advertisement


MQA beginning to see the light?

That's why I've taken to useing the term 'hints'. :)

Again, you can get similar results by the simpler method we've experimented with - simply emphasising the HF.

The big difference tends to be that commercial systems tend to get promoted and advertised. Whereas open and free methods work on the basis of people noticing them or having someone they know mention them. But sometimes the tortoise wins... 8-]
 
Jim Audiomisc said:
At present I don't think I could give a definitive answer to your question, Tony. The statements and 'answers' I've read are too ambiguous and before-the-fact. However it seems likely that some companies would be licensed to sell closed-source replay software. The uncertainy being if that would then try to block fully decoded LPCM being sent out of a computer to, say, a non-MQA DAC. And if it did try to control something like that, how successful it might be. per the 'BS' example, what happens might vary, and possibly sometimes occur because it 'slipped past' those controlling MQA.
Click to expand...

Don't think it has anything to do with it "slipping past" those controlling MQA.

See TonyW's posts on this thread:

https://helpdesk.bluesound.com/disc...sid=085586702218ab5d17679fd2b34b8d12&start=20
 
Looking at that got me:

"If you bypass our CL (Gen1) or Burr Brown (Gen2) DAC and select Optical or COAX out, we decode and the file and provide the content but the 3rd party DAC unfolds it to the best of it's ability. If that DAC is not MQA certified, it will be limited to how much it can open the file and will play at it's best. That's the beauty of MQA, if you do not have MQA certified equipment, you can still at least play the file - just not to the best of it's ability. It would be like watching a BluRay disc in a DVD Player and it still work... just not at 1080p... :D (yeah yeah, 4K I get it - I digress)."

Which seems to me to say the digital output fed to a non-MQA device *doesn't* have *all* MQA processing and 'enhancing' carried out to produce an entirely LPCM stream.

The fact that the output may be higher rate doesn't necessarily mean anything much beyond x2 upsampling was done on the 'LF' part. Alebit the upsampling might have been done with a fancy reconstruction filter to make it sound 'better'.

There is also, of course, the problem that a non-MQA DAC couldn't have the stream optimised for its reconstruction filter since spdif is a one-way transfer. The source has no idea what the device on the other end of the cable is.

The above is potentially relevant for a reason I'll add in another posting.
 
Thinking about the 'filter problem' it occurred to me that what might be going on is that the MQA process on a 48k/24 source is to upsample the LF to approximately get the intersamples. Still all LF. Then use the lsb as a series of 'addon values' to nudge onto the LF upsampled ones the pattern of the HF which was found to be missing during MQA encoding.

i.e. The encoding takes a 96k source and downsamples it to 48k. It then upsamples that again and compares it, sample-by-sample, with the original. Measures the difference sample-by-sample. Then stores those values in the LSB (albeit with reduced resolution).

This avoids the filter problem and is easy to undo. But again, isn't rocket science. So it is a trick anyone could apply in an open alternative if they fancy.

The snag is that the resulting HF in the lsb will seem like noise to a flac encoder. So compress poorly. Leaving the HF, freezing noise, and keeping a 96k file for flaccing may well work better if filesize reduction is the aim.
 
I read that as similar to the DVD situation with a limited spdif. The interesting thing is that the internal dac has an exposed I2S bus. BluRay stops these vulnerable busses
 
Jim Audiomisc said:
Looking at that got me:

"If you bypass our CL (Gen1) or Burr Brown (Gen2) DAC and select Optical or COAX out, we decode and the file and provide the content but the 3rd party DAC unfolds it to the best of it's ability. If that DAC is not MQA certified, it will be limited to how much it can open the file and will play at it's best. That's the beauty of MQA, if you do not have MQA certified equipment, you can still at least play the file - just not to the best of it's ability. It would be like watching a BluRay disc in a DVD Player and it still work... just not at 1080p... :D (yeah yeah, 4K I get it - I digress)."

Which seems to me to say the digital output fed to a non-MQA device *doesn't* have *all* MQA processing and 'enhancing' carried out to produce an entirely LPCM stream.

The fact that the output may be higher rate doesn't necessarily mean anything much beyond x2 upsampling was done on the 'LF' part. Alebit the upsampling might have been done with a fancy reconstruction filter to make it sound 'better'.

There is also, of course, the problem that a non-MQA DAC couldn't have the stream optimised for its reconstruction filter since spdif is a one-way transfer. The source has no idea what the device on the other end of the cable is.

