Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Meridian Audio's new... sub-format called MQA. (Read 145496 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

 

Re: Meridian Audio's new... sub-format called MQA.

Reply #227
Why is this even being entertained?

Because of the claim that it sounds different than the source?
Here:
True.  But if they're claiming that MQA sound better than FLAC, then we have some DSP going on also.


Not just different.  They're claiming it's better than the source.

Re: Meridian Audio's new... sub-format called MQA.

Reply #228
...but if the source used a brick wall (sinc) filter, and if the process removes the ringing from this filter without adding any more, then it's not factually incorrect to claim an improvement.

With 192kHz sampling, it's several miles beyond the limit of credibility to claim this is an audible improvement, unless someone can prove different. But in signal processing terms, if you define ringing as a bad thing to have, then by that definition this is an improvement. (Reality check: there's usually little or no original signal content up there to excite the ringing.)

I'm still getting over the idea that ringing at 22kHz has reportedly been ABXed. If anyone is going to repeat the trick at 96kHz (i.e. the cut off frequency for 192kHz sampling) then I shall probably need therapy :)

EDIT: I take it as read that all claims of "night and day" differences can be dismissed out of hand. I automatically ignore these. Even in the world of true "professionals", 90%+ of it is just imagined. I am open to the possibility that some people have concentrated on certain specific issues so long that they really do hear them as night and day differences, while the rest of us can barely hear them at all. You can see that some of us do that right here on HA with non-transparent (but really rather good) audio coding: some of us ABX it readily enough, but grade it as 4.9; others hear the same thing and give it a 3.0 because it's their life. If some people really can hear different anti-alias filters, it's got to be in this category, but 1000x more subtle.

Cheers,
David.


Re: Meridian Audio's new... sub-format called MQA.

Reply #230
They're claiming it's better than the source.
...and that's supposed to justify the following nonsense:
Quote
if they're claiming that MQA sound better than FLAC, then we have some DSP going on also.
??

Please understand that I will expect you to provide some substantiation should you tell me the above quote isn't anything more than an attempt at an educated guess.

Shy of substantiation, my response to this...
733 date=1453234747]If this is more than just placebophile BS (meaning you can actually ABX the difference),, then there has got to be some kind of EQ or other adjustment going on there by the decoder.
For the sake of argument, let's say a difference is ABX-able, that certainly does not mean there is some kind of manipulation outside of the technology as it is currently understood by *knowledgeable* third parties.

Re: Meridian Audio's new... sub-format called MQA.

Reply #231
We already know that MQA cannot be lossless despite their claims
I think that Meridian want MQA to be considered as an integral part of the mastering process.

So after mastering with MQA  ‘encapsulation’ (avoiding the term compression and the inherent question of loss), the distribution and playback process is lossless.

Of course, you could say the same about any other lossy codec employed at the end of the mastering chain.

Re: Meridian Audio's new... sub-format called MQA.

Reply #232
@2Bdecided

Is there really even vaguely compelling evidence for the audibility of ringing with 44.1ksps sampling? I mean, call me stubborn but it seems like the Stuart paper (and I think there's one other recent AES candidate?) isn't really much to support what would be an astonishing capacity for detecting heavily masked, low level ultrasonic content. I mean, I can cite more vigorously controlled papers that 'prove' ESP...

Re: Meridian Audio's new... sub-format called MQA.

Reply #233
@bandpass well yeah, but no
MQA is as much 'encapsulation' instead of compression as is lossyWAV ... it's compression, and lossy.
24/192 PCM input => MQA => output does not equal input. That's lossy.
"I hear it when I see it."

Re: Meridian Audio's new... sub-format called MQA.

Reply #234
@2Bdecided
Is there really even vaguely compelling evidence for the audibility of ringing with 44.1ksps sampling?
Not if done right according to Meridians owners manual.
But because it can be done wrong, we now have an elixir, MQA. Now everything needs to be "cured".
yay

cheers,

AJ
Loudspeaker manufacturer

Re: Meridian Audio's new... sub-format called MQA.

