In Reply to: Likely an ignorant comment by me. posted by Rick W on June 6, 2017 at 11:24:08:
>> At this point I don't know how MQA works. Is it employed during the recording stage, mixing or mastering stage? Has zilch to do with any of that and is just employed in playback? <<
The short answer is that it applied at the mastering stage. The long answer is that MQA seems to have morphed over time, and as such is something of a moving target in attempting to understanding how it works. I believe that the original intent of MQA was to reduce the file size to allow easier streaming. Some time after MQA was announced, the goals were expanded to also improve the sound quality of existing files by performing digital filtering to "correct" for "errors" created by the A/D converter. Now it appears that they have expanded the goal even more by employing digital filtering to "correct" for the "errors" in the D/A converter also. Looking at those goals separately:
1) Regarding file size reduction, this success of MQA would seem to be dependent on the sample rate and bit depth of the original source file. The clearest case is when starting with a double-rate, 24-bit file (eg, 88/24 or 96/24) that the file size is reduced by ~30% because the bit depth is reduced to ~17 bits (one form of lossy compression). Quad-rate, 24-bit files have greater compression ratios because the audio data in the very top octave is compressed using lossy methods similar to MP3, which can achieve compression ratios of 10:1. In that case a quad-rate MQA-processed file will not be much larger than a double-rate MQA-processed file. While the resolution is still reduced to ~17 bits, this situation allows for the greatest file size reduction. It is unclear if there is any file size reduction for single-rate files. It appears that Redbook (44/16) files would actually increase in file size, while 44/24 and 48/24 files may again reduce the file size slightly due to the reduction in bit depth to ~17 bits.
2) Regarding improving the sound by way of using digital filters, this has been an area of interest for many decades. Both Wadia and Theta began producing D/A converters with custom digital filters in the late 1980s. Wadia specifically focused on improving the time-domain performance, just as MQA is currently claiming to do. Historically, the use of these types of digital filters that improve time-domain performance have been restricted to D/A converter manufacturers that possessed the resources to design and manufacture custom digital filters. When looked at from this standpoint, MQA is simply a third-party company offering to license the custom digital filter they have developed to any D/A manufacturer who wishes to use it.
One interesting note is Auralic. They originally had planned to support MQA but it turned out that the MQA requirements conflicted with features that Auralic were developing. Auralic has therefore implemented their own version of decoding MQA-encoded files. See link below for more information.
As always, strictly my own opinions and not necessarily those of my employer or stuffed animals.
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- RE: Likely an ignorant comment by me. - Charles Hansen 12:57:55 06/06/17 (17)
- RE: Likely an ignorant comment by me. - Isaak J. Garvey 13:20:21 06/06/17 (16)
- Sorry, wrong nomenclature - Charles Hansen 22:19:11 06/06/17 (13)
- RE: Sorry, wrong nomenclature - Isaak J. Garvey 10:25:49 06/07/17 (11)
- It appears that MQA has managed to mislead the Sound on Sound author - Charles Hansen 12:58:19 06/07/17 (10)
- RE: It appears that MQA has managed to mislead the Sound on Sound author - Isaak J. Garvey 13:58:04 06/07/17 (9)
- Fair Enough - Charles Hansen 16:03:33 06/07/17 (8)
- one "reviewer" - DAP 16:00:05 06/08/17 (5)
- RE: one "reviewer" - Isaak J. Garvey 17:48:47 06/08/17 (4)
- RE: one "reviewer" - SpotcheckBilly12345 07:31:39 06/09/17 (2)
- RE: one "reviewer" - Isaak J. Garvey 12:03:23 06/09/17 (0)
- That's a little extreme :-) - DAP 08:27:32 06/09/17 (0)
- socks off - DAP 19:59:12 06/08/17 (0)
- RE: Fair Enough - Isaak J. Garvey 17:04:41 06/07/17 (1)
- I do not agree with the author's conclusions - Charles Hansen 19:24:26 06/07/17 (0)
- RE: Sorry, wrong nomenclature - Isaak J. Garvey 10:13:51 06/07/17 (0)
- RE: Likely an ignorant comment by me. - ahendler 14:13:41 06/06/17 (1)
- RE: Likely an ignorant comment by me. - Isaak J. Garvey 14:34:26 06/06/17 (0)