MQA Sonic Comparisions

[tmazz]

So I had a little free time yesterday and my son was home so we decided to finally sit down and do some sonic evaluations of Tidal MQA files.  Since its introduction I have heard a lot of albums released by Tidal in MQA that sounded quite good, but never really had the chance to sit down and seriously compare the MQA versions to other releases of the same album. This is very difficult to do because most MQA releases have been remastered for MQA so there is no real way to know what differences are due to the MQA technology vs what ids driven simply by changes in the new mastering. But for what its worth here's what I found.

Before I get started, a quick word about the hardware involves so you can all have a baseline of what I was listening with.  Digital to analog duties were performed by an Eastern Electric MiniMax DAC. This is the original version and is completely as it came out of the factory with no mods or upgrades. Play it using the tube circuitry engaged and it is equipped with an NOS RCA 12AT7.  Digital discs are played via a Denon 2910 connected to the DAC via an electrical SPDIF cable from Audio Insurgents. Streaming from Tidal is accomplished via a Bluesound Node connected tot eh DAC with an Analysis Plus optical cable. The Bluesound does software unpacking of the MQA files and passes them on to the DAC as standard a 96/24 bitstream. LPs will be spun on a VPI HW-19 Mk IV equipped with an SME IV are and a Sumiko Bluepoint Special cartridge  played through the phono section of an Audio Research SP-9. As you can see there are a lot of variables here, so it will really be impossible to isolate the differences driven by the MQA technology itself, but rather what I will be reporting on is simply my experiences listening to MQA in my system as it sits. Not a prefect scientific analysis, but it’s a start…..

[tmazz]

The first album I tried was the new Steven Wilson remix of Jethro Tull's Songs from the Wood. I have both the CD and LP versions of this album and Tidal recently released and MQA version. Since all three of these versions were released in a relatively short span of time I would think it is likely that  they were all sourced from the same files and there is no large mastering changes among the versions, but of course I have no way to know that for sure.

I started by listen to the CD vs the MQA version. Both sounded pleasant and try as we might neither Bobby nor I could really find any significant sonic differences between the two. If we really strained we could pick out a few really minute differences, but they were so small that I don’t feel we would have been able to reliably identify them under blind testing conditions. The LP on the other hand was a different story. The differences there were not huge, but they were readily identifiable. The improvements were mostly in the sense of space in the soundstage and the detail around the notes. On the LP the instruments and vocal emanated more from a defined are as opposed to a point in space. This was most apparent on the reverb which sounded like it was echoing around the instrument instead of coming from the point of the instrument itself.  And on the acoustic guitar I the LP let me hear more of the body resonance in addition to the sound of the strings themselves. At this point it is worth noting that in the grand scheme of things I thing my turntable setup is much high quality that my DAC and if I had a better DAC, the differences between them would likely be smaller. But as I said earlier, this is nothing more than my experience in my system.

[tmazz]

The next album that I listened to was Supertramp’s Crime of the Century. This was the first mainstream album released by MoFi in 1978  (It was MFSL-1-005, the first four LPs  were Steam Trans and elevator music).  I always considered this album a benchmark in terms of soundstage imaging, right up there with DSOTM (although in Dark Side the imaging was more than just great SQ, it was also an integral part of the musical experience. So this in mind I was very excited to see tidal release an MQA version of this album, which they stated was sourced from a 2010 24/192 high res mastering.  Well what a disappointment this was.  Te MQA version was very flat and two dimensional. Instead of the wide deep and high soundscape produced by the MoFi LP, the MQA produced more of an impression of photographs nail to a wall that was in the same plane as my speakers. However when I pulled out my CD version of this album (which ia a stock A&M pressing from the 80s), the results were even worse. Not only was the soundstage flat, but the music itself was grainy as well. I was able to find a 16/44 verson of this on Tidal, which was labeled as a 2014 remaster and it was better than the CD, but not quite as good as the MQA. And all of the digital versions lagged way behind the SQ of the LP. It wasn’t even close.

[tmazz]

The last test was using Elton John’s Goodbye Yellow Brick Road. On hand I had The Tidal MQA version, The SACD Hybrid version (of which I played the CD layer) and once again, the MoFi LP. This time the versions were not all that far apart with one notable exception, and that was the size of the image. The placement of the instruments in the soundstage was not all that much different among the versions, but what came across as strange to me was that the size of the individual instruments seemed significantly smaller in the MQA version. While on the LP Elton’s piano on Funeral for a Friend seemed correctly proportioned in size compared to the other instruments, in the MQA version it sounded like Elton was playing a little toy piano, not in tonality, but rather in the size of the image. When playing the CD layer of the SACD pressing  the images seem a bit larger (although not as big as on the LP), but the height of the soundstage seen to end at about the height of my equipment rack, which sits 32 inches off the floor. This led to an unnatural sounding vertical compression of the soundstage. Listening to the LP the height seemed to display a much more natural proportion to the soundstage width.

