MQA Users & Discussion Thread

[acg]

44 minutes ago, rmpfyf said:

 

No you're right.

 

Just you can go even further in bitrate given the same source material with MQA-compliant hardware.

(There's the rub.)

 

Ah, ok.  Thanks.

 

Bitrate schmitrate...it does not really matter so long as the mastering is the same...I think the mastering/filter stuff is way, way more important than the resulting bitrate will ever be.

[Guest rmpfyf]

2 minutes ago, acg said:

Bitrate schmitrate...it does not really matter so long as the mastering is the same...I think the mastering/filter stuff is way, way more important than the resulting bitrate will ever be.

 

I think we all admire MQA being a first attempt at a filter scheme that changes sympathetic to music content. Insofar as mastering goes, this much is a nice achievement.

[a.dent]

2 hours ago, bhobba said:

 

That's similar to what we found at a similar one I was at yesterday.  MQA was clearly better - but we didn't test a normal CD - just a ripped CD.  However I have found ripped CD's to be better than actual ones on all except one really heavily tweaked transport in previous comparisons - but opinions vary without going into the details.

 

Thanks

Bill

 

I don't know if my Vitus CD/SACD  player could be called a heavily modified transport but it is pretty good. And it consistently outperforms my version of an "idealish Music Server" with expensive PSU etc. 

 

I predict once MQA decoding to 96/24 in Roon and Audirvana is available it will be damned close. 

[a.dent]

2 hours ago, AudioGeek said:

I thought you never get the full benefit of mqa unless you have an mqa capable dac?

 

I don't think you will need to to get a significant improvement over standard CD rip FLAC. Just that first "unfold" should give most of the benefits, especially if you already have a good DAC.

[ThirdDrawerDown]

Thanks DWM

 

So then the correct reading is as follows - (fixed that for you)

Quote

Despite what MQA say, this is DRM because on inspection:

 

You must pay to receive full quality  (via license fees for hardware)

You cannot edit the audio (and continue to receive full quality)

MQA  have the power to change the quality of the audio

MQA have the power to render it unplayable

 

Certainly walks and quacks like a duck.

 

[Guest Eggcup The Daft]

7 hours ago, rmpfyf said:

 

We have that system - it's called a studio master.

Good.

Interesting that I've been told off here for making the very same point, about MQA and elsewhere. Remember, the things that MQA is trying to clean up in existing digital masters are included in the result that was signed off in the studio master. The mastering engineer may have already attempted to compensate for ADC woes and such... so the MQA sound may not be that of the master.

 

7 hours ago, rmpfyf said:

 

 

On that, there is no technical reason Meridian needs to know what DAC chip you've got to write a full software decoder.

Actually, there are. Remember, one part of the MQA process is the algorithm that claims to improve the timing of impulse response beyond that of normal PCM decoding. They need to know what the final conversion to analogue is to ensure that that response is preserved (as well as adding filters to control for the DAC chip's own "errors").

Nothing to stop further unpacking though. On the other hand, is it needed? In the discussion above, we looked at a thick line on an MQA graph and tried to work out if it meant that MQA's own argument on timing meant that only 96kHz was actually needed to meet Bob Stuart's own target for that... and the experts here don't think that impulse response stuff is necessarily audible anyway. But it is the bit that defending the patent would depend on. As you said, getting audiophiles to pay the license fees...

 

7 hours ago, rmpfyf said:

 

I would be reluctant to conflate issues like jitter etc with filtering, which is what MQA addresses best. High frequency content is arguably less susceptible to jitter whether MQA or not. 

The thing about a lot of these new technologies is that they need to buffer the data to process it. That can reduce the point where jitter is important, to the internal transfer of the processed data to the DAC hardware itself. How did jitter on a USB connection ever get allowed to be important to USB DACs anyway?

