I don’t mind DSD since I can very easily and quickly convert those files to PCM if I so chose. To date there doesn’t appear to be a (publicly available) tool out there (yet) that ‘unfolds’ MQA and saves or locally streams it as a standard 24/96 lossless file. Does anyone know whether unfolding an MQA file and saving it as 24/96 FLAC will violate the DMCA? If doing so doesn’t violate the DMCA, then my concerns about an all-encompassing MQA Hi-res ecosystem are alleviated to a large extent.
Unfolded MQA should be better than CD, after that get an MQA decoder.
Failing that, buy the 96/24 (or whatever) files as always.
The MQA and normal hi res file may cost similar amounts but you will have to make a decision as to what to buy when you have both options.
This dilemma should disappear over time. For me, I would buy MQA and wait for more decoding options down the road.
The undecoded MQA files I have heard so far easily match and 96/24 files anyway.
Just thoughts, what do I know? Lol
I don’t think there’s a need to ‘unfold’ it and save it to a ‘lossless’ 24/96 FLAC file, the end result file size is likely to be bigger.
Given that it is not a truly lossless, converting to lossless FLAC don’t seem to make it better. I would rather having it decoded on the fly to compatible decoder or simply buy the original 24/96 FLAC to avoid the hassle of finding one.
I would love it if by hook or by crook we get to a state where the MQA sign on a release indicates careful production. That to me is the key value of MQA. The other stuff seems more directed at licensing.
If 8bits are used, this only left 12bit of resolution in an undecoded state? This seemed to be a pretty agressive down scaling and one can claim it sound like 16bit?
24 - 8 = 12 ?
If one wants to use just 8 bits to encode the upper band (that is 50% of the total bandwidth), it assumes there’s not really much high frequency content to be encoded in first place. The way they seem to ensure this is:
- Use low-pass filter with corner at 30 kHz so there’s not much stuff left in the upper octave to encode
- Cut the dynamic range of the lower octave to 16-bit or less in order to have enough bits left to encode the upper octave
The result is of course far from being lossless compared to the original 96/24. It is more like 96/16. Now since this LSB encoded information becomes like white noise from FLAC encoder point of view, the efficiency of FLAC encoder used to encode the bundle is poor compared to encoding equivalent unfolded 96/16 version.
So overall:
- The process is lossy
- It is space-wise less efficient than encoding the lossy variant without the folding/encryption part
IMO, it is pointless exercise. Instead better result both from quality and transfer bandwidth consumption point of view would be to encode 96/18 version with just plain normal standard FLAC. And it wouldn’t be necessary to buy a new DAC to play such…
P.S. It would be interesting to send 0 dBFS 96/24 multi-tone signal and white noise through MQA encoder and decoder and compare those to the original. If they would allow doing it… 
4 Likes
Someone shoot me please.
3 Likes
Sorry my mistake, it should be 16bit, what I’m thinking about…
Does the 30kHz LPF applied to all incoming sampling frequencies including 24/192 and above?
The 2L samples I’ve checked that originate from DXD are really 88.2 kHz variant with that LPF applied.
1 Like
Where’s the music?
Is this a rhetorical question?
Nope. Seriously asking where ther is music to steam and buy.
HighResAudio.com has ca. 200 MQA albums, including some new ECM recordings:
[HighResAudio Studiomaster | MQA]
(https://www.highresaudio.com/studio_master.php?fids=153&cr=MQA)
Enjoy…
200 albums is a very small number. That’s a joke for a new format.
Not really - SACD / was less than that in the same period- this format is very much in its infancy - judgements in my view need to be reserved for now
1 Like
Thanks, my concern is if they indeed use one LPF with corner at 30kHz for all input sample frequencies, especially for 24/192 and above, it will have far greater loss than those done at 24/96. The higher the sampling frequency, more will be filtered by the LPF; this I suspect is to prevent over-loading the encoding/decoding(running out of the 8 bit to code upper band) process while maintaining a constant bit-rate of 1.5Mbps regardless the input sampling frequency.
It’s generally unwise to believe everything that you read on the Internet.
6 Likes
I have no problem with loss in regions not even my dog can hear let alone myself 
Why is it so important for you to have this in your files? Besides waste of disk space.
2 Likes