The difference with then and now is that all the Record Labels are backing MQA and millions of songs are already available on Tidal, downloads and MQA-CD’s
But if it does not interest or harm you, why are you so worried about MQA? Many like it, some don’t. Just enjoy what you prefer.
MQA itself upsamples. MQA files are tagged so as to tell the MQA DAC the rate to which the file should be upsampled.
The MQA encoder doesn’t encode anything above 96k in any file, even if it is 384.
The lot of those things have been explained so many times already in these threads that it’s pointless to cover it again. Especially when you say you’ve read the patents and papers and still haven’t understood them
24/176 etc are the data rates of the original files. The first three papers do a good job of discussing dynamic range, both that in the original file and that in MQA. The original file does not and cannot have 24 bits of audible data, and MQA first analyzes and then preserves the actual dynamic range of the file. As far as sample rate, the measured data in the second paper shows that audio amplitude in all cases rolls off at 1/f rate. It means there is nothing there above the noise floor by around 60 or 70 kHz, and often nothing beyond around 40 kHz. MQA preserves out to 48 kHz. It is the PCM files extending to 192 kHz or 384 kHz that are “deceptive”, because data doesn’t extend that far out in music signals. MQA is simply encoding the areas that contain audible data.
Processing of PCM files including upsampling of CDs is nothing at all like the encoding and folding of MQA, which you’d know if you read the patents.
17 bits means 17 bits with noise shaping. That gives you up to ~20 bits of real data.
You just confirmed what I claimed.
MQA claims all those bits stripped out are useless, inaudible etc.
And you just confirmed they are discarded in the MQA encoding process.
I’m not really interested in MQA’s idea of what bits are useless and how they should sound after upsampling via their leaky filters.
Not coincidence but attenuation.
Many HDCD components, including those from Arcam, attenuated -6 dB all non HDCD sources to match RMS levels of HDCD sources with Peak Extend encoding/decoding. HDCD without Peak Extend – meaning most HDCD recordings – thereby got a +6 dB advantage compared to all non HDCD sources.
[Moderated] Sometimes quantitative difference is can be mistaken for qualitative difference.
AJ
In that case, you have much to learn about the psychoacoustics of audio engineering. They are not MQA’s ideas. They are the same ideas everyone else knows, but to date only MQA has written a codec incorporating them to good advantage.
MQA filters aren’t leaky. The internet has a lot of misinformation.
I’m afraid that wasn’t the case. It was no louder than anything else. It was just smoother, every time.
Sometimes it is helpful to recapitulate earlier discussions by sometimes really competent participants to see that we have been going around in circles for years now: On Comparisons... - Doug Schneider - Critic's Corner
Doubleplusungood?
Can you please show me a way back to the 90s?
So why hasn’t it been adopted by professional studios as a standard system for recording and archiving if it is so much better than used processes such as PCM / DXD and DSD which all studios still gravitate too
First, it isn’t an archiving system, it’s principally a release format, although the A/D correction can be used directly in new recordings. About the adoption by studios, I think this is a much larger question of the music industry and the pressures on studios as well as digital services and labels regarding format adoption as streaming continues to evolve. The question has to be very complex since it starts with what labels are willing to release (and thus what studios have to record), and why, and what industry players think will grow the market and produce eventual profitability.
The Stereophile analysis is of a file given to them by MQA, unfolded. It does not prove anything. To show that MQA can actually preserve information that high would require measuring the actual digital output of the unfolding process, which MQA does not permit as they forbid S/PDIF output on DACs which can decode MQA.
Regardless, for the MQA “origami” to preserve the noise in the file you showed, in an area below the noise floor of the original, you’d need a 32-bit file!
Every time one capability of MQA is pointed out, it contracts another. MQA’s “smaller files” can’t hold the high-res data. MQA’s digital filters cause the very thing MQA is supposed to fix: time smearing. And many “high-res” masters either don’t exist (Universal Studios fire, obvious up-sampling of non-high-res music to produce high-res versions, mastering engineers and artists coming out saying that they were never consulted).
As for my listening impressions being my opinion, it’s clearly obvious when a file has been bass-boosted. Anyone can observe this for themselves. Listen to Miles Davis’ Doo-Bop album on TIDAL for a glaring example. With a good headphone or speaker system the loss of detail and soundstage depth is very audible as well, and that is just one of numerous albums where it is noticeable.