 

[Guest Eggcup The Daft]

Has anyone compared the 96k output of Tidal (or decoded MQA on a suitable DAC) against the available DXD and similar masters from 2L? Surely, a better comparison than with CD, as the master is available and we know it is also the source for the MQA encoding - this is the one point where we can definitely compare like with like.

 

I'm looking to try this myself in the next few days but will be restricted to headphones and an internal PC DAC for the test initially, so am suggesting this route for the rest of you if you're interested.

[Guest rmpfyf]

59 minutes ago, Eggcup The Daft said:

Interesting that I've been told off here for making the very same point, about MQA and elsewhere. Remember, the things that MQA is trying to clean up in existing digital masters are included in the result that was signed off in the studio master. The mastering engineer may have already attempted to compensate for ADC woes and such... so the MQA sound may not be that of the master.

 

If the master's good there's nothing to change.

 

If otherwise, the tools MQA affords can be (essentially) found elsewhere. 

 

59 minutes ago, Eggcup The Daft said:

Actually, there are. Remember, one part of the MQA process is the algorithm that claims to improve the timing of impulse response beyond that of normal PCM decoding. They need to know what the final conversion to analogue is to ensure that that response is preserved (as well as adding filters to control for the DAC chip's own "errors").

Nothing to stop further unpacking though. On the other hand, is it needed? In the discussion above, we looked at a thick line on an MQA graph and tried to work out if it meant that MQA's own argument on timing meant that only 96kHz was actually needed to meet Bob Stuart's own target for that... and the experts here don't think that impulse response stuff is necessarily audible anyway. But it is the bit that defending the patent would depend on. As you said, getting audiophiles to pay the license fees...

 

MQA does not have a mortgage on correcting response timing - there are many such means. 

 

Even if the DAC IC needs to be characterised it doesn't imply Meridian can't open the process. 

 

59 minutes ago, Eggcup The Daft said:

The thing about a lot of these new technologies is that they need to buffer the data to process it. That can reduce the point where jitter is important, to the internal transfer of the processed data to the DAC hardware itself. How did jitter on a USB connection ever get allowed to be important to USB DACs anyway?

 

No buffering, and USB is a packet format - hence USB jitter got to be important.

 

No matter, one can buffer I2S happily regardless.

[bhobba]

8 hours ago, rmpfyf said:

Done these many times Bill. It's as I suggested earlier - content-dependant.

 

Strange how your experience differs to mine.  What I quoted is typical of what I got with I would say about 30 tracks - some down-sampled from DXD and others from DSD.

 

Don't know how to explain the difference.

 

Thanks

Bill

[davewantsmoore]

9 hours ago, a.dent said:

How many of the people who opine that MQA is only about DRM

 

I've never heard anyone ever say MQA is only about DRM

[bhobba]

6 hours ago, a.dent said:

I don't know if my Vitus CD/SACD  player could be called a heavily modified transport but it is pretty good. And it consistently outperforms my version of an "idealish Music Server" with expensive PSU etc.  I predict once MQA decoding to 96/24 in Roon and Audirvana is available it will be damned close. 

 

Seems to vary.   I personally find computer audio better, but others find issues like glare, whiteness etc that I don't hear.  The transport I heard was an upgraded Wadia with very low jitter clock.  It was better.

 

Will see if I can arrange a comparison of MQA with that.

 

Thanks

Bill

 

[powerav]

Any PC sound cards out yet with MQA? 

[davewantsmoore]

8 hours ago, rmpfyf said:

.... but the freedom I once had to roll DACs at will left the building for a licensing reason. Which isn't a technical reason and shouldn't ever be a consideration for an audiophile. Or any sort of customer. Even Apple doesn't give a damn what you play iTunes content on, and that's a walled garden in the extreme. Meridian is taking the piss here.

 

There is an argument for needing to know about the DAC (ie. what it does to the audio), in order to make the type of corrections MQA are proposing.

 

Is it the very very small end of the wedge.   Probably.

 

8 hours ago, rmpfyf said:

The above renders it pretty much unable to become an ISO/IEC format, which leaves it a specialty thing.

