There are two steps to decoding MQA:
1- “First unfold” - can be done either in hardware or software (currently TIDAL native app, Audirvana, etc)
2- “Rendering” (aka upsampling with MQA filters) - can only be done in hardware (not a must, just a licensing restriction)
You are correct that this means DSP cannot be applied post-rendering. Your only option (at the moment) for DSP post MQA decoding is to just do first unfold, then apply DSP - you loose the MQA-specific upsampling here, but did gain the first unfold.
Well, bass management would matter even if you’re moving around. The vast majority of the impact of room correction is to deal with room modes in the bass.
@brian: Is it an explicit requirement that DSP must be disabled after the first unfold?
James_I
(The truth is out there but not necessarily here)
64
Thanks - the explanation I understand. What I was pointing out was Brian’s statement that MQA Ltd. has, as a product guideline, a “requirement” (perhaps not always enforced? Not sure why Brian said they were going forward despite it unless it isn’t being applied to Roon) that when any software that performs the First unfold detects an MQA capable DAC, it must then let the DAC perform the First unfold.
Precisely why the claim that all of the albums on TIDAL received careful treatment rather than being put through a generic MQA sausage maker is so ludicrous.
Audirvana lets you choose. So no this is not the case.
James_I
(The truth is out there but not necessarily here)
67
I must have misunderstood. Or MQA isn’t enforcing this guideline. I just thought it was interesting that, if it was the case, an MQA-capable DAC would then trump MQA unfolding earlier in the chain, which would then impact the ability to DSP.
I can understand MQA Co wanting to restrict any form of decoding to cases where a payee has been detected (ie a DAC that payed their licensing fee). However, I think that cat is completely out of the bag at this point.
As I said, I am not going to get involved in speculation. You do seem to have missed my point about good product.
I’m not being obtuse because there’s bad news we’re holding back. I am happy with where things are going, and I think you will be too. There is just way too much vitriol and emotion surrounding this topic for us to be spilling details before this is released.
1 Like
James_I
(The truth is out there but not necessarily here)
70
I was reading this more as a flexibility-unfriendly rule. If that is not the case, I am happy for that. Nor do you need to violate any NDAs or preview a pending Roon release if that isn’t a good idea.
Yes, looking forward to that firestorm! (I am not anti-MQA as a format; I just don’t want it to be used as any tool of control whatsover…I also find the marketing hyperbole a bit silly but you can say that about 95% of marketing…)
Why on earth would you want to do signal processing before putting the DAC to work? The DAC is exactly the path component you WANT to have processing the digital information.
1 Like
James_I
(The truth is out there but not necessarily here)
72
Upsampling? Room correction? Bass management? There’s a whole group of Roon users that use all kinds of Roon DSP.
Then your DAC will not do it’s job. It’ll just go into pass-through mode. An audio DAC is made to process digital sound files, not processed signal files. Even if you change the file information from redbook to something that’s got a higher sample rate, it’l be downsampled in the audio device by a separate processor (just like mp3’s will be unpacked and upsampled using the firmware codec, flac files will be unpacked and so on) and then fed to the DAC chip(s) for conversion to 44.1 for analog or PCM output. That’s all.
Bass management? Post-DAC modification of the analog output. Room correction? Equalization on amplifier level, not in the DAC component of your system.
If you pass from Roon to a computer’s sound card and then to a power amp or headphones, yes, possibly. All the rest, well, that’s a bit like people feeding a PCM signal to an external DAC believing this will change anything. They probably don’t realize PCM is just the equivalent of an analog signal and won’t be reprocessed by the DAC. It’ll just be passed through just like the signal from an analog output (RCA-outputs) would if it were possible to find a DAC with analog inputs.
I believe its the other way around. Depending on the resolutions handled by the DAC an incoming signal of that resolution bypasses any upsampling stage in the DAC. That is the engineering theory behind HQ Player. So far as I know DACs don’t usually downsample as you describe, they just don’t play higher res signals than they are equipped to handle.
That’s the most worrying part; it has happened to Aurender A-10 music streamer reviewed by Sterephile. Only technical measurements are able to reveal that, so keep a look out for any audio magazines reviews.
Sometimes ago I did a critical listening using Mytek Brooklyn DAC, a fully certified MQA DAC decoder. My test involved comparing the highest resolution available; 24/352.4k DXD Master vs MQA fully unfold 24/352.4k from 2L test tracks. My observation concluded they both sounded different in sonic signature. DXD sounded clean, more dynamic with better projection of spatial and soundstage. MQA version sounded sweet, with a bit of ‘warm’ but I notice the spatial and soundstage are much narrower, a bit like sound coming out from a ‘small hole’. This is quite consistent from other reviews who have heard from a fully decoded MQA vs a PCM Hi-Res.
Listening is very subjective, some like this kind of SQ but I don’t, I prefer the original master, which I feel is closer to the original performance.
That generally would make sense except it is not true… For example, the Dragonfly is limited to 96KHz input but the ESS DAC upsamples to 384 with all streams (with specific choices for MQA filters when rendering - I hope it is not just sticking to one MQA filter). Many other DACs have internal upsampling exceeding input rates (EMMLabs, dCS, PSAudio, etc).
However, once upsampled to say 384KHz with an MQA style filter, it is hardly arguable that any upsampling after that would change the sound.