Yes, they are. If they aren’t, show me the impulse response of an MQA filter, then a white noise spectrum out beyond nyquist. Technical analysis and measurements aren’t misinformation, and they have proven that MQA cannot be doing what it has been claimed to be able to do, but the opposite.
Impulse response examples are published in the papers cited above, as well as in many other sources. Technical analysis and measurements are a story requiring a deeper discussion than what can be done here. That includes anything you cite, which has to be verified carefully as to source and measurement.
How more biased can it become when the admin of the “MQA development and reviews” facebook page is entering this discussion? The same admin that has very close contact with Bob Stuart who even is a member on that facebook page. They even personally met. lol
I didn’t, what makes you think I did?
[Moderated]
16/44 mqa is the most ridiculous format, there’s nothing to fold. The ony thing that happens is that 16bits are compressed into 15bits, because they want to light up a green led and show a sample rate on the dac display, and 1 bit is needed for that. Not a single 16bit mqa has a studio dot, and the reason is obvious, they sound worse then redbooks.
I prefer the original unmangled 16/44 PCM where I still have the choice to select linear filters, which outperform the minimizing filters mqa uses.
We know how mqa is doing tests for the public:
They play a 16/44 cd against a 24/96 mqa. No wonder it should sound better then, since a 24/96mqa, coming in a 24/48 flac, is still about 2.5x the size of a 16/44 redbook. Probably they also used a linear fast filter for the redbook, keeping all the pre and post ringing, to make the mqa sound even better. They should have compared it with the original 24/96 masters and using the same kind of filters, but then mqa would sound worse.
My jaw dropped as well, when I read how they make tests invalid so Mqa comes out as a winner.
If you mean the removal of preringing, all minimizing filters can do that.
If you mean less harshness, all slow filters can do that.
Minimizing filters are non linear and actually create blur. It’s a property of such filters that makes lower sounds come first, creating a unnatural punchier bass.
Yep the result of using a minimizing filter.
I did, the PCM is way crisper and detailed. The mqa sounds damped.
I don’t know but I think you compared mqa with mp3 in your findings.
Look, Peter, he’s thinking the same.
So you admit mqa is supposed to sound worse then CD, unless you pay extra for it?
Yes on Tidal you’ll never find the original PCMs, because MQA ltd is affraid people should start comparing mqa with the original pcms and start prefering the PCMs and notice it’s all a scam.
True, there is no proof whatsoever there are more unfolds.
And then you should take a look at mqa-cd which is by definition 16bit. There everything is folded into 3 bits i.s.o. 7 bits making it even lossier then regular 24bit mqa. And the remaining 16/44 has to become 13/44 to make room for those 3 bits. MQA-CD sounds VERY bad if you don’t buy an mqa decoder. And if you do buy one, don’t expect to reach the sound quality of an orginal 16/44. Mqa-cd is showing it’s all about the money.
They are all lossy. The originals are always better. (Especially true for 16bit mqas, those are horrible)
A good summary Danny2 ! 
iFi (sorry to have to drag them into the discussion) created their GTO filter with help from (but clearly, from analysis, based on) MQA. If you look at the measurements of the output using the GTO filter in the Stereophile review of it, you’ll see that it leaks horribly. iFi Audio Pro iDSD D/A processor/headphone amplifier Measurements | Stereophile.com
The previously cited AES paper considers a shorter impulse response to be an improvement in the time domain. What it fails to point out is that an impulse response (effectively a single, full-output sample*) is an illegal, output-of-bandwidth signal. The response of the filter, splitting it into smaller frequencies, is the valid process of the filter attempting to remove out-of-bandwidth high-frequency noise. It is not something you’ll ever see in music.
*I have to note here for people following but not understanding, that PCM sampling requires more than two samples (even fractionally more) to be able to reproduce the original waveform. Thus, a single-sample at full power is an impossible signal. It’s equivalent to a signal containing infinitely all frequencies at the same time.
Hi Wim,
Are you enjoying Roon now? Maybe you can explain Danny where the cutoff frequency of 24/96 PCM is
Glad to hear you are OK
Cheers!
You have to have an MQA DAC for MQA to work as it is an end to end system, that’s why measurements need to be of the Analog resolution…