The above is potentially relevant for a reason I'll add in another posting.
Click to expand...
Having read that post it seems to me that it could mean almost anything.The "third party dac unfolds it" would seem to mean that the origami is not unfolded by Bluesound. Heaven only knows what they are passing out, perhaps raw MQA? Can they really mean that all that is missing is the revolutionary anti-imaging filter?
 
adamdea said:
Heaven only knows what they are passing out, perhaps raw MQA? Can they really mean that all that is missing is the revolutionary anti-imaging filter?
Click to expand...

Don't worry too much.

From the second link in the first post in this thread:

"In this scenario, the MQA software within the Node 2 will unpack any hi-res content and send the entire, unfolded file to the external D/A converter over S/PDIF. How much hi-res unfolding takes place will depend on the maximum sample rate handling of the Node 2’s coaxial and Toslink outputs.

Now comes a slight catch: the MQA software inside the Node 2 isn’t ‘externally DAC aware’; it knows not which DAC has been connected to its rear panel and therefore cannot apply its custom, pre-corrective filter. "

You get the origami unfolding to dual-rate. You don't get anything beyond, i.e. the MQA-style partially DAC-specific narrow-filter oversampling to whatever rate the MQA boffins found necessary for a particular DAC. In the SPDIF output case this oversampling is entirely left to that DAC.

There is more detail at the end of this page, https://helpdesk.bluesound.com/discussions/viewtopic.php?f=32&t=3021&start=20, although as usual with this sort of communications between customers, engineers, and mgmt types entropy quickly sneaks in.
 
It isn't clear to me if the x2 rate conversion actually adds back any detail above that which was in the LF. Could be a simple upsample using a standard filter shape. Depends on what isn't stated about 'generic'.

BTW I've now finally started reading the basic MQA WO Patent doc in detail. Until now I've avoid this as I wanted to approach doing an 'open' method with a clean and clear eye. The patent doc is also knee deep in over-complex statements that say little and claim all. But of course this is the nature of modern patents. :)

Near the start, though, it does introduce some details which do strike me as odd if the aim is *fidelity*. Has anyone else yet had a go at making sense of the patent? The number here is

WO2014108677A1

The GB patent I have os GB2503110A

And if anyone knows of later/other patents on this, please let me know. I'm now into a 'reading' phase as I start to put together a write-up.

I decided I needed to start a full investigation article of this by beginning with the basic 'requirements' which audio fans are assumed to wish tackled. e.g. getting a quart from a pint pot and not being happy with boring old lossless compression. Then considering the pros and cons of various approachs as well as looking at the potential 'marketing, licensing, control, and secrecy' aspects.

It'll no doubt take a a while. :)
 
Werner said:
Don't worry too much.

From the second link in the first post in this thread:

"In this scenario, the MQA software within the Node 2 will unpack any hi-res content and send the entire, unfolded file to the external D/A converter over S/PDIF. How much hi-res unfolding takes place will depend on the maximum sample rate handling of the Node 2’s coaxial and Toslink outputs.

Now comes a slight catch: the MQA software inside the Node 2 isn’t ‘externally DAC aware’; it knows not which DAC has been connected to its rear panel and therefore cannot apply its custom, pre-corrective filter. "

You get the origami unfolding to dual-rate. You don't get anything beyond, i.e. the MQA-style partially DAC-specific narrow-filter oversampling to whatever rate the MQA boffins found necessary for a particular DAC. In the SPDIF output case this oversampling is entirely left to that DAC.

There is more detail at the end of this page, https://helpdesk.bluesound.com/discussions/viewtopic.php?f=32&t=3021&start=20, although as usual with this sort of communications between customers, engineers, and mgmt types entropy quickly sneaks in.
Click to expand...
Hmmm if the remaining stage is just an anti imaging filter we're home and dry. It would be great to be able to compare the output with the ordinary 24/96 or 24/88 (whatever happened to 24/88 one hardly seems to see it these days?). Or does the unfolded top octave require decoding in some way? What does "pre-corrective" mean?

Perhaps i'm over-thinking this but the initial quote seems to be from John Darko himself and therefore a bit difficult to take seriously.
 
I recall some earlier comments in one of the MQA threads saying that the poster thought that the MQA'd and non-MQA'd versions had come from different source versions. Reading the MQA Patent makes me wonder if this was actually down to changes made by MQA. It may come down to just how they process the 'split' etc.
 
radamel said:
I'll try and listen to a Node 2 feeding a very good non MQA DAC.
Click to expand...
If you have acces to a node 2 with a digital output it would be great if you could capture the output so that we could see what the upper octave looked like (image of 0-24khz, upsampled low rez, undecoded mqa noise or hf signal)
 
I'd be interested in anyone looking for aliasing and a roll off in the 'encoded' version that the 'decoding' tries to correct. I may get the MQA versions of some of the 2L files myself at some point and look for this. So far, my impression from the WO patent is that MQA deliberately folds back aliasing and then tries to unscramble the eggs at the other end. In effect, seems to use 'lazy' filters.