 

Hence, the PCM container (which is a universal format).

 

A cynical viewpoint is that the intention is to reduce the quality of the regular PCM audio (for non-MQA people), and use the encoded data to undo that quality reduction.

 

8 hours ago, acg said:

I can stream Tidal at MQA bitrates with my non-MQA hardware now, as I assume others can already do so.  This is through XXHighEnd software.

 

Yes, but only at 96 / 88.2khz

 

I was talking about a decoder which will give the full filtering (using higher time precision), and which could be used independently of tidal.

 

 

 

[bhobba]

7 minutes ago, davewantsmoore said:

I've never heard anyone ever say MQA is only about DRM

 

Its not.

 

Probably best if I explain exactly whats going on.

 

Why do hi res recordings sound better? Your hearing peters out at 20k - mine at 62 goes to 12k and that was measured a couple of years ago so is likely worse now. So why does it sound better?

 

Much research has been done into it and the culprit is the brick wall anti-aliasing filter smearing transients. The higher that frequency of the brick wall filter the less smear you get. Now a DAC like the Direct Stream (DS) uses a filter at a phenomenal 10 times DSD - way beyond frequencies that will cause audible smearing. What you do is you put in whatever is coming in, then pad the data not that high out with zeroes, then apply the filter. Having read what the DS is doing its a bit more sophisticated, but that would be the effect:
http://www.6moons.com/audioreviewhttp://www.aes.org/e-lib/browse.cfm?elib=17501s2/psaudio/1.html2

 

But at recording its another matter.

 

MQA does all sorts of tricks to attack that issue at the source. Depending on your background there is various explanations of exactly whats done. I will give the following link, but if it's too advanced I know others:
http://www.aes.org/e-lib/browse.cfm?elib=17501

 

Just as an overview at 96k frequency it doesn't take samples but rather, using a recording at higher frequency than 96k convolves it with overlapping triangle functions (actually the function is a bit more complex than that but no need to go into that in the overview). This has the effect of a transient not being interfered with as much. It has the effect of applying a 6db slow roll off filter from about 20k at a higher frequency like 192k (depending on how that convolving is done and the sample rate of the source) that does very little time smear compared to a brick wall anti aliasing filter. I will assume 192k here - but it can be a lot higher. I prefer half DXD - but that's another story I may detail later.  It then chucks away the bits above 48k ie every second sample for a 192k source, which according to the Shannon sampling theorem means frequencies above 48k are folded back into the 48k range ie for 192k the frequencies from 48k to 96 k. But this is of little concern since due to the 6db roll-off, and amplitudes at that frequency are low anyway, it's way below the noise floor.

 

Now we need to 'unfold' it to recover the 192k. The optimum way to do this depends on the function for convolving - the triangle function is very simple - it's just linear interpolation. The way the DS works is it uses its own filter to upsample to 30bits PCM at 10 times the DSD frequency then converts it to DSD. Optimally it should use the filter that matches the convolving function it was encoded with ie simple linear interpolation for the triangle function. Indeed its even more complex than that. MQA takes into account the characteristics of whatever DA converter the DAC uses - it figures out the optimal filter from that. MQA has metadata on the convolving function used to create the file and its up-sampling needs to match that. The DS uses down-sampling and a delta zigma modulator to convert the 30 bits 10 times DSD PCM to double DSD for output. The output is very simple - it is just a filter that cuts out above about 80k. So the up-sampling to 30 bit 10 times DSD should use an optimal filter for MQA that matches the down-conversion DSD converter and output filter.

 

What Roon and Audirvana is going to do is simply provide the 96K PCM. The unfolding to higher frequencies (ie any up-conversion done), if done in the DAC, will not be optimal. How will it sound - who knows. What I can say is in my experimentation with the DS it sounded very good. Although we didn't do an A/B comparison (I will try and do one later) it was felt it was better than 96k masters. Why would that be? The lack of a brick wall filter in creating the MQA. Oh - on sources MQA encoded from masters that use brick wall filters they have tricks to compensate.