Reply #235
@2Bdecided

Is there really even vaguely compelling evidence for the audibility of ringing with 44.1ksps sampling? I mean, call me stubborn but it seems like the Stuart paper (and I think there's one other recent AES candidate?) isn't really much to support what would be an astonishing capacity for detecting heavily masked, low level ultrasonic content.
For me, it needs to be both independently repeated and understood. Saying "this is why it's detectable" is quite different from proving that is why it's detectable (if it is).

Remember the results were only 5% better than chance - i.e. one time out of 20 someone did better than guessing when trying to detect the ringing. There were enough trials to make this statistically significant, but it's a tiny effect.

I can't hear it, and I've tried. IIRC there were a couple of positive ABX results in a thread looking at the audibility of 20kHz ringing that I started years ago. Even in my own extreme samples, I could hear nothing wrong. https://hydrogenaud.io/index.php/topic,68524.0.html (the last positive ABX result is facecious ;) )

Cheers,
David.

Re: Meridian Audio's new... sub-format called MQA.

Reply #236
I still hesitate to accept they heard the filter. They may have heard simply a piece of the used hardware acting different without HF content. My bet is on the tweeter.

I remember a test Archimago created 176.4kHz music files with very strong ringing settings for linear and mimimum phase from 44.1kHz sources.
http://archimago.blogspot.ca/2015/04/internet-blind-test-linear-vs-minimum.html
A CA member that has for sure a minimum of 100 posts about the working and finetuning of upsampling filters only tried to find excuses why he can't hear it, no usable result.
The fear of ringing is really strong and works best sighted. For MQA a lightbulb indicator alone surely will do wonders.
In several forums there also seem to be a growing interest in doing custom filters for DACs.
Many claim all kind of things with these also.
So MQA is good at satisfying a trend.

I wonder if really something new could have been a 16bit/64kHz srongly noise-shaped format that in the end is smaller compressed as MQA.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Re: Meridian Audio's new... sub-format called MQA.

Reply #237

I wonder if really something new could have been a 16bit/64kHz srongly noise-shaped format that in the end is smaller compressed as MQA.
I agree. It might not even have needed to be 16/64, 14/64 maybe with a very modest increase in file size over 16/44

Re: Meridian Audio's new... sub-format called MQA.

Reply #238
@2Bdecided

Thanks for the links. The test you conducted was particularly interesting: I'd expect a number of people to be able to hear it (ringing a mere 10dB down at a relatively low frequency). I'd put some of that down to people actually hearing the ringing, and some of it down to the surprisingly poor distortion performance exhibited by a lot of integrated audio solutions at high frequencies. Hell, I was looking at TomsHardware, and almost all the audio solutions they tested - and these were on expensive boards - had high frequency IMD (RMA, not sure of the exact test protocol) at about 45-50dB down.

And of course we have Meridian's...study. Their literature (looking at their other papers too) is a mess - their citations are frequently FUD. They make a lot of perfectly reasonable claims about the ear's temporal resolution (albeit poorly cited - I particularly enjoyed when they cited, in one paper, the entire Oxford Handbook of Psychoacoustics with no fucking page references), and then stitch that to their "temporal smearing" bullshit via such luminaries as Kunchur, of "I don't know how sampling theorem works" fame. Taken as a whole, their arguments are somewhere between incompetent and wilfully misleading.

I suspect I'm preaching to the choir here. I feel like one shouldn't be too open minded when it comes to stuff like this. The ability to make a vaguely plausible argument for the super-marginal audibility of ringing at 44ksps completely disappears once you push the sample rate just a little higher...and seeing as an aggressively noise-shaped 16/88.2ksps file is about the same size as their stupid proprietary format, even when one is scrupulously open-minded I can't help but see this as really, really crappy.

Once you get past the argument that conceivably justifies slightly higher sampling rates so we can shift the filters well out of the audible range (which isn't one they're actually making), to justify the rest of it (MQA: better than even 32/350ksps+ PCM!) they rely on an outrageous collection of innuendos, insinuations and half-truths.

