MQA arrives on Tidal

[Julf]

There are a number of folk on this and other websites actively trying MQA with a variety of software and hardware architectures. Between them thay are getting closer to deducing what is happening where, what the limitations are, and what they hear. They are then generously taking the time to share their experiences both good and bad. This has helped other folk to correctly configure their devices so spreading the experience and understanding.

It would be nice if you too were actively engaging in this.

And me sharing my experiences (and things like spectrum plots) isn't engaging in this?

If you are researching MQA in the wild, please share your set-up, your musical test pieces and your findings both technical and SQ-wise.

Ah, I see, I am not sharing exactly the way you want. My apologies.

 

[IceG]

Ah, I see, I am not sharing exactly the way you want. My apologies.

I'd very much like to understand your actual experiences with MQA over Tidal in a real-life scenario, like what the thread is about. It would flesh out the hypotheses which you have posited and add empirical data to justify your thoughts.

It would be nice of you as a member of a sharing community to do that, especially as you seem to understand the theory so well given your many insightful comments about the technology.

 

[mark-hants]

Are there any analogue specialists on here who can provide a comparison with top end vinyl reproduction? I would find that very interesting.

There's a Madonna album in the Tidal MQA section which I'm fairly sure that I also have on vinyl.

I'll be most interested to put the two side-by-side.

That's with a Thorens Td150MkII/SME3009/Goldring 1042 so not what you might call "top end" but it satisfies me.

Will report back.

 

[SteveS1]

SteveS1,

Sorry, that probably shouldn't have read as if the world had agreed that to be the case.

In my experience, I've yet to be able to find an audiophile version/pressing/download that hasn't been improved upon by MQA.

Apologies for the bad writing!

Chris

No worries Chris, this is all very interesting.

 

[Julf]

I'd very much like to understand your actual experiences with MQA over Tidal in a real-life scenario, like what the thread is about. It would flesh out the hypotheses which you have posited and add empirical data to justify your thoughts.

Not sure my subjective impressions really add much value. I didn't keep proper listening notes when I did a listening session at my Tidal-subscribing neighbour, but I will try to do a new session, with proper notes, at some point this week.

 

[darrenyeats]

No axe to grind - just a dose of curiosity and healthy scepticism.

Same here.

The point is that no ABX is required - the difference is enormous and it doesn't matter which alternative master to go back to. Original CD, High Resolution download, LP... anything. No matter what version I've got of some of those albums, no matter what time I've spent curating the very best audio version (Japanese versions, 180g pressings - you name it), MQA bests it each and every time.

It just doesn't make sense, and it makes me even more suspicious that level differences, EQ, cross-feed or some other trickery is involved to create a noticeable difference. Or invalid A/B as Jim suggests below.

So using these as a comparison with MQA versions may simply fall into a snag which mimics the one people have been talking about. That switching off MQA decoding may give you the results of playing MQA-altered data *without* it being decoded. i.e. *not* was fed into the MQA encoding, or indeed, 'Authenticated'.

There are good grounds for skepticism about these "clear audible differences" being a step toward higher fidelity, rather than something else. That MQA is closed and somewhat opaque does not allay any of these fears.

 

[Rodney gold]

I just did a test with 3 sources
I used Avishai cohen trio From darkness CD.. lots of dynamics , strong bass , delicate cymbals..a cd I know well.
My system resolves very well and my room is fully treated..I wound it up a bit..well a lot...

Roon sending a 24/48 file .. IE mqa file , no decoding

tidal playing the hifi 16/44 version of the CD

Tidal playing the unfolded MQA

Firstly I cued up the 16 bit VS the 24/48 and flipped between them instantly and seamlessly and the 24 bit was superior in just about all respects .. especially bass and slam , it sounded more open ,, the levels of both were the same .. the 16 bit was more strident ..
So it seems that just by streaming the 24/48 you get some of the "goodness"

Then I cued up roon passing thru the 24/48 and Tidal decoding it .. the decoded was noticeably different and IMHO a lot better ...

Then I cued up the 16bit version in roon and the tidal mqa decoded .. Ie compare "raw" to processed
difference was huge.....

