Today MQA Audio

Sounds like an App or driver error, it's not setting the header sample rate right in either SPDIF or USB frames

Hmm - tried re-installing the app, and Hugo doesn't need a driver to be driven from a macbook, so am a bit stumped.

Hugo does not need a driver - but a driver none the less is required by the Macbook to drive audio through its USB (or SPDIF) port and the app interfaces to that driver. The driver will be part of the standard OS X build if you haven't had to install one. I suspect its an app bug.. i have no problems driving my Hugo from my MacBook - but I have not used the Tidal app - but I might give it a go if I can find my mini USB lead

I have Tidal running on my MacBook Air to my Devialet 250 Pro which I acquired last month. I stuck to some of the very old favourites available currently available in MQA, such as Bowie's Ziggy Stardust, Led Zep I and Fleetwood Mac Rumours because they allowed me to instantly hear any gain in quality. I matched some of the output against remastered versions available in Tidal and was at first a little uncertain about the MQA stuff as it seemed to be slightly smeary. However, further listening would suggest that the level of detail I was hearing caused my aural confusion. It is certainly not hugely different from remastered versions currently available and remastered music, to me, generally brings a very big gain in quality of output. 

Having recently picked up a Meridian Explorer 2, which is currently available at £129 vs the £199 it has been at for quite some time, it gives a quick way to get a way to natively decode MQA. I would say that initial impressions have been v. positive with Tidal. The main drawback of course is that you have to use your mac/pc to drive the DAC. Using this to drive my UnitiLite, has given me a great way to experience what I ultimately hope will be the setup that we will get natively out of Naim :-!

Each to their own, but in my view MQA is a really positive way to get the best of both worlds - hi-res streaming services (so I don't have to buy my library again) and in a practical way that works even over a somewhat limited broadband connection. 

Listening to "Masters" through my desktop iMac connected to an iFi DSD DAC through Audeze EL-8 headphones. It does sound as a significant improvement to me compared to "regular" Tidal and (less so but also) to red book files. Original hi-res files (ex. Jethro Tull's Aqualung Steve Wilson remix in 24/96) sounds slightly better to me in their original format than Tidal MQA.

london_r posted: 

Each to their own, but in my view MQA is a really positive way to get the best of both worlds - hi-res streaming services (so I don't have to buy my library again) and in a practical way that works even over a somewhat limited broadband connection. 

I'm not entirely convinced about that, I've noticed that the Tidal Masters albums I've tried so far tend to buffer a bit, where the regular 16/44 versions don't, on my slow rural broadband. Perhaps there are other reasons for that, but the jury is still out as far as I'm concerned.

ChrisSU posted:

london_r posted: 

Each to their own, but in my view MQA is a really positive way to get the best of both worlds - hi-res streaming services (so I don't have to buy my library again) and in a practical way that works even over a somewhat limited broadband connection. 

I'm not entirely convinced about that, I've noticed that the Tidal Masters albums I've tried so far tend to buffer a bit, where the regular 16/44 versions don't, on my slow rural broadband. Perhaps there are other reasons for that, but the jury is still out as far as I'm concerned.

I think you have defined your own problem. I have Sky internet at between 30 - 40 Mbps and I have an ethernet connection to my system. I have had absolutely no issues whatsoever with Tidal either before or after using MQA. I don't believe I have the hearing ability to pick up the subtlest of differences but I do know when I listen to something whether it is good or bad reproduction. MQA is very good but whether it is so much better than remastered I will leave that to better ears than my own.  

MQA software decoding has been generating a lot of noise on a number of forums, including this one.

Chris Connaker from The Computer Audiophile has spent 45 minutes taking to Bob Stuart about this topic today in an effort to add some signal. He hopes to write something up soon.

_{Source: http://www.computeraudiophile....dex5.html#post621288}

I suspect Naim may feel it is better to let the dust settle and evaluate the pros and cons of MQA before committing to implementation.

Reading reports elsewhere of the software decoders, they look to be consuming quite a lot of process / CPU capacity... Naim know that is usually a  no-no for best sound quality with their products because of coupling and ground plane noise... perhaps another nod to hardware decoders?

[@mention:1566878603876589] I've noticed this. I'm currently streaming a Coltrane Tidal Master on a maxed out 5K iMac and Tidal is being quite greedy! I'd hope the guys at Audirvana and Roon can do a more efficient job. The Tidal app isn't the best in many respects, but it does appear MQA decoding is demanding some power. It doesn't bode well for those who are using Roon on a NAS. 

The guys at AudoQuest have confirmed more details about the DragonFly USB DAC's MQA support. Contrary to my speculation and their published specs; 24 bit /96 kHz, they will be able to unfold all of the hi-res detail in an MQA file. The Dragonfly Red for example uses a 32-bit ESS Sabre ES9016 DAC which by all accounts is pretty capable.