 

Thanks
Bill

 

[davewantsmoore]

10 minutes ago, bhobba said:

Its not

 

Yes, I know.    

Edited February 18, 2017 by davewantsmoore

[bhobba]

7 minutes ago, davewantsmoore said:

Yes, I know.

 

We know you do.

 

Its meant for others, and explains why its best to do the final unfolding inside the DAC, at least in some cases.  It could almost certainly be done in some cases in software, but not so sure in all cases eg the DS DAC.

 

BTW I still think MQA has published a lot of confusing BS about it, but fortunately we don't need to worry about it - just have a listen.

 

I will detail mine with the Explorer 2 in another thread.

 

Thanks

Bill

[Guest Eggcup The Daft]

35 minutes ago, powerav said:

Any PC sound cards out yet with MQA? 

While it's external, that basically describes the Explorer 2...

[sir sanders zingmore]

call me naive or cynical (or both), but I'd be very surprised if Meridian or PS Audio have really found a way to stop transients from smearing.

Or at least not in any way that's audibly better than current methods.

 

[bhobba]

Just now, Sir Sanders Zingmore said:

call me naive or cynical (or both), but I'd be very surprised if Meridian or PS Audio have really found a way to stop transients from smearing. Or at least not in any way that's audibly better than current methods.

 

PS audio doesn't - its a simple brute force approach.  You take the sample, pad it out with zeroes until its ten times DSD then apply the brick wall filter.   You get exactly the same smearing but since its at such a high frequency its inaudible.

 

What MQA does is, in effect, use a 6db per octave filter during down-sampling from a master such as DXD or 4x DSD - not a brick wall one - it has virtually no ringing or smearing.   You need to transmit a higher frequency of samples but it has tricks to do that (dithering, bit stacking and folding).  That's all MQA is, basically.   Its not really that hard - but there seems a lot of BS about it.

 

Thanks

Bill

[LHC]

6 minutes ago, bhobba said:

 

PS audio doesn't - its a simple brute force approach.  You take the sample, pad it out with zeroes until its ten times DSD then apply the brick wall filter.   You get exactly the same smearing but since its at such a high frequency its inaudible.

 

What MQA does is, in effect, use a 6db per octave filter during down-sampling from a master such as DXD or 4x DSD - not a brick wall one - it has virtually no ringing or smearing.   You need to transmit a higher frequency of samples but it has tricks to do that (dithering, bit stacking and folding).  That's all MQA is, basically.   Its not really that hard - but there seems a lot of BS about it.

 

Thanks

Bill

 

Hi Bill the question was not how the smearing is reduced or removed, but whether the removal of smearing can result in improvements that are audible by human hearing. In a separate thread about the benefits of 197kHz there was a lot of argument that such smearing reduction is pointless because we can't actually hear it. Obviously Bob Stuart thinks otherwise. 

 

 

[bhobba]

27 minutes ago, Eggcup The Daft said:

While it's external, that basically describes the Explorer 2...

 

Indeed it does.

 

Its very very small - I will do a post on my experiences with it a bit later.

 

Thanks

Bill

[bhobba]

33 minutes ago, LHC said:

 

Hi Bill the question was not how the smearing is reduced or removed, but whether the removal of smearing can result in improvements that are audible by human hearing. In a separate thread about the benefits of 197kHz there was a lot of argument that such smearing reduction is pointless because we can't actually hear it. Obviously Bob Stuart thinks otherwise. 

 

That's where listening tests come in.

 

I will describe mine is a separate thread on the Explorer 2 - others can describe theirs.

 

I know for example Paul McGowan thinks its slightly worse than the Masters it came from:

http://www.psaudio.com/pauls-posts/mqa-thoughts/

 

Personally for what its worth I think Paul doesn't like it (in the sense of checking what filter changes will do) because it would require licencing fees and slow down the software releases they do for their DS - they do it quite frequently and it gets better with each one.

