Jump to content
IGNORED

MQA Software / Hardware Decode Etc... Questions


Recommended Posts

Hi Guys - I'm writing up my MQA article on hardware and software decoding. I know there have been some attempts at explaining how it works, but I also know there have been many more questions as to how it works.

 

Please post questions about software and hardware decoding in this thread and I will address them in my article. If I don't have the answer, I'll talk to MQA ltd. and get the answer.

 

I'm seeking the facts and trying to help everyone. I'm not interested in the arguments about MQA. Those are for another thread.

Founder of Audiophile Style

Announcing The Audiophile Style Podcast

Link to post
Share on other sites

How could be custom DSP used with MQA content? It seems any custom DSP between the software decoding (MQA core) and DAC would cause the audio stream not to be recognized as MQA in MQA capable DAC, so full MQA decode wouldn't be performed.

 

As custom DSP I mean for example room acoustic equalization, headphone crossfeed or binaural processing, simple equalization, multichannel to stereo etc.

 

Many SW players like JRiver, foobar2000 and others support VST or their own format of DSP plugins. Any chance to use them with hardware MQA decoding in DAC?

 

Any chance to get full decode or MQA renderer functionality in software? That could allow to apply DSP on fully decoded stream.

 

Any chance to get at least MQA software decoding (MQA core) as VST plugin or as DSP plugin in popular software players? If that will not be available from the MQA company, then I expect that some non original MQA decoders will appear as choice in these players after some time ...

Link to post
Share on other sites

I posed these questions is another thread but I'll paste here (with minor edits):

 

- pro/con of full hardware decode vs software+hardware decode vs software only decode

- why are 44.1 and 48k tracks upsampled (according to my non MQA DAC) when running the Tidal software decoder?

- does the hardware MQA decoder also do custom upsampling (see previous question) of 44.1K material, or does it just leave that to the DAC upsampler (assuming its an upsampling DAC)?

 

A bit wider than just technical MQA stuff, other open questions include:

- what about room correction?

- why did miniDSP get denied a MQA licence? (OK we won't get a direct answer on that).

- Without miniDSP being MQA enabled, what hope do we have of affordable room correction that works with MQA? (excluding partially decoded MQA which yes, can be fed into miniDSP)

- how will it all work in the new MQA release of Roon? (I don't expect Bob to answer that, buts its a question we are all asking!)

- why are there upsampled tracks on Tidal Masters? Isn't the whole point of MQA avoiding that? (probable answer: the labels can do what they want with MQA, within reason)

Link to post
Share on other sites

-Would DACs that have a bandwidth limited interface but DAC chips capable of higher sampling rates actually be able to unfold to the full capability of the DAC chip? Example, Meridian Explorer 2 is limited to 192khz PCM playback, but the DAC chip it uses supports 384khz. So would a 2L 352.8khz MQA file unfold to 352.8khz resolution on this device?

 

-What performance gain is there to using an MQA dac on 44.1khz MQA source material? The reason I ask is because there is no unfolding, so no advantage there that MQA brings. The ADC deblurring is already taken into account in the file, so non-MQA dacs would benefit from that already thus no advantage MQA dac would bring.

 

-It seems software MQA decoders like Tidal are upsampling 44.1khz MQA source files to 88.2khz. Does hardware MQA also upsample source files originally 44.1khz to a higher rate internally in the DAC? If so, what rate would it run at? How would this be different from DACs that already internally upsample 44.1khz content anyway, like the Explorer 2?-MQA talks about having better temporal response than even high sample rate PCM files. Where does this benefit get applied? At the source file de-blurring stage such that it benefits all DACs? Or the dac profiled rendering stage? Or a combination of both of the previous? Or somewhere else?

 

-In a 44.1khz source encoded as MQA, does it still have the temporal response MQA claims exceeds that of 192khz pcm? How does it achieve that?

 

-In an unfolded MQA file, is the high frequency information restored exactly matching what a high sample rate PCM would achieve? Such that if we ignore the DAC profiling part of things, an MQA decoded file would sound the same as that same file with the same deblur corrections applied but delivered as high sample rate PCM?

 

-The DAC profiling part of MQA is the piece that is not very clear how it works. Can this be explained in more detail?

-How does it work?

-Does it depend on analog circuitry of DAC or just digital behavior?

-Can this DAC compensation potentially be used regardless of input being MQA or plain PCM? If no, why not?

