is MQA going to be the new Standard? Is it better than DSD and PCM, opnions.

[Kefas]

MQA should be great for portable devices due to it's file size, but does it really sound as good as claimed;

MQA accomplishes the seemingly impossible—delivering extraordinary sound quality in a form that easily integrates into the existing music distribution infrastructure, is backward compatible with existing hardware, and serves all applications, from smartphone streaming to high-end home downloads, with a single file.

http://www.theabsolutesound.com/articles/beyond-high-resolution/

 

[LenWhite]

According to the current month AS article (pgs. 76-77) regarding new material, MQA can "take the listener one step further back in the chain to the signal before the master recording". RH compared a Peter McGrath original recording in 88.2kHz/24-bit vs. an MQA version, RH states the latter was much better in terms of timbre, dimensionality, differentiating individual instruments, and emotion.

RH also asked MQA to encode a file containing one of his 1988 DAT recordings (44.1/16-bit), providing the A/D converter information used to make the original recording. RH states his comparison of his original file to the MQA file, the MQA version had improved instrumental timbre, dynamic attack, specific location, air and bloom, resolution, and soundstage dimensionality. Other recording professional testimonials are included in the article.

Of course the thought crosses my mind this is just another industry manipulation to sell more audio equipment and recordings. And I would certainly prefer my current DAC be upgradeable via firmware rather than buying both new hardware and software.

But for me to buy into MQA, the process will have to easily demonstrate dramatically improved playback on recordings, demonstrate that other formats (e.g.; PCM) won't piggyback on MQA labeled products such as they have with SACD "DSD", and MQA recordings must become ubiquitous unlike SACD's. And since MQA is proprietary, licensing fees may seriously curtail recording companies from using this process, similar to Sony's DSD fees.

I bought into the SACD phenomenon and have lots of discs. It reaffirmed the recording and mastering process are more important than the format. While native DSD and hirez PCM recordings arguably have brought sonic improvements, it's a mixed bag.

Perhaps if MQA is made available to all recording studios and artists, is easy to use, and affordable, I can see it may be successful. Somehow the industry is going to have to demonstrate they're really on-board with creating high quality, high resolution audio recordings. Maybe MQA is the answer, time will tell.

 

[rbbert]

Since Len duplicated his post from the 'Shark I will do the same

As per Bob Stuart, MQA is basically a "lossless" (but not really) compression scheme for PCM audio. If it sounds better than a hires PCM audio file, there is something else going on. And since it is not really lossless compression, why can't we get the true lossless PCM files instead??

 

[fas42]

I've started to look at a sample of MQA "fiddling", from 2L. Have the hi res, CD and MQA versions of a demo track - and, surprisingly, the MQA variety sounds "better" on laptop speakers than the CD - no MQA decoding happening on this old fellow - so, what's going on?

On closer inspection, the level of high treble frequency energy is markedly less for the MQA than CD, it's "rolled off"! The peaks of treble energy have not been touched, at a first glance, but in the areas where the treble is "quieter", the MQA is even quieter!

To me, what's happening is that in the area where low cost gear will tend to create unpleasant IMD from excess high frequency content, that the the level is effectively reduced by the MQA processing - hence, in many cases the MQA version will always automatically sound "better" ...

Depending upon my interest, and that of others, I'll take a closer look at the underbelly of this beast ...

 

[Ken Newton]

I'm pretty certain that MQA is founded on PCM.

It seems to me that the key sound quality issue targeted by MQA is the minimization of temporal smearing caused by ringing anti-alias and reconstruction filters. In order to utilize the non-ringing filters MQA does, with their soft rolloff slope, the system bandwidth has to be dramatically expanded so that those filters have spectrum to reach their stop band frequency in order to sufficiently suppress any would-be alias products. Which, I suspect, is why the extra bandwidth can be encoded with low dynamic range and folded in to the noise floor of the audio band. The ultrasonic bandwidth isn't captured for the fidelity of any ultrasonic content there, it's captured simply to prevent aliasing until the soft slope filters reach their stop bands. In this way, MQA adds little to no temporal smearing while also effectively suppressing recording aliases and playback images (which are opposing system constraints) without wasting bandwidth and storage space.

DSD is another matter. Whether or not MQA it's subjectively better than DSD will be in the ear of each listener.

 

[Bruce B]

As per Bob Stuart, MQA is basically a "lossless" (but not really) compression scheme for PCM audio. If it sounds better than a hires PCM audio file, there is something else going on. And since it is not really lossless compression, why can't we get the true lossless PCM files instead??

+1000... it can't sound better than the original file. If it does, then something fishy is going on and it's altering the file in some way. Not cool... Reminds me of the BSG thingy...

 

[Ken Newton]

+1000... it can't sound better than the original file...

Bruce, this is an intriguing question. I do not know exactly how MQA does this, but from the little I've read, I do have an idea of how they might go about doing it. MQA may apply deconvolution to an pre-existing file of samples resulting from a ringing brickwall FIR anti-alias filter. Deconvolution can reproduce the original signal samples nearly as they were prior to having been run through the FIR anti-alias filter. Apparently, this is very possible to do if one knows the kernel coefficients (ascertainable via it's impulse response) of that anti-alias filter. Once the original (prior to brickwall filtering) samples of been recovered, a different anti-alias filter function can be applied, such as MQA's non-ringing filters.