 

Thanks

Bill

Edited February 18, 2017 by bhobba

[Guest rmpfyf]

1 hour ago, davewantsmoore said:

There is an argument for needing to know about the DAC (ie. what it does to the audio), in order to make the type of corrections MQA are proposing.

 

Sure. But that's knowing about the DAC IC as opposed to licensing a structure around it. 

 

It's also the hysterically thin edge of the wedge and frankly not something that can't be done on an upstream FPGA, appropriately configured. Or a licensed hardware predecoder on an IC. Or, with enough grunt and for certain output sample rates, a PC. Or in many other ways that aren't 'here are your choices of approved white boxes, enjoy'. 

 

1 hour ago, bhobba said:

it was felt it was better than 96k masters. Why would that be? The lack of a brick wall filter in creating the MQA. Oh - on sources MQA encoded from masters that use brick wall filters they have tricks to compensate.\

 

That's the issue in hype, Bill. I'd not assume there no filtering in creating MQA. Or that 96k is a studio master. Or that 'tricks to compensate' are Meridian's sole purview. 

 

If you can hack the bandwidth, there are other ways that aren't closed. If.

 

[frankn]

A couple of comments posted by Ted Smith (in PS Audio Forum) regarding MQA and DS.   Certainly there is as much discussion on PSA as there is here. 

 

Upsampling in software before the DS means you are listening to the sound of that upsampler rather than the upsampler in the DS.  You may prefer one over the other (and either may be right for you) but most people would contend that part of the advantage for the DS over MQA is it’s upsampling of CDs.  Tho you may prefer the sound of HQPlayer at the outset I’m not terribly surprised that you like MQA compared to it.  Also there’s no such thing as “absolutely no USB noise”: all connections have noise, we just pick a noise level that we find acceptable with tweaks.  (And at times with different tweaks we are just picking noise with one character compared to noise with a different character.)

So far we can’t get the noise from MQA to not affect the sound of the DS in a negative way.  If (and at this point probably when) MQA is allowed to run before the final DAC we may be able to let MQA fans enjoy what they like from MQA by having MQA running in the playing software or the streamer, etc. without degrading what the DS does well.

Any connection generates noise, on average each connection makes tradeoffs as to which noises they minimize.  I2S is good in that it’s balanced and has the clocks separated from the data – on the other hand it’s only designed for short connections (i.e. no cables.)  That doesn’t mean that I2S with cables is necessarily bad, but different choices might have been made if that were the goal.  EMI radiation and pickup (and hence the length of the cable), quality of connectors and interfaces between connector pins and wires, bulk impedance, signal levels, etc. all affect jitter and noise down a cable.  With I2S those effects can be more easily handled than, say, S/PDIF, etc. but they are still there.

Possible to do MQA in the DS’s FPGA – possibly, but ignoring all of the IP and company to company logistical issues there’s still the fact that everything in an FPGA affects the noise of the whole and that just the presence of the MQA code in the FPGA, even if it’s not being used, will cause more jitter and noise in the FPGA.  With a certain amount of energy and time I can make more of a difference in the quality of the FPGA output doing other features than ameliorating MQA’s negative footprint (especially if the MQA code ever changes).  It’s much more logical to put the MQA decoding upstream where it only affects the streaming code.  And FWIW we already know that some upstream MQA implementations (including those from MQA itself) won’t make the average DS customer happy.  That doesn’t mean that we wont find a good place to decode MQA, it just means that the obvious places don’t work well yet.

I can’t speculate what other people like or don’t like – but I have a good idea what things in an FPGA do to the average feedback about DS releases.

[a.dent]

14 hours ago, davewantsmoore said:

 

I've never heard anyone ever say MQA is only about DRM

 

You obviously don't read the conspiracy theorists posts on CA.

Edited February 19, 2017 by a.dent
grammar