Do I really need an MQA-compatible DAC for MQA tracks?

Roon Software Discussion TIDAL
1

Now that the MQA community is growing…

Has anyone done a careful listening test with MQA tracks playing through an MQA DAC in two different modes:

A. Roon/Tidal unfolding only, and play as PCM 24/96

B. Pass through Roon/Tidal and unfold+render through the MQA DAC

Does B further improve the sound? If so, is it consistently the case for all MQA tracks or only for higher res MQA tracks (e.g. 24/192)?

Please do not comment if your comparison was between different DACs. Please do not comment if you don’t believe MQA.

I still don’t have an MQA compatible DAC, so I need some statistically significant evidence to buy one. My humble BorderPatrol DAC SE is so glorious that I won’t be easily tempted for another DAC.

So far, my impression is that the quality of the master (or ADC process) determines most of the sound quality, so software unfolding may be sufficient…

1 Like
2

I have a Meridian 818v3 + DSP8000SE speakers.
It can do full MQA but when i do Roon unfolding the Meridian does not do the final unfolding, so this exactly the test you are asking about.
I found no significant difference.
For speakers (including other rooms) I use DSP room correction so I have to do Roon software unfolding. For headphones I could use full MQA mode but it made no difference.

As a consequence I have not made MQA capability a requirement for other gear I am buying. E.g, I have some Chord DACs.

4 Likes
3

Maybe I wasn’t using the right filter, but I did not like the sound of MQA unfolded by roon and then upsampled by HQPlayer. Qobuz sounded better in this setup.

I dropped Qobuz because of limited library so now I’m using my ifi dsd micro in MQA render mode and the MQA tracks sound much better. Can hear clear improvement with fully unfolded files. I still feel Qobuz hires files upsampled by HQPlayer beats MQA fully unfolded but will wait until Qobuz has better library.

4

I’m fairly convinced that “careful testing” is an oxymoron as far as sound quality is concerned.

For what it’s worth (not much), I think your observation is, at the Brits say, “spot on.”

However, there are numerous intelligent, articulate folks that swear by MQA.

1 Like
5

I’m going to try and withhold my MQA bias here and give you a technical answer based on my understanding of MQA…

A MQA file uses some of the 16 bits of redbook to hold info about the file. It actually reduces the “CD” quality of a file down to 14 or 13 bits (somewhere in there I forget the exact number).

The first unfold (which can be done in software and is what Roon is doing) takes those bits and reconstructs, at best, a 24/96 kHz file. You get the “MQA mastering” as well as the reconstructed 24/96 higher resolution file. It should be noted that anything beyond the ~22kHz frequencies is reconstructed from a “lossy” combination of compression and CODEC magic. The high resolution portion of the file does use compression to reduce file size.

Adding a MQA DAC includes the proprietary MQA filters in the DAC as well as an additional reconstruction (unfolding) of up to 24/192. There is a hardware requirement here because there is a decode operation which cannot be accomplished in software in real time.

It’s this additional decode to the 192 resolution plus the MQA filters in a “MQA DAC” that will resolve the entire file. Do you need a MQA DAC? It really depends on if you are onboard with MQA and like the way the MQA filters sound. A comparison in the method you suggest may be problematic because no two DACs will sound the same. Your question is one of buying a MQA compatible DAC or not. A MQA DAC means it’s using MQA filters so it’s going to sound different than a non-MQA DAC and that is regardless of the files resolution. Basically, I’d find it very difficult to compare a second unfold to the first unfold as there are too many variables involved during the second unfold which will color the sound. Either you’ll like the way the sound gets colored or you will not. That has everything to do with the DAC + the DACs reconstruction of the compressed high resolution + the MQA filters at the second unfold.

But, the short answer, if you’re all in on MQA you’ll want a MQA compatible DAC. If you’re already getting excellent sound and are not ready to be all in on MQA then I suggest you keep what sounds good. Technically, of the two, a 24/xx file vs. a “MQA’d” 24/xx file that’s been folded and now needs to be unfolded and with both files coming from the same master… the 24/xx non-MQA should sound better as the high resolution (those beyond ~22kHz) has not been compressed. However, and again, applying the MQA filters may be preferred and therefore can sound better to some listeners.

There is also the question of how you get your music? If you primarily stream Tidal and want the highest resolution possible then a MQA DAC is required. Tidal is all-in on MQA and that is their high resolution format.

Personally (bias re-engaged): I am currently shopping for a new DAC and have decided that MQA is not a requirement for me. I prefer, purely on a technical level, DSD or PCM without the MQA process. However, a handful of DACs I am considering include MQA anyway but it isn’t influencing my purchase decision.

