advertisement


MQA beginning to see the light?

The optimum action would be to be patient until there's more content out there and kit to check it with.

I've been on the sidelines of this business for 30 years now. Sometimes I crossed the lines. For instance, back in the old days when practically no-one had a clue about publishing on the internet I approached a certain Bob Stuart to help spread the word of ARA. (Remember ARA? No, not AR-A.) I've seen many trainwrecks from inside and outside. MQA as it was tauted originally was such a potential wreck. I write either from knowledge, or at least from what is highly plausible. No conjecture. I have no time for conjecture.

And when I write that I am concerned, you can bet there is a reason for concern. Under such circumstances sitting down and waiting patiently is not likely the right course of action.
 
Here's the exact quote from Bluesound/MQA:

"When MQA is previewed and approved in the studio, a ‘digital out decoder’ is one of those preview options. This preview process ensures all applications of MQA playback are fully approved and authenticated. MQA is ‘one’ music file and will be delivered to many hardware devices. This MQA Decoder “knows” its playback environment and unfolds the stream to the ability of the playback device. When decoding for a digital output, the stream is optimised for a generic 96kHz-capable DAC or a subsequent MQA decoder."

Can anyone "de-blur" that for me? Does it mean what comes out of the Bluesound is as good as it gets? Or does it mean that what comes out of the Bluesound is not as good as the full-fat MQA decode available only to those with Meridian DACs?
 
I've been on the sidelines of this business for 30 years now. Sometimes I crossed the lines. For instance, back in the old days when practically no-one had a clue about publishing on the internet I approached a certain Bob Stuart to help spread the word of ARA. (Remember ARA? No, not AR-A.) I've seen many trainwrecks from inside and outside. MQA as it was tauted originally was such a potential wreck. I write either from knowledge, or at least from what is highly plausible. No conjecture. I have no time for conjecture.

And when I write that I am concerned, you can bet there is a reason for concern. Under such circumstances sitting down and waiting patiently is not likely the right course of action.

If I may direct you to this thread again:
http://www.pinkfishmedia.net/forum/showthread.php?p=2833188#post2833188

Your comment was "fees will be punishing...."
Do you KNOW what the fees will be?
If not, then this is a classic example of the conjecture that you've been throwing around in these forums on the subject.

I've got to ask. Do you have commercial reasons for wanting MQA to fail?
Otherwise, I can't understand your level of criticality bearing in mind that we still do NOT know all the facts.
 
Mqa offers companies drm, it offers us nothing that doesn't exist already. Let them show high bit rates sound better, then let them show identical masters sound better as mqa, until that they have nothing to sell us.

People who won't pay for music won't pay for it no matter how you sugar coat it, to e they realised and moved on.
 
Afraid the above is quite misleading and obscures the problem.

It is a *format* which differs from LPCM in that some of the bits presented as if LPCM are there for other purposes and alter what would otherwise be LPCM info. Thus it will not play correctly *as* LPCM in an LPCM player that has no awareness of MQA.

Standard LPCM wave files and Audio CD data are defined to be LPCM. Not altered by MQA. Failing to make this distinction and understand its implications may lead us back into a mess of the kind HDCD often propagated in practice.

Maybe Bob Stuart's lying. However, I don't believe he has a track record of it, so when he has directly said MQA is not a file format, I'll go with that. I think this is an attempt to make things better for all and not to confuse people.
 
Mqa offers companies drm, it offers us nothing that doesn't exist already. Let them show high bit rates sound better, then let them show identical masters sound better as mqa, until that they have nothing to sell us.

People who won't pay for music won't pay for it no matter how you sugar coat it, to e they realised and moved on.

SQ>
Sorry I don't agree.
MQA "might" encourage mass availability of music approximating the quality of 24 bit. If it does that, I'll be a happy bunny, and that's very much where I'd like things to go.

I do have a reasonable level of HD music now, along with 50 or so SACDs and genuinely believe that they "can" add value.
 
When decoding for a digital output, the stream is optimised for a generic 96kHz-capable DAC or a subsequent MQA decoder."

Can anyone "de-blur" that for me? Does it mean what comes out of the Bluesound is as good as it gets?


Where did you get that quote? It adds a tessera to the mosaic, and ties in with my earlier 'conjecture'.


What it means? With a dual-rate (88.2 and 96) digital output you get some of the MQA benefit (the dual-rate-signal-in-a-single-rate-file), but not the full fat (the quad-rate-timing-in-a-single-rate-file). But that first step is of course much more important than the second. (The more so as it allows one to extract the dual-rate signal and then post-process it oneself to quad-rate.)
 
MQA "might" encourage mass availability of music approximating the quality

Yes. Just as HDCD for a while pushed production values. Just as perhaps DSD is doing today. Whatever the root cause, anything making people think twice and opt for the better sounding production is valuable. That doesn't mean the format itself is superior. It is a mind thing.

Maybe Bob Stuart's lying.

He is not. He is very careful with his wording. Would make an excellent PM. Fans read his words this way, and critics (or dare I say objective people) easily read them another way.