My superficial knowledge of deconvolution pretty much ends there, but I'm left with the following burning question. Might the deconvolved signal now contain alias products unless the output sample rate were increased up to the original ADC native rate? Logically, it would seem that it should.

 

[YashN]

Bruce, this is an intriguing question. I do not know exactly how MQA does this, but from the little I've read, I do have an idea of how they might go about doing it. MQA may apply deconvolution to an pre-existing file of samples resulting from a ringing brickwall FIR anti-alias filter. Deconvolution can reproduce the original signal samples nearly as they were prior to having been run through the FIR anti-alias filter. Apparently, this is very possible to do if one knows the kernel coefficients (ascertainable via it's impulse response) of that anti-alias filter. Once the original (prior to brickwall filtering) samples of been recovered, a different anti-alias filter function can be applied, such as MQA's non-ringing filters.

I believe the MQA process does take into account the ADC used during recording and reverts the effects of that. I think this requires knowing which equipment was used during production, so in cases this information isn't available, that part of the process won't function as intended or won't be used.

Someone must know and explain this much better than me, after a while I haven't kept too much abreast as it seems the wide availability of the technology is slow to ramp up IMO.

 

[bmoura]

+1000... it can't sound better than the original file. If it does, then something fishy is going on and it's altering the file in some way. Not cool... Reminds me of the BSG thingy...

From the MQA demos I've heard, the sound from MQA encoded and decoded music is different. But you'd expect that from processing that includes lossy and lossless compression, etc.
To my ears, a better bet is to stay with the 24/96 PCM and DSD 64 versions of the music.

 

[rbbert]

From the MQA demos I've heard, the sound from MQA encoded and decoded music is different. But you'd expect that from processing that includes lossy and lossless compression, etc.
To my ears, a better bet is to stay with the 24/96 PCM and DSD 64 versions of the music.

I don't think lossless compression would be an issue...

 

[bmoura]

I don't think lossless compression would be an issue...

But MQA doesn't stop at lossless compression. It also uses lossy compression, filtering and other techniques.

 

[Folsom]

I don't like DSD. Delta-sigma DAC's often sound fine, but for whatever reason DSD - that is similar in nature to Delta-sigma/classD - does not. I don't like classD either so... who knows why a Delta-sigma DAC doesn't bother me.

That being said, I welcome MQA if it gets us off the DSD train.

 

[Steve Williams]

I don't like DSD

what bothers you about DSD

 

[Folsom]

It sounds sort of soft and weird to me. Where as the best PCM is more solid and believable. Either can be fatiguing, but it seems some popular DSD "DAC"s are less so... which gives an easy rise to the DSD format popularity it seems.

 

[Ken Newton]

I don't like DSD. Delta-sigma DAC's often sound fine, but for whatever reason DSD - that is similar in nature to Delta-sigma/classD - does not. I don't like classD either so... who knows why a Delta-sigma DAC doesn't bother me.

That being said, I welcome MQA if it gets us off the DSD train.

All current production sigma-delta modulation (SDM) DAC chips I can think of, except for DSD of course, are a hybrid SDM and multibit converter. They typically feature core D/A units (quantizers) of 5 or 6 bits, whereas DSD is 1-bit. This approach is intended to marry the lower distortion delivered by SDM with the lower native quantization noise of multibit, easing the required order of the sigma-delta modulator circuit.

I too have noticed a soft sounding presentation via DSD as compared to PCM. Experimentation has shown me that PCMs characteristic solidity is at least partly due to it's brickwall reconstruction filter. So, while I'm unsure that DSD's soft presentation is not a coloration, I'm equally unsure that PCM's solid presentation is not a coloration.

 

[Geardaddy]

It sounds sort of soft and weird to me. Where as the best PCM is more solid and believable. Either can be fatiguing, but it seems some popular DSD "DAC"s are less so... which gives an easy rise to the DSD format popularity it seems.

Through what dacs? DSD sonics will depend on chip architecture, etc...

 

[Folsom]

Some don't have chips. I listened to Lampizator for the longest of all DSD.

 

[Don Hills]

... MQA may apply deconvolution to an pre-existing file of samples resulting from a ringing brickwall FIR anti-alias filter. Deconvolution can reproduce the original signal samples nearly as they were prior to having been run through the FIR anti-alias filter. ...
... Might the deconvolved signal now contain alias products unless the output sample rate were increased up to the original ADC native rate? Logically, it would seem that it should.

This is my concern, too. My naive understanding is that deconvolution can't restore the signals that caused the filter to "ring" as they were removed. There's also the problem that the signal has likely gone through at least 2 A-D stages - from source to the multi-track, and from the mixdown to the 2-track master - with likely different filters each time. I can see I'll need to do some learning and experimenting. Time to install and learn Octave, methinks.

 

[Geardaddy]

Some don't have chips. I listened to Lampizator for the longest of all DSD.

So nothing you have lived with. Just off-the-cuff impressions.

 

[Folsom]

Call it that if you wish. I call it hours and hours.