You donāt subscribe to the fallability of the human auditory system then?
I wonāt argue with the science, except I will in a sense. Here are my concerns with taking this at complete face value:
(1) We had the compact disc, which by todayās standardās isnāt particularly compact. High Definition TV (720p/1080i/1080p) is not high definition relative to 4K, and on and on. What science holds as completely unassailable today becomes tomorrowās half-truth, and so on. Point being, there is in my mind a possibility to a likelihood that future science will either consider your statement untrue or at least not the whole story regarding human sound perception. Thus, I do not think, as either an archiving format or something to āinvest in,ā that any format should dispense with any information - at least not to the extent possible within reason of storage capabilities.
That said, I think MQA is a pretty amazing concept as something that can be used for streaming and thus something I can hear from Tidal without heavy investment. Iām not happy about the requirement for the machine with the blue light, but if partial unfolding in Roon improves the sound of what we get from Tidal, and maybe even represents an improvement over some of my CD-FLACs, then thatās a win in my book. It just should not supplant other formats that donāt dispense with any data, double-blind testing be damned - and eventually Iād like to see it do everything but the DAC endpoint correction without requiring proprietary hardware.
(2) Those of us that grew up in the more punky-alternative half of the music world have a lot of skepticism about the record companies that hold the rights to much of the lifeblood of our hobby. And as I wrote above, Iāve repeatedly purchased the same titles looking for the best sounding versions as they come out. Some really were a revelation, others not so much. But still, many of us have a natural skepticism and feel that the record companies always want to hold something back to sell us something again later. MQA still has that feel, even if itās an improvement in some ways.
For those of us that want to retain our own, non-streaming library, a version that preserved all the data would be preferable. So thatās not to comment negatively about MQA being available through Tidal or other streaming services; itās just to state a concern that I hope MQA doesnāt supplant non-proprietary, lossless codecs.
(3) Finally, I do find the marketing of MQA to be a bit over the top. By telling us that weāll hear what was heard in the studio says it will never get better than this ā this is The It. For the reasons stated above, Iām skeptical about such an assertion. I think itās overblown. To say itās better than what you have, sounds really good, evokes an emotional response youāve not in your own listening room, all good. But please, itās too hard to believe that itās identical to the studio experience, and it feels like marketing speak, not sincerity.
6 Likes
Excellent reflection.
I totally agree with everything you said.
I tested audirvana with MQA files and I really liked the result.
Do you have any predictions as to when Roon will be able to unpack via software the first part of MQA files as well??
1 Like
I have heard from confidential sources ( Dr. Linn & Co) it might be released Summer 2020.
(ten characters to meet minimum post requirement)
The year 2525, Zager&Evans
The trouble with this question is that anyone with any information isnāt at liberty to make a prediction, meaning any predictions will be from those without information. The devs have indicated it is a priority.
2 Likes
Iām just catching up with this. Actually, if I understand, the claim is that it is lossless (in the traditional digital-file sense) up to the first fold, which would be 96kHz sampling, 48kHz audio. Itās only above thatādouble the audible range and nearly double the extent of musical information on recordings with exceedingly wide bandwidthāthat the compression becomes lossy.
Cheers.
James, nice post. To be clear though, MQAās inventors/developers have been very clear about the distinction between archival formats and transmission formats. Their chief publication (which however doesnāt mention MQA) indicates that they (Stuart and Craven) believe the output of a multibit sigma-delta ADC would be an appropriate archival format, although for existing recordings itās typically 24/XXX PCM, or DSD, or analogue tape. There are, howeverāmust beāintermediate stepsāmastersābetween the archive and the transmission. MQA steps into that territory via āwhite-gloveā remasters, and thereās often a decision as to which masters are the most authenticāthe most appropriate for MQA encoding and transmission. Distinctions arenāt as clear as they could be.
Jim
1 Like
Bob Stuart did mentioned about ā17 bit resolutionā of MQA compared to the āAirā and he believes no recording at the moment can achieve true 24-bit resolution. A very interesting revelation from the recent interview belowā¦
3 Likes
3 Likes
If you follow JAās measurements in Stereophile, itās clear that the best digital processors today manage 19-20-bit performance. Thatās not new. Thereās a big difference, though, between 17 and 19-20 bits. Thatās a factor of 4-8.
JA wrote about a power amp that was remarkably quiet, deep black backgrounds, all that talk. Normally you show amp noise in dB, but he described it as 21 bits. The quietest amp ever. Noise was 21 bits below 300 Watts. If you listen at more reasonable levels, you may use 30 W or 3 W on average, with 300 W headroom for peaks, so now amp noise is a lot worse than 21 bits below the music.
And room background noise level is way, way noisier than that.
Just because something can be measuredā¦
Iām not sure what youāre getting at here. I donāt recall that review/those amplifier measurements, and Iām not sure it has anything to do with the point here, which is about measurements of the actual noise floor/dynamic range of the best current DACs. Unless this, maybe, is your point: 20 bits is a hell of a lot of dynamic rangeāin which case I would agree. It does however provide some context for the discussion, about the relevance of 24 (or 32)-bit files, and that 17-bit figure for MQA (itās easy to forget that 17 bits is double what CD can do), so itās a point worth making.
Best,
Jim
Yes, thatās what I meant. 20 bits is a hell of a lot, so is 17, arguing about the difference doesnāt seem really important.
OK, right, agreed. 17 is a lot, 20 is a hell of a lot, 24 is more than is meaningful, unless you plan to do some work on it in the studio, in which case the extra is useful. Itās good for people to know that.
If the best recordings can achieve 20 bit resolution from studio to home reproduction, why canāt MQA design to match that resolution?
Because they reserve the lower 7 bits (or so) to encode the high frequency content (i.e., above 22kHz or so).
The claim is, though, that with proper processing, the reconstructed signal actually approximately 21 bits.
Where do you get 21 from? The claim is itās 24 bits see this.
From the above link you can see that it betters that resolution.