Is it time-up for MQA

[Clive197]

It's been reported on a HiFi+ Blog that MQA is badly flawed and is not a lossless format at all. Tests have been done by HiRez Audio that is said to prove that MQA is only 17bit and not 24bit. That puts the cat amongst the pigeons.

[walpurgis]

I wondered where MQA was going. Things seem to have been a bit quiet. Mind you, I've not been actively chasing info.

[SLS]

Apparently Universal and Warner signed up recently. But then Universal specialise in low-res music, so they would probably be delighted with this HiFi+ Blog report.

[SLS]

I've just read the blog.

http://www.hifiplus.com/articles/highresaudio-to-stop-offering-mqa/

The bloody details are here
https://www.xivero.com/downloads/MQA-Technical_Analysis-Hypotheses-Paper.pdf

Xivero are the designers of leading music analysis software such as Musicscope, they know what they are talking about.

It's actually from the Managing Director of Highresaudio. HRA are a superb business and I buy their downloads - the best quality music I own and they really do check everything. I once complained about a recording of Shostakovich Sy. 2, there was actually nothing wrong other than a greater dynamic range than I had ever had on a recording.

So MQA may be emperor's new clothes after all. What a surprise.

[Rothchild]

MQA was never pitched as lossless, was it?

My understanding is that it uses lossy data compression on the pointless* part of the spectrum in a hi-rez recording (ie the bit much above 20kHz). It's kind of like FLAC for 20Hz to 20kHz and MP3 for 20kHz to 48kHz (or whereever it goes up to).

Not that I care one jot, nasty proprietary junk (with a closed specification) for the benefit of commercial streamers and the like. Not created in the interest of music fans and too far behind the curve to hit home; bandwidth and storage just get cheaper, so the requirement to 'save' them reduces at the same rate. Expecting and using open formats will do us all favours in the long run.



*subject to endless circular argument on the internet

[Macca]

What's you opinion on the claim that MQA also fixes 'timing errors'? I mean you can guess mine, but you know more about it than I do.

If it does do something like this then that does give it a USP ordinary 'hi res' doesn't have.

[Manicatel]

I've played with MQA vs cd rips a fair bit.
Using Tidal, Bluesound etc.
On some tracks, I hear very little difference. On other tracks I get more of a step up. On others, I hear a difference without being able to decide whether it's better, or merely different.
I'm still a bit unsure as to whether it's just a re-mastering effect on some older material that was recorded all analogue anyway, & my tech knowledge is nowhere near deep enough to truly understand what's genuinely going on.
However, if you already subscribe to Tidals Hifi level, it's a no cost option, so have a play with it.....what's to lose?

[coolblue]

First MQA encoded CD , details are here : http://www.stereophile.com/content/mqa-encoded-cds-yes#kgkwcvEvTeZUTmb3.97

[sq225917]

Mqa specifically used the term lossless in all their promo, though clearly admit to attempting to spin out a new meaning, as in audibly lossless, as opposed to data lossless. Of course they've never presented blind A/B to back it up.

I have no interest in streaming from the web so mqa has no appeal to me. The timing claims are easily disprovable rubbish.

[StanleyB]

First MQA encoded CD , details are here : http://www.stereophile.com/content/mqa-encoded-cds-yes#kgkwcvEvTeZUTmb3.97
If MQA is supposed to be for streaming, what is its purpose as far CD reproduction is concerned?

[Simon_LDT]

If MQA is supposed to be for streaming, what is its purpose as far CD reproduction is concerned?

Sounds like they're trying to do some sort of newer 'sacd' format whereas it will play in all CD players as standard redbook, but then hook it up to an MQA capable DAC and you'll get a higher resolution and all other gobbledygook.

Really hope this doesn't become the standard. We already have 16/44 redbook which is perfectly fine and for lossless & higher-res we already have Flac which has been around for many years and is fine. This all just smells of a money grab to me, they'd make a killing just on licensing costs around the world. Can never trust the big labels to give a damn other than to line their own pockets.

[Dynamics]

It's not worth it for me, nor hi res really as there isn't much out there in decent hi bit rate formats for new music anyway. Formats are rubbish against the bits that do the sound in speakers and amps, and frankly these can yield much bigger benefits than worrying if you have or haven't got the latest kit to decode mqa or dacs to do it etc.

[Yomanze]

You'd think that proprietary digital formats would've gotten the hint by now. SACD did. HDCD did. DVD-A did. MQA will fail simply because it'll never become a defacto standard. Doubtful that DSD will either despite the massive push after SACD has failed. IMHO the major push will be towards 24/96 files, which in itself is contentious.