If I may direct you to this thread again:

Your comment was "fees will be punishing...."
Do you KNOW what the fees will be?
If not, then this is a classic example of the conjecture that you've been throwing around in these forums on the subject.

There is conjecture and there is conjecture. Meridian is owned by an investment group. These want to see dough. If you launch a new format and want it to be open source then you declare it open source. Instead we have had secrecy, hand waving, and the Auralic debacle. This makes it plausible that a strict licensing scheme was in the original plans. Maybe not as draconic as HDMI, but likely still bad enough to keep small companies and open source efforts (Volumio, Moode, Rune, piCorePlayer, ...) out of it. That would be bad.
But I admit there is a bit of guessing there. I am not privy to their business plans. Maybe Santa Claus is their chairman.

When it comes to signal theory, however, my 'guessing' is of an entire different order and you'd better well pay attention. Unless you can say yourself straight-faced that you can build an MQA-like scheme in less than a week. (Not that this is that much of an achievement.)

I've got to ask. Do you have commercial reasons for wanting MQA to fail?

Not in the least.

MQA failing massively would be a good outcome.

MQA being a total overnight success would be a good outcome.

MQA floundering a couple of years while fragmenting industry and confusing the few people interested in this even more would be a bad outcome. They were heading this way. The audio lemmings following their trail happily ...
 
Where did you get that quote? It adds a tessera to the mosaic, and ties in with my earlier 'conjecture'.

From post 14 in the computeraudiophile thread you gave a link to in your first post. Might be worth you engaging the author directly on whatever forum it was quoted from. If someone's giving answers there's less need for conjecture.
 
Werner, I have to take issue with the idea that Bob Stuart has "fans". Paul McCartney has fans, Johnny Depp has fans. Bob Stuart, not really. He's a respected engineer whose work prior to MQA has done much in the field of hi-fi. No, I don't think he's lying either and he's clearly got a commercial imperative to monetise his work. No problem there, it's how many things work. I think we need to give this time. Make no mistake, if the music companies want to imprint everything with DRM, they will and they've ****ed things up often enough that nothing would surprise me when it comes to confusing the market. Fingers crossed that this is a step in the right direction.
 
The issue here is not really about the potential benefits, as most of us here have never heard MQA working as it should and even those who have only a limited amount of content to check.

The optimum action would be to be patient until there's more content out there and kit to check it with.
The threads on the subject at the moment are frankly less helpful as they're full of conjecture, mostly negative.

That pretty much sums it up.

MQA seems to be a lot of hot air, one minute it's being punted as a streaming compression system, the next on something to do with timing.

One day it will launch and it will either be a huge success or a complete failure. The judgement will be purely commercial.

From what I hear, I favour the latter as I do not see what major problem it is trying to solve.
 
When it comes to signal theory, however, my 'guessing' is of an entire different order and you'd better well pay attention.

They say pride comes before a fall, and this statement oozes hubris.

There's no gentle way to say this, and I'd still be silent if you hadn't emerged with the above quote. I had generally thought you had a fair idea what you were talking about until I came to a rather different impression from posts 306 & 310 here:
http://www.pinkfishmedia.net/forum/showthread.php?t=170431&page=21#post_message_2833094

You describe your process as lossless and producing a 75Mb flac instead of the original 91Mb. You know what lossless means, so one might think you've got a valuable contribution to make to improving FLAC, except that you then go on to say what you've done...

And when you said "Add A and B. Enjoy" you stopped there and didn't go on to run a file compare did you? Because it didn't pass. Did it.

Sorry, but I'm not impressed that you appear to be vaguely aware of the outline of a quadrature mirror filter. Even if you'd got the details right, you seem unaware of its limitations or that your hf treatment destroys information.
 
They say pride comes before a fall, and this statement oozes hubris.

Thanks for the contribution.

And when you said "Add A and B. Enjoy" you stopped there and didn't go on to run a file compare did you? Because it didn't pass. Did it.

That was a proof of concept, done in minutes, just to show that the existing FLAC algorithm could be tricked into higher compression by spectrally moving part of the data around.

quadrature mirror filter. Even if you'd got the details right, you seem unaware of its limitations or that your hf treatment destroys information.

Of course special care is needed where the signal is split into the baseband and the upper band to be folded. I always wondered how MQA tackle this, whether they would attempt symmetry, or whether they would apodise the baseband (there are claims that undecoded MQA would be superior to CD) and perhaps sacrifice a bit of the band right above Fs/2.
 
Just so I can keep up is Ceanothus' objection that the process will not be lossless as there will be some imprecision (limit on the precision?) in the filtering? I assumed something would be lost in the redithering.
 
I think all the engineers here understand that MQA is a process plus a container format, so a lot more than a new file format

Hi, that may well be the case, especially as I'm not an engineer! I think I'll bow out as I'm not qualified to have engineery arguments. The final thing I would say is that MQA seems to be "a good thing". The ability to illegally listen to music already exists and the music industry won't be putting the genie back in the bottle. No matter how hard Hollywood tries, film copy protection usually lasts a short time before a work-around for Blu-Ray etc is figured out. Consequently the idea of something that sounds better even if played back on existing gear and even better with what BS described as a "white glove" scenario can't be a bad thing.