[tmazz]

At this point I think I need to add that I have heard quite a few MQA files on Tidal that I thought sounded quite good on their own, but unfortunately I did not have access to other versions to compare them too.

[tmazz]

So where does that leave my opinion of MQA. Well I don’t think I did enough listen, nor was I able to isolate enough variables to form a conclusive opinion of the overall sound of MQA, but I have heard enough to convince me that this technology  may have some potential,maybe not as the ultimate sound quality standard, but at least as a vehicle that could bring better sound to more people. However, what I am convinced about even more after this experiment is that the mastering of an album has far more impact on the sound we eventually hear than does the storage and transport technology used to get it to us. It is the old garbage in garbage out theory. A high res format can do nothing to help an album that is not mastered, or remastered well. And as we have seen in the past with a lot of albums that are being sold in high res formats, just because the file contains more bits, it does not always follow that it will sound better.

[tmazz]

And this is by now means meant to be a final answer on this subject as the technology and the hardware supporting it is still relatively new and still evolving. But t rather I put this out there as a first stab at the topic and hope that others will jump in with their experiences.

[richidoo]

That's so cool Tom!
Thanks!

[rollo]

Tom the Eastern Electric DAC has MQA decoding ?


charles

[HAL]

Tom,
Is TIDAL doing the MQA decoding now?  Thought I read they had it working.

One thing from going to the MQA talks by Bob Stuart is the DAC reconstruction filter as part of the chain should be a minimum phase apodising style, or at least minimum phase.

I checked the ESS9018 DAC datasheet and there are two filter modes.  One is Sharp Rolloff and the other is Slow Rolloff.   There is no definition of the type of filter like Linear Phase or Minimum Phase styles.

You might see if the Slow Rolloff is used in the Eastern Electric MiniMax DAC.  Probably more likely to not be a linear phase filter.

Just a thought.

[tmazz]

Tom the Eastern Electric DAC has MQA decoding ?


charles

No it does not. The MQA decoding is being done in software by the Bluesound Node, which in turn passes the digital signal to the DAC in the 96/24 format.

 I spoke with Bill at Morningstar to see if there was a possibility of any updates to the EE DACs that would provide MQA functionality, unfortunately he said that they had no plans to do so.

[tmazz]

Tom,
Is TIDAL doing the MQA decoding now?  Thought I read they had it working.

One thing from going to the MQA talks by Bob Stuart is the DAC reconstruction filter as part of the chain should be a minimum phase apodising style, or at least minimum phase.

I checked the ESS9018 DAC datasheet and there are two filter modes.  One is Sharp Rolloff and the other is Slow Rolloff.   There is no definition of the type of filter like Linear Phase or Minimum Phase styles.

You might see if the Slow Rolloff is used in the Eastern Electric MiniMax DAC.  Probably more likely to not be a linear phase filter.

Just a thought.

The Tidal desktop app will do MQA decoding, but I an streaming to  my stereo via a Bluesound Node which is streaming from Tidal directly and doing the decoding itself.

I would think that the  the minimum phase filter recommendation applies to DACs that are equipped with MQA Hardware decoding. I wonder if it makes a difference in DACs used downstream from a device that does MQA software decoding, since the output of those devices is a standard 24/96 bitstream, which is not really MQA specific.

As for the EE DAC, as far as I know the filter type is not user selectable, so while it might be interesting to know, the filter is what it is and I would not be able to change it without buying a whole new DAC.

[HAL]

Tom,
Cool.  Thought I had read they have the software MQA decoder working for TIDAL.

With my dspMusikLCD crossover, the DAC's have selectable filters.

Those make a very big difference here.  My preference is for anything but the linear phase filters.  This follows Bob's talk about the MQA filters.

Nice comparison!

[Nick B]

Tom.
Very thorough and nicely done. For the reasons you stated, i never tried to do listening sessions regarding  MQA. Do you feel that Bob Stuart has provided enough information regarding how MQA works? I’ve read a couple of articles that were quite critical because of that.
Nick

[_Scotty_]

Here is a link to James Tanner's comments and conclusions from the Bryston circle' Post #93
http://www.audiocircle.com/index.php?topic=151245.80
My concerns about MQA are in regards to its best case recovered resolution of 17.2 bits – a loss of 6.8 bits from the 24 bit original and the potential for DRM problems.
 Of course DSP and crossovers are nixed by MQA from the get-go.
Scotty