Now that Roon offers Qobuz integration it has become easy to do a SQ comparison between Tidal MQA or Qobuz HiRes. Making sure that the respective album release dates at both services are closely aligned (within 1-2 weeks) should also ensure that the releases come from the same master.
My setup
I have been playing around with this for the last days streaming tidal and qobuz over the Roon/AIR-integration (asynchronous, bit-perfect) to my Devialet Expert 220 Pro. I don’t use any DSP or filters and the sound is played over my B&W 802 D3 speakers. As the Devialet is not MQA certified the MQA files are unfolded within Roon (Only first unfold).
So, what sounds better?
To be honest my expectation was that Qobuz will win hands down as my Devialet DAC can not unfold MQA further and the uncompressed HiRes files from Qobuz should have the theoretical advantage.
That was not the case though. After comparing over 50 songs over several hours now (Classical, Jazz, Pop, Rock) I would have to call it a tie with maybe an even so slight advantages towards MQA.
As there is probably expectations bias involved (no blind test) I recruited my wife and younger daughter as well. They don’t have any clue about file formats and I didn’t elaborate on the setup at all.
I only asked them a simple question: does one sound better or do they sound the same? They were not forced to pick sides.
The results here was either a tie or slightly tilted towards MQA. Examples:
The Doors, Riders in the storm: MQA better -> sounds a bit more realistic (the rain drops)
Víkingur Ólafsson, Concerto in D Minor, BWV 974, 2. Adagio -> MQA sounds better as it is more relaxed
Margo Price, Midwest Farmer’s Daughter -> was a tie, no difference
So what are your experiences?
I would like posts only from people who actually listened and compared in a solid set-up. I would also like to exclude a general format discussion; I am well aware that MQA might have some lock-in effects and is not open source.
I’ve tried both Tidal MQA and Qobuz High Res. And I couldn’t agree with you more: it’s a tie. Sometimes the MQA files sound better, sometimes the High Res files sound better.
Definitely a tie here too … MQA tends to seem louder, but when you have a higher res file on Qobuz it’s quite excellent. They are also a tie in catalog depth so far so I can’t quite figure out if I can really ditch Tidal for Qobuz which I’d love to do as the mobile experience is so much better … not too mention I don’t have to clear past forced hip-hop selections every time I login.
I personally think the “lossless” vs. “lossy” discussion is totally besides the point.
Every digital signal converted from analog is “lossy” as it is not possible for it to contain all the original information…also HiRes PCM no matter what the sampling rate. Only data created originally in the digital realm (computer code) can be transported lossless.
So in my view the only valid question is: how good of a job does a digital codec do to convey the original analog performance?
That’s the main difference between Qobuz Hi-Res vs Tidal Master! Both are source from the same recording digital masters but Qobuz Hi-Res streams the exact master copy of the original, while Tidal Master goes to an additional processing to reduce the amount of bit rate it needs to transmit. In the end of the day, you can’t get the exact digital master copy (bit perfect) that is intended in the first place because data has been modified to such a high degree. Whether this translates to the difference in SQ, you have to listen for yourself.
This is exactly what Qobuz wants you to listen; the exact master copy of the original, nothing less, nothing more.
You have understand any A/D conversion is always based on finite representation of the analog signals and that’s your valid answer.
I’ve listened to MQA vs Hi-Res since it’s early conception and I have to say consistently I’m able to hear the difference even with recent Qobuz integrated into Roon. It is not matter of which sound better, rather they sound slight to some different depending on the music contents.
As listening is very subjective indeed, I prefer not to draw conclusion here. Given Tidal Master vs Qobuz Hi-Res, I will go for the unadulterated lossless streaming which is obviously Qobuz.
For me Qobuz easily wins. I’ve tried full MQA decoding on my ifi micro and also roon doing first unfold and sending to HQPlayer 128dsd.
Doing full decoding the difference between the two is smaller but I still prefer Qobuz. With roon doing the unfolding Qobuz is clear winner. Feels like a slight vail lifted, vocals are more precise, better sense of space, larger soundstage, and deeper bass. MQA is flatter in comparison and some seem tuned differently. Like a filter has been applied to give smoother sound.
I’ll be dropping Tidal once Qobuz goes live in the US, should be on the 14th. Just hoping most of missing albums get added as Qobuz promised.
Be advised that placibo effect is always lurking to mislead us. I changed something so it’s got to sound better.
The link leads through a non-mathematical review of the theory of digital signals and the practical implications and realization of digital signal recording and reconstruction to analog. Oversampling is used because the band-limiting and reconstruction filters realizable at the Nyquist frequency ring and introduce phase shift in the passband. Raising the sampling rate above the Nyquist frequency lets saner filter designs be used to band limit the signal being recorded and to reconstruct the audio for playback. Mike Moffat of Schiit talks about this in the interview article https://www.audiostream.com/content/qa-mike-moffat-schiit-audio
Many of the apparent benefits of newer high res recordings result from going back to the original tracks, cleaning them up, and remixing them more skillfully than the original engineer could given the studio tools of the day. The Beatles reissues show this in spades. They don’t sound better because they are 96k/24 but because Giles Martin had better tools than George Martin with which to equalize and mix the original session tracks into a stereo record. A better mix is a better mix however it is properly encoded for storage or transmission and however it is properly reconstructed for playback.
One minor quibble is that we can’t assume that Tidal MQA and Qobuz Hi-Res are starting from the same original digital master. I purchase High-res music all the time and regularly see labels, especially in Classical offering the same recording at 24/96, 24/192, and even DSD64, etc. So what was the original that Tidal MQA started from, and what was the original that Quobuz Hi-Res started from? Neither of them equal 24/192 in any case. And is either of them, then, still “lossless?”
I’m pretty shure that a lot of times the source both services use is completely different. Sometimes there is hardly any difference and sometimes it is a big difference. But i have heardexamples where MQA seems to be missing detail because it sounded rather dull and I hear many Qobuz hi res where I sounds like an upsampled cd file with some eq in the top frequencies.
I personally believe, to get the best result, that the quality of the source is way more important than the format used.
The next important thing is probably the person at the buttons when the transfer is made.
If you have both Tidal master and Qobuz Hi-Res as what I presently subscribed, it is easy to compare if both are from the same album, artist, label and sampling frequency.
It is uncommon but possible for music labels to have multiple mixed down masters for distribution. For example certain masters were recorded initially in 96k and later re-mastered in higher 192k sample even they are source from same analog master tape.
Lossless here is the way it is being stream from the master copy. The master copy (stereo mixed down) is always lossless. MQA takes the lossless master copy and created its own lossy copy for streaming.
I’m new to this. I have both Tidal and Qobuz (I just used NordVPN to sign up) and it streams fine to the USA without a VPN. I just spent the last hour listening to the videos. Monty is very compelling and easy to listen to. The net is, I can’t much tell one over the other. That is a comment on me as much as the setup. I’m using the Chord Qutest which does not do the MQA unfolding. I think I cared, but am now pretty convinced I shouldn’t. The part I was not aware of was the conversion of points to a SIN wave.
Way back early 80s I’d be amazed how my CAD system (Anvil-1000) could construct a mathematical spline through most sets of points. I would dot/scribble all over my CAD screen and sooner or later, mess that computer up, and some dots created a too-hard to do solution.
So I agree a perfect analog representation can come from some digital points. I’m not yet convinced all random noises the ear may hear can be reproduced mathematically, but the good news is I’m convinced I couldn’t hear the difference.