Either the Black or Red models look to be a good way to test hardware vs. software MQA decoding at the relatively low cost of USD 99 (Black) or 199 (Red), or nothing if you already have one.

I'm using the Red together with an iPhone for personal listening on the go. It's great :-) Soon it will be even more useful.

AQ with Tidal running on a phone with 24/96 would be a good setup but as far I understand the masters are only available in desktop

Simon-in-Suffolk posted:

Reading reports elsewhere of the software decoders, they look to be consuming quite a lot of process / CPU capacity... Naim know that is usually a  no-no for best sound quality with their products because of coupling and ground plane noise... perhaps another nod to hardware decoders?

Hi Simon that is why Rob Watts also had to split his 1 million tap solution between DAC (DAVE) and the 1 million tap scaler (Blu 2) as he said exactly the same thing, the power required by the FGPA (10 amps) in the blu 2 scaler would interfere with Dave.

Maybe Naim would have to go for a 2 or 3 box solution themselves to avoid this issue.

Simon-in-Suffolk posted:

Reading reports elsewhere of the software decoders, they look to be consuming quite a lot of process / CPU capacity... Naim know that is usually a  no-no for best sound quality with their products because of coupling and ground plane noise... perhaps another nod to hardware decoders?

But if the software decoding was 'Off board', in a 'transcoding, MQA decoding, UPnP server', so all the additional processing is transparent to the Naim player as it is just served WAV or DSD64 as it does currently with, no further processing in the streamer board or DAC is required.

Or the processing is done in Roon Core, running on a NUC of which the recommendation is for i5 with 8GB RAM, with Naim as an Roon Endpoint, being served WAV for PCM formats or native DSD64 for DSD formats.

However here, if you wanted to use an external DAC, such as a Hugo device, you would probably be better off using a USB endpoint to the Hugo ��  

This is true

simes_pep posted:

But if the software decoding was 'Off board', in a 'transcoding, MQA decoding, UPnP server', so all the additional processing is transparent to the Naim player as it is just served WAV or DSD64 as it does currently with, no further processing in the streamer board or DAC is required.

 

this is true, but then we revert to the scenario of unmatched encoding and decoding reconstruction filters in the ADC and DAC.  Now I know MQA is actually a pick list of different features that you can selectively use, so if one is simply using the 96/24 compression and expansion part of MQA format then a seperate software decoder as you will work fine... and you simply using MQA as a hidef compressor / audio enhancer algorithm... and it should sound nice for many MQA masters for many people's tastes.

Dont get me wrong I can see in part what MQA is addressing with its lossy hidef encodings, and I admire their patent, there are so I have discord other potential ways of achieving a similar outcome. One thing is for sure standard hidef PCM is very inefficient, as much energy and therefore stored info is simply representing the higher frequency bit dither with no musical or information content. Further this noise has a Gaussian distribution so doesn't compress with FLAC or ALAC.. it's very inefficient... MQA is one approach to address this.

Interestingly, I have no idea of the merits or demerits of MQA coding and decoding but Tidal seem to be going for it big time. I think on release there were about 150 albums available but today there are over 400 with the complete Led Zep and Yes catalogues as far as I can tell.

So, am sitting listening to Led Zep I hammering out at pretty high volume. No jitter, no background distortion, no buffering issues, just clarity. Wonder what it would be like if the wife lets me buy Focal Sopras or PMC twenty five.26s?

Simon-in-Suffolk posted:

 

 

this is true, but then we revert to the scenario of unmatched encoding and decoding reconstruction filters in the ADC and DAC.  Now I know MQA is actually a pick list of different features that you can selectively use, so if one is simply using the 96/24 compression and expansion part of MQA format then a seperate software decoder as you will work fine... and you simply using MQA as a hidef compressor / audio enhancer algorithm... and it should sound nice for many MQA masters for many people's tastes.

Dont get me wrong I can see in part what MQA is addressing with its lossy hidef encodings, and I admire their patent, there are so I have discord other potential ways of achieving a similar outcome. One thing is for sure standard hidef PCM is very inefficient, as much energy and therefore stored info is simply representing the higher frequency bit dither with no musical or information content. Further this noise has a Gaussian distribution so doesn't compress with FLAC or ALAC.. it's very inefficient... MQA is one approach to address this.

Then surely the Tidal MQA solution is no(t materially) different to playing standard hidef files, which will not have been constructed using matched ADC/DAC filters either. 

I think that what you are saying is that it is a pity that the ability to match ADC/DAC filters that MQA offers is not available unless you have a MQA DAC, because this would take the SQ beyond normal hidef. 

I tend to think that the HW/DAC solutions have a down side as well as an up side.  Like Simon-peres says, they draw significant current close to the DAC  (and avoiding that sort of thing was one of the reasons given for why Naim prefers WAV files to FLAC).  Doing that work all upstream will insulate the DAC from that, especially if you pass PCM through a decent reclocker to remove much of the PC noise and jitter.  