Re: Meridian Audio's new... sub-format called MQA.

Reply #239
One question to the experts when we are at it. This apodizing seems to work linear phase up to a frequency and minimum phase for higher frequencies up from there.
Lets assume the linear part ends at 12kHz. Isn't the pre-ringing of that part of the filter not at 12kHz? So you may have less ringing but much nearer or inside the audible band?
I remember a spectral pic of iZotopes intermediate phase filter with an impulse exactly showing this.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Re: Meridian Audio's new... sub-format called MQA.

Reply #240
Mixing linear and min phase is not different from just filtering an initially linear phase filtered file again with a min phase filter or vice versa.
The linear phase filter will just ring at its frequency. If the min phase filter used has some attenuation at that frequency then the "previous" ringing will also be attenuated accordingly. Of course the min phase filter will also ring.
"I hear it when I see it."

Re: Meridian Audio's new... sub-format called MQA.

Reply #241
Thanks! So using changing phase behaviours in one process can lead to other problems but may produce nicer pics in some areas you want to look better on funny graphs.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Re: Meridian Audio's new... sub-format called MQA.

Reply #242
There's always a tradeoff. With an added min phase filter it's phase shift and/or roll-off.

"I hear it when I see it."

Re: Meridian Audio's new... sub-format called MQA.

Reply #243
Got one, thanks! Above iZotope pre-ringing 50 that must be linear to ~15kHz and below SoX b92 -a


Edit: Somehow the embedding does not always work. Direct link:
http://www.directupload.net/file/d/4249/fa9ez2tl_png.htm
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Re: Meridian Audio's new... sub-format called MQA.

Reply #244
There's always a tradeoff. With an added min phase filter it's phase shift and/or roll-off.

But the official story is that phase shift is inaudible, even at loudspeaker crossover frequencies. Which is why no one bothers about it. Verging on the ultrasonic, surely it must be of academic interest only..?

Re: Meridian Audio's new... sub-format called MQA.

Reply #245
But the official story is that phase shift is inaudible, even at loudspeaker crossover frequencies. Which is why no one bothers about it.
It is best to read more, post less (nonsense).
Quote
1) Even quite small midrange phase nonlinearities can be audible on suitably chosen signals.
I believe JJ once mentioned "percussive" sounds, maybe acoustic guitar, but I'm too lazy to search and that's your job anyway.

And now back to the miracles of MQA...


cheers

AJ
Loudspeaker manufacturer

Re: Meridian Audio's new... sub-format called MQA.

Reply #246
Charming, as ever.

Possibly cut-and-pasting from the same web site you looked at:

Dr Floyd Toole:

" It turns out that, within very generous tolerances, humans are insensitive to phase shifts. Under carefully contrived circumstances, special signals auditioned in anechoic conditions, or through headphones, people have heard slight differences. However, even these limited results have failed to provide clear evidence of a 'preference' for a lack of phase shift. When auditioned in real rooms, these differences disappear.. ."

I am intrigued that one set of industry gurus say that microsecond-level timing differences are highly audible (MQA), and others say that we can heavily distort the signal with phase shifts (in non-corrected speakers for example) without it being audible.

Re: Meridian Audio's new... sub-format called MQA.

Reply #247
Possibly cut-and-pasting from the same web site you looked at
My quotation is directly from the AES site/paper I linked. Please spend more time there.
I know Dr Tooles opinions and summaries quite well thank you. As well as how often they are misunderstood and misquoted.

Dr Floyd Toole:

" It turns out that, within very generous tolerances, humans are insensitive to phase shifts. Under carefully contrived circumstances, special signals auditioned in anechoic conditions, or through headphones, people have heard slight differences. However, even these limited results have failed to provide clear evidence of a 'preference' for a lack of phase shift. When auditioned in real rooms, these differences disappear.. ."

But the official story is that phase shift is inaudible
One of these things is not like the other, which one is it, can you tell?
(Hint, you've misunderstood the last part).
Loudspeaker manufacturer

Re: Meridian Audio's new... sub-format called MQA.

Reply #248
Sigh.