 

-What is the breakdown of the importance of each piece of the MQA chain in terms of contribution to the final sound quality?% DSP deblur of original source, % unfolding into high resolution, % rendering output with DAC profiling = 100% MQA quality

Link to post
Share on other sites

I'd like to learn some more on how the output of the triangular sampling kernel data in the ADC is decoded at the DAC end. This seems to be one of the most important components of the temporal de-blurring process. It appears that a pulse can have data spread over 2 samples. How is that reconstituted as 1 pulse with greater timing precision than the sample rate in the DAC? up to10 times greater than a 192Khz sample rate according to MQA

kernel.jpg

Link to post
Share on other sites

Thanks to abrxx and bcwang for your posts. Those are key questions for me and I never could have expressed them as well as you did.

 

Exactly why the DAC-specific customization is so essential to the resulting sound quality is something I've been asking since the day I first heard about MQA. Is this anything more than just making sure that the standard vendor-supplied filters don't mess up what MQA claims to fix?

 

Not having clear answers to these questions just fuels skepticism about the technical merits of MQA and efforts to figure out exactly how it works.

NUC7PJYH/AL --> Berkeley Alpha USB --> Jeff Rowland Aeris --> Jeff Rowland 625 S2 --> Focal Utopia 3 Diablos with 2 x Focal Electra SW 1000 BE subs

 

i7-6700K/Windows 10 Version 2004/HDPLEX 300W/HDPLEX 400W DC-ATX --> EVGA Nu Audio Card --> Focal CMS50's 

Link to post
Share on other sites
If the TIDAl app (and later Roon) decodes to 24-96 and given the Dragonflys are limited to 24-96, what's the point of the Dragonflys being MQA DACs?

 

I guess, because you could get the software decoding part when playing from apps that don't do SW decoding, e.g. Mobile apps or anything other than Tidal, Audirvana + V3 and later Roon.

 

 

Sent from my iPhone using Computer Audiophile

Owner of: Sound Galleries, High-End Audio Dealer, Monaco

Link to post
Share on other sites

It bothers me if some Tidal Masters which software decode to 24/96 are just files which have been up-sampled to 24/96 rather than true Hi-Rez. Do we have specific examples?

I've seen Tidal Masters which are limited to 44.1 (Bowie's Young Americans). That keeps it honest. I really hope labels are not passing off upsampled files as High-rez withon MQA files. Surely one of the main points of MQA was that the file was authenticated?

 

Further up-sampling in the DAC is another matter; but how do we know, with and MQA DAC if the signal reaches 192 or above whether that's been achieved by dac up-sampling or whether the file was originally at that rate before MQA folding.

 

 

 

 

Sent from my iPhone using Computer Audiophile

Owner of: Sound Galleries, High-End Audio Dealer, Monaco

Link to post
Share on other sites
If the TIDAl app (and later Roon) decodes to 24-96 and given the Dragonflys are limited to 24-96, what's the point of the Dragonflys being MQA DACs?

DAC-side deblurring, i.e. elimination of pre-ringing and post-ringing.

Peter Lie

LUMIN Firmware Lead

Link to post
Share on other sites
It bothers me if some Tidal Masters which software decode to 24/96 are just files which have been up-sampled to 24/96 rather than true Hi-Rez. Do we have specific examples?

 

According to Archimago, Madonna's Like a Virgin is fake hi-res.

Link to post
Share on other sites

I have a few questions:

1. What do I need MQA?

2. How can I turn off MQA if I prefer to listen to the bit perfect original?

3. Why is it called "Master Quality" when it has been reported that it's a hi-resulution lossy version of the master containing artifacts on top?

4. If there is a way to eliminate the impact of the a-d converter and d-a converter why not offer just that part of the scheme?

5. Has anyone studied the impact of the embedded extra data on the sound quality of the original file?

Link to post
Share on other sites
but how do we know, with and MQA DAC if the signal reaches 192 or above whether that's been achieved by dac up-sampling or whether the file was originally at that rate before MQA folding

 

Oh that's easy: the echt hirez won't have the temporal blur so it will be easy to tell the difference. Simples.

You are not a sound quality measurement device

Link to post
Share on other sites

In my case, I've not become a real streaming person (yet?) if/when I stream music, "party quality" is good enough, I'm not all that concerned about it being the best quality. My listening room has no capability for streaming music anyway. When I listen to streaming audio it's through mid to low quality amplifiers/speakers anyway.

 

My understanding..and maybe it's wrong.. Is that MQA encoded recordings have embedded within them (Metadata?) describing various technology used during the original recording process. e.g. the ADC used.

 

Knowing this information up front and sending this info along embedded within the MQA file, would allow the MQA capable DAC to adjust it's own filtering etc. during playback to in effect compensate for the characteristics of the recording process.

I'd assume this "compensating" could ultimately be done directly within an application e.g. JRMC.. but I'm just "ASSuming".

