. It’s a marketing term, and has no actual recognized technical meaning.
Measurements of MQA show that it adds distortion and abnormalities to the audio stream-it doesn’t remove them.
If you like how it sounds, no problem. Just don’t promote falsehoods and Orwellian double speak.

5 Likes

Peter-
Please stop using the misleading term “third unfold”. There is no such thing. It’s a misdirection to try to make people think there’s some special third level off decompression to higher res going on. There isn’t.

MQA files are compressed. There is one unfold. Everything after that isn’t an unfold or decompression, it is simply upsampling via the MQA filters. Doing upusampling in two stages to higher rates doesn’t “unfold” anything and doesn’t create a non-existent “third unfold”. It’s upsampling.

5 Likes

I could care less, but I did some looking around… the third unfold is sourced from mqa uk…

How MQA works

I still like hqplayer’s DSD256 renderings better.

Of course it’s from MQA. All part of their misleading marketing of their product since day one.

3 Likes

MQA do in fact show the de blurring by demonstrating the almost elimination of pre and post ringing within the signal.
That’s what it does…

3 Likes

This is what it says it does, a “demonstration” is still wrapped up in patents and marketing obsfucation. Without open and clear documentation we’re left arguing over individual interpretations of MQA double speak. The best way to get my confidence would be publish the specifications and move the conversation out of the shadows.

1 Like

They are not wrapped up in patents and marketing obfuscation. They are clearly explained in several white papers that have been cited here over and over. If you look at the published impulse responses of the filters, you will understand what is being said and what is meant by deblurring. However, ask yourself whether you would include that kind of signal processing in discussions on a consumer-oriented website like MQA’s blog, where 99.99% of consumers woud have no idea what you’re talking about.

1 Like

Show me an end to end digital comparison of the data vs PCM, following a complete chain of MQAs DSP, all unfolds, however many we think there are now. That’s one of the few things I thought was clear :wink:

Something I have saying for a long while but the marketing blurb is just blinding the truth

Well, you’re asking something that no algorithm designer would ever do. Just publish every detail, show measured end-to-end design data, and give the whole thing away free to competitors? The level of secrecy in most algorithms is far greater than with MQA.

“Unfolds” are their choice of wording to describe successive stages of signal processing. Since each is based on data included in the encoding and signal bits, it appears reasonable to name them successive unfolds. A lot of the internet argument calling ‘unfolds’ and ‘deblurring’ lying or deceptive marketing is more about just refusing to accept MQA’s choice of wording.

1 Like

The evidence is in the Analog resolution compared to the Analog input. That’s where your comparisons should be made. All the digital stuff in the middle is not relevant. Then I hear many musicians who know the Analog input, it’s their creation (art) who say MQA delivers their intent. That’s makes me happy with the system, a system that works…

1 Like

O yes it is…! , it’s not the first time a high-end audio company understand the problem on a higher level than others. 22 years ago HDCD was introduced and there are many similarities with MQA.

Both Keith O. Johnson and Bob Stuart understand the problem of digital filters and how to deal with it. The result is a serious and audible improvement.

Upsampling is part of unfolding, but you certainly need 3 folds and unfolds to reproduce a frequency spectrum up to 100 kHz in the analogue domain.

1 Like