Back to the sidelines for me!
 
Here's the exact quote from Bluesound/MQA:

"When MQA is previewed and approved in the studio, a 'digital out decoder' is one of those preview options. This preview process ensures all applications of MQA playback are fully approved and authenticated. MQA is 'one' music file and will be delivered to many hardware devices. This MQA Decoder "knows" its playback environment and unfolds the stream to the ability of the playback device. When decoding for a digital output, the stream is optimised for a generic 96kHz-capable DAC or a subsequent MQA decoder."

Can anyone "de-blur" that for me? Does it mean what comes out of the Bluesound is as good as it gets? Or does it mean that what comes out of the Bluesound is not as good as the full-fat MQA decode available only to those with Meridian DACs?

It might mean that *studios* can have kit to convert back to plain LPCM having fully 'decoded' all MQA info. So it could be something provided for studios, but not for the consumers.

The problem here is a that a lot of what is quoted from MQA statements is ambiguous. The use of the term "format" is another example.

Wave LPCM is a defined format which dictates that the samples *are* plain LPCM. MQA hijacks some of those bits and uses them for conveying *non*-LPCM info. Thus a standard LPCM player that doesn't recognise MQA will treat the bits as being part of LPCM samples, altering the results.

From the POV of a real-world engineer, that means the file or stream is not wave LPCM format. It is something else.

What I can't tell is *why* these ambiguities seem to be appearing in MQA statements, answers, etc. They may be a lack of thought, or the result of someone having become so acclimiatised to what they are trying to do with MQA that they don't see the ambiguities. Or someone wanting to 'dumb down' explanations/answers because they don't want to befuddle non-technical readers. Or some other reason.

However - unlike MQA - anyone who wishes could start using the open alternative *now*. The programs now exist and are in the public domain to generate the files and play them. At the moment you'd need some computing 'nous'. But it is also open for someone to make desktop file generators and players, etc.

The point of this isn't necessarily that everyone will rush out and do so. It is that as and when MQA appears people will know the alternative exists and is available for adoption. And that it has already shown similar levels of file size compression, etc.

The only thing it lacks is the mechanicsms for covering the details in secret sauce and controlling access in exchange for added payments to IPR owners of the methods used. What it adds is the ability for anyone interested to completely decode and check the results and develop *improvements* as and when they wish.

FWIW as I've said before, I have a high regard for the audio skills, etc, of Bob Stuart and Peter Craven. They have an impressive track record. So none of my concerns have been about the sound quality potential of MQA. Indeed, having spent time experimenting with an alternative has let me agree that it should be quite feasible to reduce the 'high rez' file sizes by amounts akin to what MQA seems to be offerring. My interests are in people have a free and informed choice and being able to use and check the results and improve them *without* having secret sauce get in the way. Up to each person then to decide if that attracts them or not. Your choice, not mine. :)
 
Of course special care is needed where the signal is split into the baseband and the upper band to be folded. I always wondered how MQA tackle this, whether they would attempt symmetry, or whether they would apodise the baseband (there are claims that undecoded MQA would be superior to CD) and perhaps sacrifice a bit of the band right above Fs/2.

The problem of both keeping the LF section 'unchanged' for LF-only replay *and* perfectly splitting and recombining the HF info for 'high quality' was what made me feel that simply emphasising the whole thing made more sense. Much easier to 'correct' and ends up providing the full range anyway for files and streaming.

Of course, this wouldn't do for stuffing the 'enhancements' onto an audio CD. But personally given the shambles that HDCD often became, I really *hope* we *don't* get MQA'd Audio CDs *unless* the magic addons are shoved into a distinct 'partion' of the disc a la non Red Book discs. Even then, all too likely that the MQA process will alter what would otherwise have ended up being played by a standard LPCM player/DAC. So in my view, a bad idea. Another example of vague ambiguious semantics like calling HDCD 'compatable'. :-/
 
That was a proof of concept, done in minutes, just to show that the existing FLAC algorithm could be tricked into higher compression by spectrally moving part of the data around.

Who cares how long it took? You claimed "a process that is totally reversible, i.e. lossless". It's nothing of the sort, and you show no sign of understanding why not.

If, when it came to signal processing, your guessing really was of a different order then you'd know you removed information in your ultrasonic quantisation and, faced with less information in the audio, FLAC produced a smaller file. Nothing to do with spectrally moving data around. And you'd also know that even if you hadn't removed that information, putting your two halves back together wouldn't get exactly what you started with.

Is this an interesting result? I think not - you could have completely thrown the ultrasonic file away and got yourself down to 52Mb. Or used MP3 on the original and got yourself something far smaller. It would have been interesting if it really was lossless, as you falsely claimed. And very surprising as FLAC does do a good job - to "trick" it you'd need to discover something genuinely novel about redundancy in audio.

You're clearly not an authority on signal processing. Don't claim to be one.
 
MQA is not lossless.
The FLAC and WavPack experiments we have done are bit perfect to 20kHz to the 16 bit level and 20 bit almost up to 10kHz.
Resolution goes down towards 48kHz just like MQA does
 


advertisement


Back
Top