Good luck and happy listening.

16 Likes
6

I thought I made the question pretty clear but still people go tangent. I don’t need a long technical thesis which I can find plenty online. My specific question was: in your subjective experience, does that extra MQA reconstruction filter beyond the first unfold make a noticeable improvement in SQ? I just want to sample statistically meaningful data (say a dozen) to make up my mind whether I should give it a try… Thank you.

2 Likes
7

The Tidal MQA files are 24 bit in my experience. MQA Cd would be 16 bit.

1 Like
8

If you go all in and get Meridian Gear like Chris has with a DSP5000 or better, then yes it will make a difference. If not, then no.

But, statistical data pulled from some people in a thread is no way to choose to spend a lot of money. You should listen to what you are going to buy first.

2 Likes
9

People are only enjoying the debate here, it’s supposed to be fun and mostly is. Personally, I enjoy MQA through my Meridian DSP SE system and a Bluesound Pulse 2.
The decision as to wether it’s worth it will be an individual choice. I would suggest spending time with MQA not the A/B stuff. Then go back and see if you miss it.

2 Likes
10

In my opinion the MQA files do sound noticeably better fully unfolded. The highs are cleaner and soundstage is more open. Just unfolded with roon and upsampled sounded duller. Thought regular non-MQA files sounded better upsampled.

4 Likes
11

You really need the second ‘unfold’ to make sense of MQA.
IMO, just listening to the first unfold without the second makes no sense at all, and if that’s the case you may as well avoid MQA altogether, and stick with lossless FLAC etc.

2 Likes
12

Why doesn’t it make sense when most MQA tracks are originally 24/96 masters? Even Bob Stuart claims that the software unfold will get much benefit.

13

I liked first unfold when I only had that for sure. Full MQA sounds just right here with the stuff I listen to.

14

Because just listening to the first unfold on its own will bypass the second ‘unfold’/filters/decoding. It just doesn’t sound ‘right’ IMO.
If you’re not planning on using an MQA DAC, then as I’ve said before, I’d just avoid MQA and stick to lossless FLAC. Alternatively, if you plan to listen to MQA, then I’d recommend listening via an MQA DAC/decoder.

1 Like
15

From Archimago, in 2017… i.e. no breaking news here:

"Objectively with the songs I examined, the software decoder works well to reconstruct what looks like the equivalent 24/96 download."

and

"Bottom line: TIDAL/MQA streaming does sound like the equivalent 24/96 downloads based on what I have heard and the test results"

https://archimago.blogspot.hk/2017/01/comparison-tidal-mqa-music-high.html

This applies to the 1st unfold only (up to 96kHz)… he’s done plenty of analysis on the stuff after the 1st unfold, which doesn’t need repeating of course.

So I don’t really understand the comment:

1 Like
16

Because accessing a second ‘unfold’ using an MQA DAC will finish the decoding, beyond 24/96 to 192 etc.
If you’re not going to use a second-stage unfold, I’d stick to lossless FLAC, IMO.
I’ve listened to first-unfold MQA, and haven’t been impressed with it compared to lossless FLAC.

17

Post was corrected shortly after I posted.

18

You probably weren’t comparing the same mastering… I’ve done similar objective analysis to Archimago… it’s quite easy to do the digital capture of the Tidal 1st unfold…

"Objectively with the songs I examined, the software decoder works well to reconstruct what looks like the equivalent 24/96 download."

and

"Bottom line: TIDAL/MQA streaming does sound like the equivalent 24/96 downloads based on what I have heard and the test results"

19

Yeah, you posted those conclusions from that article twice. I can read… :wink:
And that’s just his ‘take’ on it. We all have our own.
And I’ve compared several recordings that would appear to emanate from the same masters. In all cases, the second-stage MQA unfold was needed to really compete with the lossless FLAC stream. After second-stage unfolding, it’s often very difficult to tell the difference between MQA and lossless FLAC, and I found myself preferring the MQA stream in some, but not all cases.
With just a first unfold, there’s really was no competition IMO, and lossless FLAC won every time. Although I appreciate that these findings could by system-specific.

20

Not sure if you missed it but he used the word ‘objectively’… as I mentioned, it’s quite easy for anyone to objectively verify…

"Objectively with the songs I examined, the software decoder works well to reconstruct what looks like the equivalent 24/96 download."

If you are talking about subjective preference that’s fine but my original quote was of your words:

"You really need the second ‘unfold’ to make sense of MQA.
IMO, just listening to the first unfold without the second makes no sense at all, and if that’s the case you may as well avoid MQA altogether, and stick with lossless FLAC etc."

That didn’t make any sense to me, respectfully, but I’m probably missing something.