After those lovely folk at Tidal released a whole bunch of MQA-encoded goodness, I felt the need to rush out and replace the trusty Explorer1 with a new MQA-capable Explorer2 ... which opened up the opportunity to conduct an A/B test between the software decoder in the Tidal Mac client and the Meridian hardware decoder in the DAC. Here's what I found.
Firstly, the equipment ... there's a 2015-vintage MacBook Pro with the Tidal client installed, driving an Explorer2 over USB, which has its analogue output connected to a Stax SRM-007t tube amp via a Kimber cable, in turn driving Stax SR-007 Mk 1 electrostatic ear-speakers. So it's a pretty high-resolution system - and I like the neutrality of the Stax electrostatics because they're a good match to the innate "cleanness" of the Meridian DSPs in the living room.
In all cases, the Explorer was set to use Exclusive Mode and to Force Volume in the Tidal settings, so the client is just passing bits to the DAC, with analogue volume control on the Stax amp.
Then it was a matter of queueing up some music from Tidal and letting rip, whilst using the Tidal audio settings to toggle the Passthrough MQA switch. This changes the MQA unfolding from being done in software on the Mac through to being done in hardware on the Explorer2. I used a mix of albums I know really well, and which I have local Redbook and hi-res versions of for comparison purposes.
The short version: the hardware decode is materially better. You can hear the difference best on very complex music, or strident guitars or piano, particularly on albums that aren't that well recorded - the hardware decoding simply untangles everything better. There's more space around the instruments, the localisation and separation is better, and the mix sounds more natural. Badly recorded albums are less tiring to listen to; brushes on hit-hats no longer sound like white noise ... audience members on live albums sound like actual humans.
The difference is rather like transitioning from 16/44 to 24/96 ... definitely the same music, not a vastly transformational change, but very much worth having. And it became more apparent the longer I listened to the hardware decoding - switching back to software decoding added a subtle blur and sheen to things. So that was a few hundred dollars well spent, then.
And just to extrapolate a bit, in a completely unwarranted way ... in my view, software decoding of MQA (hopefully in the next version of the Roon client) is going to be a big step up in my legacy G68/DSP5200 system. It will give me perhaps 90% of the benefits of MQA at basically negligible cost, so it's an utter no-brainer to do this.
However, the Explorer2 (which, let's face it, is still a budget DAC) graphically underlines the fact that hardware decoding brings a bunch of extra goodness to the table. I'm assuming, therefore, that the improvements that would come from an 818v3 and a nice set of 5200SEs would be even more marked, and when the opportunity comes to do the requisite upgrades, I will definitely head in that direction without a moment's hesitation.
As always, your mileage may vary!
Cheers
Office: MacMini + Roon -> MS600 -> AR LS-9 -> M60s (1993 vintage) Home: [AppleTV -> HD621] + [Roon -> 218] -> G65 -> DSP5200HC + DSP5200s + DSP3200s Being a JAFA in Auckland, New Zealand
Great post Kent. Later this week I'll try a decoding comparison between Tidal and my 818v3. However, I wonder if my venerable but rebuilt 6Ks will be as revealing as your Stax headphones.
Ken
Last edited by Kswanson; 2017-01-1004:44.
Main: G68J, 5K2s, 5Ks, HSU subs, Pio 151FD, Dish VP722, Roku, Wadia I70, Anti Mode Dual Core, Oppo 103D Vanity Lite Office: SBT, Berkeley Alpha DAC 2/MQA, Magico S3 Mk II, Boulder 2060, Lumin U1 Mini streamer, Roon.
whilst using the Tidal audio settings to toggle the Passthrough MQA switch. This changes the MQA unfolding from being done in software on the Mac through to being done in hardware on the Explorer2
Thanks Kent in Welly for this most insightful and helpful post.
The quote which I have extracted clarified things for me perfectly and explains many of the confused findings reported on other web forums.
It suggests a potential MQA capability hierarchy within a Meridian system architecture thus:
(1) software decoding externally in Tidal, Roon or (TBA) Sooloos into any Meridian device (except 818v3) = MQA minus anything above 24/96 and minus DAC correction to non SE DSPs or analogue speakers
(2) hardware decoding in the 818v3 = full MQA on analogue outputs, and MQA minus anything above 24/96 and minus DAC correction for non SE DSPs (*)
(3) hardware decoding in the 818v3 plus SE DSPs = full MQA
I cannot see where external software decoding into SE DSPs via a Meridian or other device would fit - this is an important question for G-series and 861 owners.
(*) hopefully the "average DAC" settings in the MQA decode will approximate to the Meridian non-SE DSP DACs
Last edited by Gianni; 2017-01-1009:31.
In the spirit of Spinaltap, I would like to confirm that I am not involved in the corrupt theft of a global sport.
It should be possible to get full fat MQA into Meridian active speakers too e.g. M33 & M60. I am considering the following for an M33 pair and M1500....
Main system - Sources (Kodi-HTPC, Vero 4k, FireTV, Sat) into Trinnov Altitude 32 (8-16) (+Roon), Trinnov Amplitude 8 amp & Vivid Audio (Giya G3, V1.5) and REL Stentor III speakers. Others: F80 + Raspberry Pi and other Pi systems.
Exactly right Ian. Though it will be interesting to hear from dealers who are able to make comparisons as to whether ultimate MQA is worth the considerable price difference.
Ok, thanks. I think most modern, quality power amps would fit that spec.
Main system - Sources (Kodi-HTPC, Vero 4k, FireTV, Sat) into Trinnov Altitude 32 (8-16) (+Roon), Trinnov Amplitude 8 amp & Vivid Audio (Giya G3, V1.5) and REL Stentor III speakers. Others: F80 + Raspberry Pi and other Pi systems.
Remember that much of the full fat of the SE speakers is down to the high resolution amplifiers and beryllyum tweeter.
Se we are back to the capabilities of the analogue signal path.
That's interesting and something that I hadn't considered. Should components in the chain be MQA certified in a similar fashion to THX? Presumably the legacy DSP speakers fail the MQA test on the analogue and digital side. Will headphones meet the requirement?
If this is the case, I can't wait to hear how fantastic my 5200SEs will sound when connected to a 218 as my old Meridian in walls sound fantastic and are noticeably different (improved) when fed with MQA through an explorer 2.
I am a bit more MQA confused now but thanks for the direct s/w to h/w decoding comparison. I am surprised at the release of a software decode (assume it includes a generic DAC profile)at this stage of development as it seems to short circuit any h/w solutions. However I am still hopeful/expecting a non-SE DSP solution which includes a Meridian DAC profile to get low fat MQA out of non-SE speakers. Will the bit rate limit of the DACs in these speakers limit the MQA resolution theoretically available to such an extent that an analogue solution would be better is what I would like to know.