So
MQA 2448 is NOT the same as the 16 bit version sonically wise .. It is not remastered however..
just streaming the 2448 unencoded gives you a SQ boost vs the same in 16/44
Decoding it takes it up a step higher.
Ostensibly adding a MQA dac should take you another step forward

I am of the opinion that every change is positive , hence my step forward / higher comments.

 

[mattgbell]

I understand that, but it is annoying that they keep making the same points over and over again in this thread about are you comparing like with like etc.

You seem be arbitrarily reducing the scope of this thread. The thread's title is 'MQA arrives on Tidal'. There's nothing in the thread's title (or indeed in the first post) specifically about listening to MQA.

I'd have thought any comments on the light that this new 'arrival' sheds on MQA would be relevant to the thread, whether they're theoretical questions about how MQA works or anecdotal reports of listening.

 

[Rodney gold]

Julf and all others..take up your queries with MQA..no one here can answer them for you ..we all stumbling in the dark and are not privy to the source code and workings of MQA or how they have processed the stuff.
I at least am curious enough to test various combinations to find whats best...the mechanism is not under my control , I have read all the web documents and seen all the videos so am not clueless about how it works , but there is nothing I can do to change or criticize it.

 

[Stunsworth]

A couple of videos I stumbled across over the weekend regarding how MQA works...

[YOUTUBE]https://www.youtube.com/watch?v=r_wxRGiBoJg[/YOUTUBE]


[YOUTUBE]https://www.youtube.com/watch?v=T5o6XHVK2HA[/YOUTUBE]

I'm not an audio tech, so remember, don't shoot the messenger.

 

[Jim Audiomisc]

Encoding a digital master that has been produced at digital fullscale, with MQA folding most likely adjust gain a little to avoid clipping once folded or on unfolding, as the process is not 100% reversible.
MQA state that the authentication verifies that the distribution is bit perfect, but carefully don't mention what the decoder does.
This gain change will be subtle but might be audible

That's an interesting point that hadn't occurred to me.

I'm familiar with the problem that downsampling and reducing the bandwidth can sometimes actually increase the peak level. Thus risking some resulting samples going out-of-range.

But the 'origami' might also do this because it relies on 'folding back' energy from the HF into the baseband.

Not something I've considered. But it does imply that it may be necessary to scale down some source material prior to MQA encoding to avoid a clipping problem. TANSTAAFL.

 

[Julf]

Julf and all others..take up your queries with MQA..no one here can answer them for you..

You seem to be in a habit of speaking for everybody.

 

[Jim Audiomisc]

Julf and all others..take up your queries with MQA..no one here can answer them for you ..we all stumbling in the dark and are not privy to the source code and workings of MQA or how they have processed the stuff.

I may well ask them about my questions sometime. However I suspect the bar will be that for me to really understand what they're *actually* doing to the audio info when MQA encoding/decoding I'd need to know those specifics of the encoding/decoding process which they have kept under wraps as proprietary. Without that I can't really analyse the system beyond using the open info they have disclosed in their patents and papers.

As it is, they've published enough about the 'origami' that we could see the risk of HF anharmonics being generated by that mechanism - and which Julf's spectra now seem to show. Beyond that, we may need to have the source material for the audio to compare sample-by-sample to know how to find any other potential snags or limitations.

And if they said, "We'll tell you under confidentiality" that would probably prevent me from explaining exactly what I might find out about any problems.

And underlaying all this is one rather more basic issue from my POV. Namely that *if* the point of MQA is to deliver higher audio quality/fidelity to the end-user via lower stream rates and file sizes, then alternatives like 'bitfreezing' do that OK *without* altering the actual audio information in ways that MQA does using 'origami'. So is more faithful to the source. Given this, the question becomes: Why do companies like Tidal choose MQA, which they presumably have to pay for it, whilst bitfreezing is free and requires no special decoders or DACs, or might generate artifacts like the origami can?

We're not entirely in the dark, Nor stumbling blind. But some significant details are being veiled.

 

[cporton]

You seem be arbitrarily reducing the scope of this thread. The thread's title is 'MQA arrives on Tidal'. There's nothing in the thread's title (or indeed in the first post) specifically about listening to MQA.