No doubt an elegant solution that gives us the best of all worlds will be forthcoming.  I have a feeling it will be expensive and involve carefully engineered power supplies. 

We may have top wait longer for audio nirvana, but in the meantime it seems to me that Tidal's goal of being able to stream hidef music without increasing the size of the stream is achieved and this is a very good and welcome thing. 

The advantage matching of ADC and DAC surely applies to any digitisation process, not just MQA, but in reality is unlikely as people buy DACs on the basis of either budget or their assessment of sound, so an MQA DAC might be best if that is all you will listen to, but only IF in all other respects it provodes what you want out of it in terms of sound, but not if you want also to play other digital sources.

Also - though I don't pretend to be an expert on this and bow to anyone who is - doesn't this only apply where the MQA ADC has been used on original analogue source material, meaning that  in the case of music recorded digitally it is already 'stamped' with the profile of the ADC(s) used during recording, so for the best sound would need filters matched with that in the domestic DAC?

Yes, at present I have no intention of replacing the some 35k HiRes PCM/DSD tracks (2,761 albums, over half of the music on my server) I have collected and have available to play without any streaming subscription. Just like the rack of vinyl.

However where MQA does offer improvements is for the streaming services, and anything that gets away from 256/320 lossy streams has to be good. And shows there are better formats that CD, which was a technology compromise back in 1984, and maybe offers direction for the mass market than DVD-A & SACD missed.

What's also interesting is, once you have tagged the Tidal Master in the Desktop app, it is listed under 'My Music' on the Tidal iOS app, and plays back as 'HIFI' - but is it still going to be better than the CD format version - certainly sounds good over a pair of headphones while out & about.

So is Tidal going to update the iOS apps to enable MQA decoding on an iPhone, true mass-market exposure of HiRes.
Also what is Apple going to offer for HiRes over Bluetooth, given the lack of headphone jack on the 7?

MartinCA posted:

 

 Like Simon-peres says, they draw significant current close to the DAC  (and avoiding that sort of thing was one of the reasons given for why Naim prefers WAV files to FLAC).  Doing that work all upstream will insulate the DAC from that, especially if you pass PCM through a decent reclocker to remove much of the PC noise and jitter.  

 

Damn - mixing up two people there - I don't suppose Shimon Peres ever said anything of the sort and is unlikely to now.  I meant Simes-Pep, of course.  

MartinCA posted:

Then surely the Tidal MQA solution is no(t materially) different to playing standard hidef files, which will not have been constructed using matched ADC/DAC filters either. 

well materially it is different, remember MQA hidef is ultimately lossy... the algorithms including its associated decimation is discarding certain info.. now MQA says  this info has minimal useful musical information.. and their patent goes through this, but none the less info has been discarded and with MQA artefacts potentially added on 'unfolding' or decoding. Regular high definition just does do this, however it is exceedingly inefficient with respect to the contained audio information especially at 192/24.

MQA would appear to focus on temporal info more than frequency for a given resolution, and there is merit for this.. our temporal hearing and hearing of pitch appear separate and our temporal hearing is not effected by old age like pitch is.. the AES has some good papers on this.... but regular hidef PCM has frequency, noise and temporal information all in one inefficient but un altered soup... One has to accept in any processing such as with MQA where you attempting to extract info from the soup, you have created a subjective assessment and created a lossy function... and in the limit I have no issue with this at all and it could and should sound good... but it is not the same as pure unalderated hidef so it is materially different 

The other thing that it is interesting and happening elsewhere is that with new apodizing processors allowing larger more effective reconstruction filters ( such as from Chord Electronics and others) there is a lot more we can get from RedBook without ever going near hidef... we are in interesting times... so MQA is not the only show in town.. and they don't all require new masters or licensing 

Simon

Simon-in-Suffolk posted:

MartinCA posted:

Then surely the Tidal MQA solution is no(t materially) different to playing standard hidef files, which will not have been constructed using matched ADC/DAC filters either. 

well materially it is different, remember MQA hidef is ultimately lossy... the algorithms including its associated decimation is discarding certain info.. now MQA says  this info has minimal useful musical information.. and their patent goes through this, but none the less info has been discarded and with MQA artefacts potentially added on 'unfolding' or decoding. Regular high definition just does do this, however it is exceedingly inefficient with spect to the contained information especially at 192/24

Simon

Come Simon - no need to pick me up on that.  My use of 'not materially different' was meant to convey that I recognise that there are differences, exactly as you describe here and elsewhere.  But I defy the vast majority of people to be able to detect those differences, which is what I mean by 'materially'.  I can't and I have a decent set-up.  I am with MQA on this in terms of minimal musical information being lost.  Most of us have set-ups that are sonically compromised in far greater ways than this, and, as you point out yourself, the filter matching has more impact.  

Sorry, it's just some might think it was materially the same, I'm with you in understanding it isn't.