I am sure you are aware that downloadable or streaming files can go through the process of conversion from analogue to digital and back again more than once before the final processed master file is created. Manipulation of the signals between these processes do impact the recording and can degrade the original audio capture.
Why do you think that "streamable files can go through the process of conversion from analogue to digital and back again more than once before the master file is created"?
The fact that the fingerprint of legacy ADC's & DAC's is known and corrections can be made to account for them is an established principle.
I'm not sure what you mean by "legacy ADCs/DACs"? Top studio ADCs/DACs were audibly transparent over a dozen years ago, cheaper (project studio) ADCs/DACs were audibly transparent by about 8-10 years ago and I don't know of any commercial studios which use ADCs/DACs older than these. In other words, there is no "fingerprint" of studio ADCs/DACs and so obviously, there is nothing to "correct"! Stating otherwise is certainly not an "established principle".
When you question file size compared to what, you must know the answer to that if you have been following the thread. MQA reduces the file size of a hires file to that similar to a CD file for transference purposes and restores it to full bit perfect PCM on decoded playback ...
Yes, I do know that but this highlights the contradiction in your posts. On the one hand you effectively state the principle that what happens below the noise floor is not "describable" (is "audibly indistinguishable") and is therefore irrelevant, a principle I would agree with generally by the way. On the other hand you state that MQA reduces a hi-res file to CD file size and then restores it again. These two statements contradict each other because a hi-res file is audibly indistinguishable from a CD format file. In other words, just converting a hi-res file to say 16/44.1 is audibly indistinguishable to start with, there is no audible benefit to restoring a 16/44.1 file back to hi-res. Furthermore, the better lossy codecs which already exist also achieve "audibly indistinguishable" but with file sizes smaller than MQA promises. Therefore, if you are talking about "audibly indistinguishable" and file size, MQA offers no audible benefit and a larger file size, the opposite of what you appear to be arguing!
As you are clearly quite clued up on recording studio technology, if you can spare the time, (I am not being cynical), please read the computer audiophile Q&A article and tell me what answers are not correct, what factual inaccuracies can you identify. I would be very keen to know as I again, don't understand all of it.
I could go through the article point by point but it would take many hours and many pages. So I'll deal with a few specific points and generalisations of the inaccuracies:
"
The steady dissolution of the album as a creative work and reduction in inherent sound quality from vinyl or CD to MP3 is more than unsettling" - This is inaccurate. A high 320 MP3 has far higher "inherent sound quality" than vinyl. However, there's no doubt that there can easily be an "actual" reduction in sound quality between many 320 MP3 releases and many older vinyl releases. This is because, as the article sort of mentions, the current economic climate of the recording industry dictates far less resources (time, money, talented personnel, facilities) are invested in making a recording today than in the past. This can (and almost always does) impact at least one (and commonly every!) aspect of the music recording creation process; composition, arrangement/orchestration, performance, recording, editing, mixing and mastering! Given the same resources as was once the case, a modern 320 MP3 release would have a higher sound quality than vinyl! In other words, any sound quality reduction is due to the quality of the (resource limited) master and is NOT inherent to the format itself! As a general point, the interviewee consistently (and deliberately IMO) mixes and confuses these two COMPLETELY UNRELATED considerations. He also (again deliberately IMO) erroneously lumps all MP3s together. A 128 (or lower) MP3 can usually be easily discerned from lossless but not so with a modern high bitrate MP3. This is a common trick employed by those trying to market something as better than MP3.
As far as audible quality is concerned, given an equal quality master Fig. 1 should be very different indeed! The relationship between reel to reel and LP is fine but reel to reel should be slightly lower than ideal. CD, DVD-A/SACD should both be the same, equal to ideal and therefore slightly higher than reel to reel. A download (given say a 320 MP3 OR 256vbr AAC) should be invisibly lower than ideal but still a little higher than reel to reel.
"The problem that MQA is addressing is how to transport an analogue signal to another time or place. It is the analogue signal from the mixing desk that the producer heard and that is the signal that you want to reproduce at your loudspeaker." - This is essentially just another example of the same thing, confusing/mixing up different practises, terms or considerations. This is a particularly bad example though, for these reasons:
1. Typically mixing desks do not output an analogue signal in the first place! Typically, mixing desks are either physical digital (digital in and out) mixing desks or virtual mixing desks (still obviously digital in/out). Analogue mixing desks are a rarity and becoming rarer.
2. Neither a producer nor any other human being can hear an analogue signal! Human beings (which include most producers
) can only hear acoustic signals. What a producer actually hears is therefore the output of their speakers in the studio environment and this is what they base all their production decisions on. This is why mastering and mastering facilities exist in the first place! The analogue signal derived from the digital files and the so called ADC "fingerprint" is complete nonsense in relation to what the "producer heard", unless of course you've got exactly the same speakers as the producer, installed and calibrated in exactly the same acoustic environment. This is going to make about a thousand times more difference than any inaudible "fingerprint"! This fact also makes Fig. 2 a nonsense. A CD played in a mastering studio compared with the original master file will sound identical! However, they do label fig. 2 as "Notional" sound quality and I'm sure many audiophiles do have the "notion" that the master file does somehow sound superior to the CD. 10 mins with an ABX test in a mastering studio would easily demonstrate that "notion" to be false!
"[1] In general, the MQA system can reach in excess of either 23-bit dynamic range capability or [2] 3–6 bits below the content noise in the audio band." - Two obvious points here:
1. There is no analogue signal to store or transport which has anywhere near 23 bit dynamic range!
2. Given that the "content noise" floor on most commercial recordings is around -40 to -50dBFS that's about 8bits of dynamic range. 3-6 bits below means that MQA would be 11-14bits. Even the widest dynamic range commercial recordings pretty much never exceed 60dB (10bits) in which case MQA would achieve 13-16bits. This is less than (or at best equal to) standard 16bit lossless or indeed a high bitrate MP3.
"Even in these great 2L recordings we don't often see hall/microphone/ADC noise below the 16-bit noise spectral level -- not surprising given the fundamental thermal limit for microphones." - Actually, we NEVER see noise below the 16bit level in a real recording. The very quietest mics achieve a dynamic range of about 90dB which is 15bit, on top of that we have to add acoustic noise from the hall and noise from the mic pre-amp. In practise we rarely see a recorded dynamic range of more than about 12-13bit. And, this is deliberately reduced down to 10bit or less during mixing or mastering anyway.
However, I agree with the idea of introducing some actual practical and audible limitations to the discussion, the audiophile world could really do with a great deal more of that IMO! My problem with the interviewee is that he applies these physical and audible limitations when it suits the purposes of selling his new codec and then does precisely the opposite to suit some other marketing point. For example, increasing noise outside the audible band doesn't matter but improving impulse response outside the audible band does? And, "temporal blur" is a meaningless term he appears to have just invented and the numbers in fig. 3 and 4 also appear invented or if not completely invented then certainly manipulated to worse case test scenario rather than what happens in practise.
In conclusion, the whole thing comes across as a typical audiophile bunch of marketing BS. Arguably more sophisticated marketing BS than many but still marketing BS nonetheless. It could be a successful new codec, because it offers some potential marketing opportunities to a number of different segments of the industry; content distributors, retailers and hardware manufacturers. Beyond these potential marketing benefits there appears to be no chance of any actual audible or convenience benefits over what already exists. This is certainly what all the misrepresentation, deliberate confusion of issues, invented terminology and double standards indicates but, I can't be absolutely 100% certain until more info is forthcoming of what's actually going on "under the hood" and the creation tools are available to run some independent real world tests to ascertain if there is anything more than just marketing BS to it.
G
