They use analog.
If you’d like a high profile recent example, look no further than Pink Floyd (others would probably know better, but I think everything before A Momentary Lapse Of Reason was completely analog).
Xekomi, but they could still use a digital master made from the original analogue master, it seems like a big task to actually generate new digital masters?
Oh, thought your question was about wether MQA only deals with DDD recordings (if you remember those…). As far as I can tell, no one knows what they’re mastering from, other than vague marketing-speak, if that’s the question. No one knows who’s authentifying the masters, either, or which masters are used, or anything else of the sort… And don’t forget that UMG’s vault burned down in 2008.
Given the number of releases, I’d be surprised if they bothered with new transfers (I’d love to be proven wrong, though… not for the white glove releases, eh, for the millions of new tracks). And I’m not sure new transfers would even be of use in most cases.
I think there is only one MQA A/D converter (Mytek?) out in the wild so you have to assume almost all MQA is generated from a digital source. Add in the man power needed to do A/D is pretty heavy.
Yes they didn’t go back and redo everything from analogue tapes.
And certainly not for the 1 000 000 Warner conversions. It is very striking that all the new MQAs have the exact same track lenghts (expressed in number of samples) of the HiFi PCMs they replaced.
Of course you cannot check this anymore because they were so smart to remove the original PCMs while doing so (leaving us no chance to compare; you might wonder why and you should).
But the PCMs are the same as those on Qobuz. The FLAC headers can be different, but that’s all.
The MQA encoder for the 1 000 000s of Warner 16b 44.1kHz tracks was just a batch converter:
non MQA PCM in => MQA PCM out
With a few options like :
(1) studio light [on/off] (do warn the artist or he might leave like NY did)
(2) filter: [1… 32]
(3) MQA-CD [on/off] (warning : destroys the PCM even more)
(4) Wrap all in 24bit FLAC even the 16bit MQAs (they already did this)
For the “white gloves” they might have pulled out an analogue tape or two,
but for all the others not. They just used the digital copies (masters) that were already available.
In the beginning of digital era, some recordings were done digital only (no analogue tapes)
so there was nothing else to use. If you think of this:
If a recording was digital only 16b 44.1kHz; then there are no frequencies above 22.05kHz recorded
(Nyquist theorem); which means no region B; no region C; nothing to fold !!!
So what exactly are they folding then in 44.1 (and 48kHz) MQAs? The non-recorded bits ?
4 Likes
Good points, Wim - are they inventing new bits perhaps?
1 Like
I’m pretty sure bits only come in two flavours, quantum computing excepted 
1 Like
Nothing is impossible for the MQA crowd.
3 Likes
That reminds me of an old joke:
“There are 10 types of people who understand binary. Those who do, and those who don’t”
4 Likes
I already thought, let Hans Beekhuyzen explain it, then even magic becomes possible! 
4 Likes
MQAnon believers 
3 Likes
Stealing that
They must be, and you can see it in Roon when you turn on the “core decoding”
(core decoding = 1st unfold)
The signal path for a 16bit 44.1kHz MQA (so made from a 16bit 44.1kHz PCM)
suddenly shows 24bit 88.2kHz after the 1st unfold (why? to restore unrecorded frequencies?)
Even Gandalf from LOTR cannot explain.
Well, this is what “Bob” says:
“1. When MQA encodes a 16b 44.1kHz Master the resulting full MQA file is also 44.1kHz/16b. Despite being 16b, this file contains all the information for decoding and rendering. These MQA encodings also contain all of the information accessible when playing the original master and in some cases more.”
Now back to my sarcastic question: bearing in mind what is stated above by “Bob”, how can a MQA encoding of an ORIGINAL 16b/44 master contain “in some cases more” information?
That is why I think you do have a point here - and no MQA patent is able to explain it.
When MQA encodes a 44.1/16 file then they cut off 3 bits (noise floor) and use them for MQA information.
If you do the 1st unfold, the MQA information in bit 14-16 is cut off, it is upsampled to 88.2/24, but if you play it without decode/1st unfold, these 3 bits are garbage, hum, noise.
That’s where the statement comes from unprocessed MQA is 13 bit.
I hope I remember correctly.
As far as the sampled signal as a representation of the original goes, you can’t. Depending on your increasing levels of (dis)ingenuity I can think of a few explanations.
- upsampling/increase bit depth, there’s ‘more’ information but you’ve just derived it from the source, it’s not better quality but may aid downstream processing;
- channel coding, redundant data used to error check, again there’s ‘more’ information and it’s useful, but it’s derived from the source; or
- metadata, so new information about/relevant to the source, but anyone can augment.
One aim is (lossy) compression that loses/distorts some of the original information and without specifications it’s informed guesswork.
They refer to “white glove” treatments. Like on the 2L comparison website …
There you have an original master from a 16b 44.1kHz DAT tape. (so a digital recording)
And they made a 24b 44.1kHz MQA from it !
Real magic here, and still “original”?
In other words, and with due respect to both of you, this is just more MQA doublespeak, also known as BLABLABLA. Something that is originally mastered in 16b/44 CANNOT be sonically “improved” through upsampling - so Wim’s point above is correct: there is absolutely no sonic reason to upsample to 88. And even the exclusion of the so-called three “redundant” bits makes zero sense (other than for bandwidth purposes), since these bits will not make any sonic difference one way or the other.
Yet it looks nice on the screen, when you see that a file is now being “upgraded” to 88 as if magic would occur. And “Bob” adds to the obfuscation when smartly referring to “more information” as if the MQA processing would make the file sound better.
2 Likes
Neither original nor beneficial in any sonic sense.
1 Like
So for 44.1kHz MQAs (and same story for 48kHz MQAs)
First “unfold” = upsample x2 to 88.2kHz
Second “unfold” = upsample x2 to 176.4kHz or even more (*)
… all this to restore air. (Master Quality Air?)
(*) “Generally, MQA decoders automatically ‘unfold’ as far as the hardware permits.”
(bobtalks about mqa playback)
Well… he did use apostrophes around the word ‘unfold’ which probably means he wanted to say something else. After all, they must keep the curtains closed and the smoke inside the building.