Ah, there we go: Ethan Winer’s Null tester.
3 Likes
How does that make sense, when in the next sentence I reference comparing tracks with their Qobuz equivalents as well as Qobuz downloads?
Yes, but I didn’t have that information from reading your original post, which, just to be clear, is what I was replying to. I was going off this:
I guess that would sort have scientific value, if a bizarre way to conduct one’s life, but with Warner developments, playlists have had certain tracks upgraded to MQA, but I guess you knew that. So in that instance I am not conducting a double-blind trial on myself, but just listening and thinking ‘that sounds really good’ and when checking it had been upgraded to MQA.
1 Like
When you make assumptions about Tidal, are you really open-minded? The bias is there from the outset.
Devialet Expert 1000 and Hugo 2 relay on Roon decode; Meridian Ultra Dac, decoding and rendering. Weirdly, I rather like the Roon decode with Expert 1000 Pros as much as the Ultra Dac.
I do, Andrew. When I notice something unexpected, spontaneously, I consider it has value. I try in all matters to try to maintain an open mind, even when reading your posts.
Gosh, where does one start when considering those that affirm their opinion is a matter of science! Still, Joseph Priestly, let alone Newton showed that scientists were complex flawed men, and if I were amongst them, I would be proud.
But the sheer level of dogma and certainty that pervades the MQA critique is poor science. Indeed it is the certainty and revulsion of any alternative view that is the realm of belief and not science, and that way lies folly.
And so I express what consider is true, and you will respond in due course…
I think you’re only assuming I was making assumptions about Tidal. As far as I can tell, I wasn’t. I was advancing an hypothesis.
I have mixed (entirely subjective) feelings about tidal MQA.
I am dubious a like for like bit rate / sample rate of MQA vs. PCM is going to be superior. If there is a perceived improvement then surely one could get the same with filters applied to a regular flac file. You can’t add musical data that isn’t there.
This suggests the 16/44.1 MQA files are a bit of a scam, if they are simply converted redbook. Though if they are actually taken from higher res masters and MQA’d to 16/44 which unpacks to 24/88 then maybe it’s not a scam, though this does seem unlikely considering the speed that Warner is adding them, and considering many only exist as 16/44.
On the other hand for music that is available as MQA 24/96, 24/192 or even 384, then this does suggest tidal is offering a better quality then they had previously. Even if anything over 24/96 is lossy compressed, that’s still roughly 24/96 plus some extra compressed bits, it’s hard to deny there is more musical data there. And for those of us that don’t have access to Qobuz then this is an improvement over what was previously available.
And yet, if I had to buy MQA files, I wouldn’t. I don’t want to buy into a proprietary file format.
I live with them, am happy to get them at no extra cost, but would likely switch if an alternative existed.
Edit. Also I believe it’s only a matter of time before a tool is released to fully unfold them in software at which point it would be easier to make definitive comparisons.
1 Like
Hello, I would like to precise what I think is a misunderstanding on the mqa format and the ‘missing bits’.
I will try to explain how these missing bits are lost and why I think they are not a real problem (I will try to make it as little technical as possible).
As we all know the bits will give you the accuracy on which a sample is recorded, 44100 or more times per second.
Let’s take the example of a TV. The bits mimics quite well the resolution of a TV: 1080, 4K, or 8K.
Generally, when we say we reduced the bit numbers that means that basically the loose accuracy of recording or to take the image of a TV, we have bigger pixels, meaning we go neared to 1080 than 4k.
This is not the case for MQA.
In MQA, the sound that is recorded is in 24 bits, but some very low level noise are not reproduced. If we translate that in a TV, that means that what you see will be in pure 4k or 8K. But, some parts of the screen won’t be reproduced. The part of the screen reproduced will be EXACTLY 24 bits.
Now, the parts of the sound that are not reproduced are part that are supposed to be beyond our capacity to hear. It is your personal opinion to think if it is to he trusted or not.
I personally trust but thats only me. And myself also I must admit I prefere MQA over just PCM (a very very very old technology with it’s own limitations by the way).
1 Like
MQA. I sometimes like the sound of it. I sometimes don’t. Depends on the recording, the mastering, and the time of day. I know it’s much better than Mp3 and the like, and MQA 24/44 sometimes better than 16/44 PCM. I do prefer minimum phase filters that MQA/Bob Stuart pushes.
It’s fun to have the option to try out new technology, that’s for sure.
2 Likes
But that’s the weird thing: they remove parts that they think one cannot hear, just to replace it with sounds that we know you cannot hear. In your analogy that would mean that we sacrifice resolution of the image to replace it with UV information.
3 Likes
Ok, so maybe we can set a protocol up.
Maybe if the curious were to give a list of files that sound better as MQA pre second unfold, an enterprising soul with Audio Hijack, a Tidal account and a Qobuz account could kindly capture them to file, at roon output. These could then be piped into Audio Diffmaker.