 

This would be of interest to me, as, if in fact there is a benefit here I would ultimately consider the purchase of an MQA capable DAC and I'd purchase MQA recordings... Then I'd have my library with WAV/FLAC/DSD/ of various sampling rates/bit depths along with MQA encoded recordings. The DAC would of course need to do seamless switching between all the formats.

Link to post
Share on other sites

BTW I had a thought yesterday ... are there different types of MQA material? We know there are at least two:

 

 

1) Archive recordings that have gone through an automated deblur algorithm

 

2) Archive recordings that have gone through a "white-glove" remaster (where ADC/DAC details are known)

 

But what about material custom recorded with a MQA ADC? (I think Mytek have just released one).

 

Its clear that a deblur where the ADC etc are known must be better compared to the one where the software is making guesses. But what difference does using a MQA ADC bring to the table?

Link to post
Share on other sites

The article was sent to MQA last night. As soon as I receive a response I'll publish it on the front page.

 

 

I should add that I tried to address the questions where appropriate.

 

The article isn't a technical deep dive but rather something for people who just want to enjoy MQA an understand a little bit about what's going on.

Founder of Audiophile Style

Announcing The Audiophile Style Podcast

Link to post
Share on other sites
In my case, I've not become a real streaming person (yet?) if/when I stream music, "party quality" is good enough, I'm not all that concerned about it being the best quality. My listening room has no capability for streaming music anyway. When I listen to streaming audio it's through mid to low quality amplifiers/speakers anyway.

 

My understanding..and maybe it's wrong.. Is that MQA encoded recordings have embedded within them (Metadata?) describing various technology used during the original recording process. e.g. the ADC used.

 

Knowing this information up front and sending this info along embedded within the MQA file, would allow the MQA capable DAC to adjust it's own filtering etc. during playback to in effect compensate for the characteristics of the recording process.

I'd assume this "compensating" could ultimately be done directly within an application e.g. JRMC.. but I'm just "ASSuming".

 

This would be of interest to me, as, if in fact there is a benefit here I would ultimately consider the purchase of an MQA capable DAC and I'd purchase MQA recordings... Then I'd have my library with WAV/FLAC/DSD/ of various sampling rates/bit depths along with MQA encoded recordings. The DAC would of course need to do seamless switching between all the formats.

I think you are misunderstanding how MQA works. It's complicated. I will try my best understanding. I might even be right for a change. But, if I am not, someone should let me know.

 

For the ADC end, I believe that the MQA audio signal data itself is corrected in the MQA remastering process to achieve deblurring of the recording. This is why an MQA recording can be played via an non-MQA DAC and still get some of the benefits of the deblurring - those at the recording ADC end. Those corrections are NOT passed through as metadata to be applied by the MQA DAC, since MQA's ADC corrections were already applied to the signal.

 

At the playback end, deblurring of just the DAC conversion is applied by an MQA DAC via its filters. The metadata passthrough from the recording does not affect D-A, but it does turn on the little MQA Authenticated light where appropriate in one of its indicator colors.

An MQA DAC can also perform the origami unfolding of higher rez versions of the music that are buried in the noise floor of MQA recordings.

 

A non-MQA DAC cannot do the unfolding, apply DAC deblurring or indicate Authentication info. However, the unfolding of hi rez "layers" in the MQA recording could also be done in software on PC streamers/players, such as Tidal's app. Unfolding to higher rez could also be done in DSP in a Mch preamp, for example, ahead of other processing - like room correction - prior to a final internal DAC.

Link to post
Share on other sites

Chris- this is not exactly germain to what you are asking for but...

I just received a new Meridian Explorer 2 the first time I connected it to my mac mini it work flawlessly with A+ ver 2.6.5 and with TIDAL and Qobuz desktop apps and with ROON.

Now the E2 doesn't transmit any sound with A+(for either TIDAL or Qobuz) in use no lights are on BUT the sound/music bar at the top of the app moves so music is playing but i get no sound to my headphones or my avr and I do not have them connected at the sametime. TIDAL and Qobuz work flawlessly using their respective desktop apps.

I emailed Damien- he responded immediately then I sent him screenshots of the A+ 'debug' info and other shots of A+-he couldn't figure it out.

I called Meridian in the USA their tech person couldn't explain it but sent specific instructions on how to connect E2 with A+ and still no sound.

Do you or anyone else have any idea what happened or how to fix this?

thanks

Link to post
Share on other sites
Do you or anyone else have any idea what happened or how to fix this?

Try this:

Disconnect your Explorer2. Reboot the Mac mini. Power cycle the Explorer2. Reconnect it to the Mac mini.

 

If that doesn't work, uninstall your Audirvana and reinstall it.

Peter Lie

LUMIN Firmware Lead

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×
×
  • Create New...