Once I can get a handle on the details I may try writing a demo prog that lets people apply similar low order filters and generate similar 'effects' so they can give a listen to a before-after for such processes and decide if they like the 'MQA effect'. :)

If the encoding *does* add a lot of aliases *and* roll down HF it might be no surprise if - listened to as LPCM - the results sound different and may seem to have come from a 'different version' at source! Lacking the details of the actual MQA filters it might well lead to someone preferring one version over another because of things like the change in responses. Or the addition of a 'spray' of aliasing.

I'll keep reading... No doubt there is more to it. :)
 
I don't see how 44.1 can be intelligently expanded to 88.2 and then further expanded with a second round again using the MQA data.
I am thinking that my theory of taking alternate samples as the lpcm part and then calculating the missing samples and encoding the difference between this calculated value and the actual value would fit a lot of this description. The aliasing in the lpcm part would be largely inaudible unless the master has unusually high levels >22k
 
Jim,
I know you're a Dr and an ex electronics engineer and lecturer. From some of your other writings over the years, you clearly have a LOT of knowledge.

So just what is going on here. What I'm seeing is:
- Virtually non-stop posting in a forum that you've just just 3 months ago, mostly to criticize something that is as stated "NOT finished"
- You weren't happy with a single thread, and are now re-iterating the same glib remarks over multiple. e.g. "lazy" filters. You're a well known engineer and are using throwaway terminology suitable for a politician
- In an earlier post (84), you happily make a comment that MQA software decoded and output to an "unknown" DAC wouldn't be taking full advantage of MQA. Why is that news? Of course it won't and anyone reading this troll level thread would know that
- Then in this thread, you make a point that the MQA is changing the source file. Of course it is. MQA state quite happily that they're trying to account for the sonic elements of the ADC and DAC. So WHY is this news and why try to paint it as a "bad" thing.

Seriously, this is some of the worst trolling I've ever seen of an audio product.
It's just a listing of snide remarks and criticisms, some hidden some not.
If we KNEW the full story, I could accept that if it were all true and relevant. However, we don't yet have it, so WHY are you banging on so hard about it?

I'm going to ask this VERY seriously. Please, state categorically that you are NOT being paid to make so much noise in this forum against MQA.
 
My suspicion is that MQA is degrading the source file by raising the noise level and folding in aliasing. This would be reversed by playing through a full MQA chain. As the industry is proposing to use the MQA version for all 44.1 and more, this would be a bad thing.
My only financial interest is that I can continue to design dacs.
Are you a MQA front?
 
David,

Happy to answer that in full:
- I am a Meridian user
- I have NO financial ties to Meridian, MQA or Roon and certainly don't receive any payment of any type from them. For ref, I'm an employed IT person

My only active interest in MQA is the "potential" which I believe it has for facilitating the availability of a much wider catalog of higher quality audio via streaming.
Again, read the term "potential". I've never heard it and have no intention of making up my own mind until I've done enough comparisons myself to make that decision.

My point here is the apparent level of destructive criticism of a potential solution prior to us having even the full story.
As I've repeatedly stated previously, if MQA doesn't deliver sonically, then it deserves to die an utter death. However, I don't want the audio fraternity to ignore it because of a small number of individuals taking what appear to be cheap shots at it, often just re-iterating stuff we already know, but attempting to do so in a style just a little too reminiscent of the Daily Mail.
 
Well, nature abhors a vacuum. If Meridian were more open about how MQA works, there'd be less need for people to speculate.
 
Mr_Sukebe said:
sn't deliver sonically, then it deserves to die an utter death. However, I don't want the audio fraternity to ignore it because of a small number of individuals taking what appear to be cheap shots at it, often just re-iterating stuff we already know, but attempting to do so in a style just a little too reminiscent of the Daily Mail.
Click to expand...

For heaven's sake, there is an entire industry printing uncritical regurgitation of marketing copy, and you're worried that it will fail because of a few people printing genuinely questioning copy about how the benefits of product could be made open source (and a few snipey comments for fun)?


The Daily mail is more or less devoted to the pursuit of pseudo-scientific drivel in para-oncology, crime statistic analysis and of course its championing of the MMR nonsense. Hardly reminiscent of crazy people who wonder what could be going on with a closed system for specially treating the bit of a signal you can't hear.
 
This unnecessary MQA crap makes the Daily Mail a worthwhile read by comparison. Can't you put energy into failed SACD revivals or something? I know, let's call it Super MP3...four out of five members of the public can tell the difference, honest.
 
You must log in or register to reply here.


advertisement


Back
Top