This should provide a rather good explanation as to what’s going on, if anything, and one that I’d guess most can agree on.
For those who argue that the answer can only be with a second unfold, or in the analog domain or whatever, I’ll let you get in touch with Ethan Winer, and ask him under what conditions he’d be willing to null test at DAC output.
5 Likes
Ok, sorry if I was not clear.
You don’t loose resolution.
Thats the important point you need to understand.
You don’t.
If your TV got a few dead pixels you don’t loose resolution. The part of the image rendered will be rendered exactly as it should.
What will happen is you loose a part of the image.
Now about what parts of the image is lost and what is gained in MQA:
What you gain is frequence and resolution.
In a reasonable size you can get 192/24bits when before itwas not possible.
What you loose is very quiet sound, I dont think it is a problem. But you can think differently, it is a question of taste.
Another point i would like to highlight is how MQA is sophisticated compared to PCM or DSD. They are both 40 to 45 years old technologies. Can you imagine that? There was a bit of refresh by pushing the specs but basically far better can be done…
To me having an highres format that try to save place makes sense. A lot of people disagree, I know. But the truth is if you like music. You will never find what you are looking for in DSD or PCM high-res. We can be sad about that but it is a fact. MQA with Tidal managed to be far more present in the market in little time than other high-res formats.
1 Like
I think that most people here perfectly understand what MQA does, and doesn’t do. Maybe something is lost in translation…
Lossy codec: codec avec perte
Losless codec: codec sans perte.
MQA is a lossy compression algorithm: le MQA est un algorithme de compression avec perte.
To put it into picture terms, MQA is taking a 4:4:4 signal, and degrading the chroma to 4:4:1, before dithering it back to 4:4:2 (first unfold), then 4:4:3.66666667 (2nd unfold), with a sharpness algo thrown in.
You can argue that it’s irrelevant, because nobody needs 4:4:4, and nobody sees billions of colours, and the added sharpness is good, but that’s about as far as you can go. You do loose chroma resolution.
6 Likes
The lossy was relevant for mp3.
Not much for mqa.
If you downsample from 96/24 to 44/16 you loose no?
The lossy is only relevant from your source.
If you take an analogic recording, a 44/16 will loose more than an mqa because of the analogic to digital process.
Now is a pure dsd 256 better in theory than mqa?
Oh for sure, I fully agree.
Do I care if there is literally no dsd available?
That’s the point.
Size does matter.
1 Like
Where’s your data, and what allows you to so unambiguously state that this is a technical issue that has to do with storage sizes and not a licensing issue, where studios are maybe, just maybe, more comfortable in distributing degraded files (not to get into the whole DRM argument) ?
(Really, how much HiDef does Tidal have compared to Qobuz and Amazon, and, since you claim an attachment to sophistication and compression, let’s keep aside the abomination that is 16/44 MQA ?)
1 Like
Neither would it be for lossless formats with comparable performance to MQA, which is is about as efficient as FLAC would be at the equivalent truncation (so 17 bits). And that’s not even taking into account the more efficient, but computationally intensive formats like Monkey’s Audio.
MQA: it’s what plants need, because computational power and bandwidth are SUCH massive problems on modern mobile devices.
I think that is a good idea, as so much of the debate revolves around bit rates etc, rather than examples of what sounded good to that listener.
My original post was related to the fact that previously I was rather more preoccupied with bit depth and sample rates than what the subjective experience of listening to a certain version was.
I have started to question whether the focus on the ‘origami’ and unfolding detracts from other aspect’s of MQA that actually make a greater difference to the listener’s experience than bit rates. If so, it does radically change a great deal of what we think is important about PCM.
So the 16/44.1 MQA albums that I find do sound so good, for me, raise questions about the ‘gold standard’ of PCM.
But before we can even get to that, it does require an attempt to get pre-conceived ideas out of one’s head, to reduce expectation bias. That is why if I notice something when I may be focussing on other issues I trust that any preoccupations with bit rates are peripheral.
I hope this is making some sense…
Before waxing lyrical, have a look at what diffmaker does: if there is a difference, we’ll know, and we’ll also know what it is.
The reason I suggested that is that it leaves little space for pre-conceived ideas. Since you mentioned that Roon’s first unfold is good enough for you, maybe you’d be willing to stick your neck out and start giving examples 
.
I will note that for the more… enterprising, rumour has it that there are illegal rips of Tidal material available, complete with MQA auth. I wouldn’t be surprised if some of that material were also available on Qobuz in unadulterated HiDef.
Now, for the MQA faithful, would such material be acceptable to you, or is it something that’d be a problem in your mind ? Given that MQA’s authentication has been proven to be easily defeatable, and that an evildoer might’ve messed with the files, I’d completely understand if you considered this to not be hallal, of course.
1 Like