I'd have thought any comments on the light that this new 'arrival' sheds on MQA would be relevant to the thread, whether they're theoretical questions about how MQA works or anecdotal reports of listening.

Ok. Thanks to those who shared experiences. I think I'm out now and apologies if I've missed the point of the thread with my suggestions around splitting it.

Chris

 

[onlyconnect]

The point is that no ABX is required - the difference is enormous

If the difference is enormous, ABX will readily show that, so it is extremely valuable. We do need an easy way to compare like with otherwise like though and that seems to be a problem at the moment.

Tim

 

[onlyconnect]

Maybe so close that most peoples' ears will "accept" 16/44 though still wondering just why it is that very good vinyl, a dated, anachronistic and thoroughly flawed medium, can sound so marvellous in a way that 16/44 never could.

Did you read this?

http://www.bostonaudiosociety.org/bas_speaker/abx_testing2.htm

Tim

 

[Pedro83]

An album I have changed my mind on today is 'Blue' by Joni Mitchel. Comparing the MQA against the DCC master, the latter wins. This contradicts my first impressions but it is clear the DCC is slightly superior upon further listening. I tend to listen A vs B for a few minutes each, make notes. Then repeatedly listen to each album from start to finish which for myself at least often yields a more accurate finding. I try to let my subconscious do the talking/action and decision making when possible.

As many have indicated, this MQA has exceeded a lot of expectation, mine included although until now all I have had shoved down my throat is that is it a complete game changer. I don't agree with this in entirety but it could be much worse and let's face it, there are an awful lot of happy people speaking great things about it which is great. I like it but it isn't the complete game changer it was promised to be when I first heard about it. I am certainly looking forward to a much larger catalogue, it becoming available on the iPhone/iPad app, assuming it will (?). This would be fantastic but I haven't read anywhere when/if this is something to come in the future.

I can see how people who haven't looked for originals, out of print offerings from japan (the CP...'s), vinyl rips etc, thus comparing average pressings with MQA are excited by MQA as it is good, very good. It is something I could live assuming the catalogue goes well into six figures with but as someone who has been researching, comparing many variants of albums it isn't the end, IME.

As an Audirvana user I will be picking what betters my current library and add it to 'My Music' within Audirvana for a great way to listen to not only MQA but my cherished library from one place which is fantastic. I look forward to Audirvana 3.0, hopefully Damien will have managed a way of MQA decoding putting less stress on the CPU as this may be a problem for users with old machines. My 6 core Mac Pro is pushed to 20% and this is quite a lot all considered. No biggie for me but I would be interested to know what is the minimum required for MQA with the current app. I suppose offsetting the decoding to an MQA DAC would help this, there are obviously some extensive calculations going on during the decoding of MQA.

Anyway, the future looks promising and I look forward to it very much whilst it advances and more options are brought to the table.

 

[onlyconnect]

An album I have changed my mind on today is 'Blue' by Joni Mitchel. Comparing the MQA against the DCC master, the latter wins.

Thanks for the report.

Let's note that the DCC is a plain old 16/44.1 digital transfer that has not been "de-blurred", just mastered with care from the best available sources.

Tim

 

[onlyconnect]

Would be interesting to capture the output with an ADC as a plain old PCM recording.

Then compare the sound of the recording with the MQA stream.

Might try it, if I decide I can trust Tidal with CC info.

Tim

I've just done this, using two tracks that I know well. My sighted impression is that they sound different, ie. even after a second ADC the "Master" version sounds different from the "Hifi" version. Will do some more testing.

Tim

 

[gazjam]

Run Roon with Tidal Hifi and have been listening to the MQA versions of albums i know.
My M2Tech Young dac doesn't do MQA so comparing Tidal "master" albums to both 16/44 Flac and 24bit versions of the same album.

Tidal Master albums sound better than 16/44, but they would probably being 24bit?
Whether upsampling, MQA processing or whatever it sounds better to me.

24Bit Flac streamed from my hard drive sounds better to me than Tidal MQA interestingly enough.
Lots of variables (probably too many) to factor in, so just use me lug 'oles to compare and dont worry about it.

Its cool though geting better quality albums from Tidal, no wish to get an MQA dac.