[Jimbo]

You'd think that proprietary digital formats would've gotten the hint by now. SACD did. HDCD did. DVD-A did. MQA will fail simply because it'll never become a defacto standard. Doubtful that DSD will either despite the massive push after SACD has failed. IMHO the major push will be towards 24/96 files, which in itself is contentious.
Contentious because a lot of 24/96 doesn't sound any better than 16 bit.

[jandl100]

Ha!
So Golden Ears* Justin user211 was right when he said he could hear no difference with MQA.
And all those folks who trotted out with 'major hike in sq" - it was just down to expectation bias.
Sweet. :D

--
* Actually, not really Golden Ears, just totally honest with himself. Not a guy who goes in for self-delusion

[Firebottle]

Doesn't the technical spec actually show that it degenerates to an equivalent of 17 bits due to using the lower bits to 'fold in' the extra bit depth?

Sweet indeed Jerry :D

[Macca]

That won't make a blind bit of difference to the sound though. The only USP I can see is the so called 'digital deblurring' which would seem to be some form of eq. People reporting tighter bass and a cleaner sound overall, but only on some recordings. I'm guessing this is similar to what they do to old music when used on film soundtracks, just dial in a bit of extra punch. Not very purist of course.

[Dynamics]

As regards to mqa I don't think it will take off in current climate as the music industry has been on a decline for some time with the demise of good cd sales from about 2000 and Internet xoming in from around 2000, rise of file sharing and free steaming music. It's hard enough for streaming in conventional mass market formats let alone in mqa. Also most streaming firms operate at a loss so willingness to invest in new formats is doubtful. I hope mqa does though.

The thing that will take mqa and great formats get up to speed, is having a foundation of more people into streaming such there is more money in the industry to invest in niche formats and ensure these are distributed and available. I'd like to see a market where nearly everyone has a music streaming service. It would help all of us wanting these great hi res formats.

[jusbe]

Doesn't the technical spec actually show that it degenerates to an equivalent of 17 bits due to using the lower bits to 'fold in' the extra bit depth?

Sweet indeed Jerry :D

Yes. Here's Charles Hansen (Ayre Acoustics) thread (https://www.audioasylum.com/audio/digital/messages/18/184054.html) at 'the other place'. And, to quote:

<< >>If you play a Tidal Master and do not have software decoding enabled it is still delivering a high quality 24/48 stream.... which is better than CD quality.<<

Actually this is not the case. MQA is delivered in a 48/24 FLAC container. However the seven or eight Least Significant Bits (LSBs) of the container are used to store the "folded" dual-rate audio data (encoded losslessly) and the quad-rate audio data (encoded with lossy compression). They *replace* the low-level bits in the original 24-bit file. reducing the resolution of MQA below 24 bits.

If one starts with a full 192/24 true high-res source file, MQA encoding/decoding allows three possibilities:

1) Hardware decoding of MQA can perform both unfolds. The result is 192 kHz sampling rate but with a resolution of something less than 17.2 bits maximum. It appears that the number of lower bits required to store the dual- and quad-rate information is dependent on the signal levels in those upper bands. The MQS AES paper specifies that program material of a string quartet results in only 17.2 bits of resolution after the first unfolding (dual-rate source).

However according to James Boyk of Caltech, the string family (violin, viola, cello) has the lowest amount of signal power above 20 kHz than any other orchestral instrument besides an oboe. Boyk's article measures cymbals as having 30 dB (1000x) more power than a violin. Without further information it seems that the baseband audio resolution could be reduced as much as 6 bits fewer than the example of the string quartet.

2) If using a non-MQA DAC, the Tidal app allows for a single unfold via software, delivering up to 96 kHz sample rates, but again the resolution is still limited to a maximum of 17.2 bits (depending on the levels of high-frequency energy in the track).

3) If no decoding is used the 192/24 source file will be delivered with at a 48 kHz sampling rate, again with a maximum of 17.2 bits (depending on the levels of high-frequency energy in the track). For the published example of a string quartet, this potentially about 1 extra bit over Redbook. It is unknown at this time, but possible that music with high levels of high-frequency energy (eg, lots of cymbals, percussion, synthesizer, even brass or piano) would have *less* than 16 bits of resolution in the baseband - possibly even as low as 11 bits. >>

[sq225917]

Mqa only exists to give labels a chance to sell the same shit again, to sneak in DRM by the backdoor, to generate hardware and software sales due to tech churn, and to line mqa's patent pool pockets.