I support Bruno for a simple reason: he designed a superb ADC, as evidenced by the sound quality of many Channel Classics recordings. Bruno's goal was sound quality and he achieved it. Stuart's goal appears to be useless bit bumming and (possibly) latching onto a royalty stream for unnecessary products, based on an obsession he has had since his side lost the hi-res format wars.
Anyone heard about Meridian's new project called MQA
- Thread starter thedudeabides
- Start date
I support Bruno for a simple reason: he designed a superb ADC, as evidenced by the sound quality of many Channel Classics recordings. Bruno's goal was sound quality and he achieved it. Stuart's goal appears to be useless bit bumming and (possibly) latching onto a royalty stream for unnecessary products, based on an obsession he has had since his side lost the hi-res format wars.
Needs to be viewed in the whole context of the discussion between myself and Stehno, this goes back to the posts I referenced a little while ago.
Again they said (which ties in with the technical aspects Amir mentions and I am as well):
It cannot be clearer than that, sorry Tony.Grimm Audio for LS1 said:
This does not necessarily conclude just how critical filters/downsampling/upsampling/transcoding/dithering/etc are in the context of the complete audio industry but that there is a problem (not the problem); this is not just consumer but all aspects of the chain-processes-workflow associated with a recording and its distribution.
That said Hi-fi News has measured seriously dubiously released digital albums that fundamentally ruin their hirez quality from many different labels and distributors, all down to the way it was handled by the various people-chain involved; surely consistency is what we are after, including for CD releases that must come from a better master file.
If it is an obsession as you say, then you should also complain about Grimm Audio (because that quote aligns with Bob Stuart) and the associated engineers, along with those engineers and other manufacturers that create digital filters (time and frequency,impulse response and echo-ripples/phase) or discuss the issue and challenges with dither/upsampling-downsampling/etc.
It is a complex issue to do digital correct from the start to the end, and consistently (this is where it falls down as some of those poor released hirez are from some surprising labels - not all are bad but the point here is consistency)
Of course this is only one aspect of MQA.
Please note I am not necessarily advocating MQA is essential, just that it seems the music industry as a whole cannot be bothered to resolve these complex and intertwined challenges.
Look how long it has taken the music industry to get its act together implementing dither correctly, and then consider that for all the music industry from the elite niche (the very small number that take care from recording-mixing-mastering) to more mainstream hirez and standard releases or streaming.
Cheers
Orb
Last edited:
I addressed the issue of phase response in an earlier post. Apparently I did not make myself clear. The phase response on a digital recording is just one aspect of the choice a mastering engineer must make when making a recording. His job is to produce a result that sounds good on the majority of systems. In particular, his job is to produce a recording that sounds good when played back in his mastering studio, as this will be the sound that will be approved by the producer and/or musicians. If there is a problem with the ADC then the engineer can adjust this according to his taste, but he can only evaluate the result based on what he hears, and that will be through his DAC and its filter(s). There is no way to automate this process based on the contents of the master file. As an engineering process there is insufficient information (e.g. filters used in the DAC) to know what the intended sound of the recording is. And even if this information were present, the final judgment involves tradeoffs and this is an artistic process, not something that can be automated by a scientist or engineer, AES fellow or not.
The above concerns the "M" and "Q" parts of the MQA marketing smoke. Now I get to the "A" part. Cryptographic technology that turns on a green light will not make up for lack of integrity in the music and audio industries. Cryptographic certification has value only to people who trust the certifying authority(s) to be honest and competent. MQA exists in the context of a corrupt recording industry and a corrupt high end audio industry. Lack of integrity can not be solved by technology.
Ah ok,
but then most mastering engineers do not understand the filter implementation associated with ADC/DACs and coefficients (to them do they really care the difference between minimum/linear/steep brickwall), what they may do is select a product they like the sound of (appreciate this can be to get as close to as matching the sound to the speaker monitors)
If it his/her job to produce a sound that is good on the majority of systems, who is also responsible for the selection and when to use dither/upsampling/downsampling/etc?
As I was saying before, many cases dither is not necessarily implemented correctly, and there are plenty examples where hirez digital files are actually done in a flawed way beyond that (as per measurements shown by Hi-fi News).
So I agree in principle with what your saying, but in reality there is a disconnect between what the staff do (some or many no idea myself) and the actual digital technology.
Regarding knowing what filter was used, MQA can have profiles built for the associated with ADC, where this may fall down is if ADC can implemented multiple choices and whether notes explain the setup used for a recording (I get the feeling the answer is no and I guess why your a bit sceptical with your background in audio), anyway ADCs and DACs ideally should be perfect in both time and frequency domain as part of the judgement to what the intended sound of the recording should be but that is more from a concept and discussion than current reality.
TBH I also feel that if people cannot see the issues-challenges involved here with PCM through the whole chain for hirez and standard, then why even bother with DSD as it is not needed and pretty redundant.
I do agree that there is a difference between PCM and DSD, but logically part of that comes back to some of what this thread touches upon.
Cheers
Orb
Last edited:
What I didn't realise about MQA, until I read the Harley review, was that MQA was more than just a clever way of delivering high quality audio at lower bandwidth - it also addressed some aspects of digital filters. I guess that I should have realised this from Stuart's AES paper but I didn't.
I'm of the opinion that Bob Stuart is correct in his general guiding principle - that the learnings about psychoacoustics are the necessary underpinnings needed for advancing the goal of audio reproduction i.e providing as realistic an illusion as is possible. To my thinking, the most important statement in the whole article is "Meridian maintains that the standard metric for digital audio - sample rate and the number of bits in each sample - are less important than two new metrics 1) absolute stability of the noise floor & 2) the amount of "temporal blur" in the digital system.
I don't know yet how effective MQA will prove to be to our auditory perception - I wait to hear a consensus about the sound when using this technology. But I'm fairly well convinced by my own various experiments that point 1 - "absolute stability of the noise floor" - is crucially important. What strikes me is that this noise floor stability is one area which seldom gets measured using suitable test signals & measuring techniques.
Some other things that struck - Figure 2 "impulse response of typical 192/24 Vs MQA reminds me of an interesting thread on DIYAudio which investigated various iterations of digital filters for the Soekris DAC - it's a long thread but gets interesting around the time that the patent paper from Michael Pflaumer (of Pacific Microsonics & Berkeley Audio Design fame) is discovered http://www.diyaudio.com/forums/digi...ilter-brewing-soekris-r2r-60.html#post4284615
After that some interesting Impulse Response graphs are investigated for each of the filters designed & it guides the design for quite a while. However, by the end of the thread, it seems other criteria become more important in the audible differences between filters. An interesting thread!!
So, IMO, Stuart may be on the right track & it remains to be seen if MQA will prove to be a sufficiently audible improvement to become mainstream?
I'm of the opinion that Bob Stuart is correct in his general guiding principle - that the learnings about psychoacoustics are the necessary underpinnings needed for advancing the goal of audio reproduction i.e providing as realistic an illusion as is possible. To my thinking, the most important statement in the whole article is "Meridian maintains that the standard metric for digital audio - sample rate and the number of bits in each sample - are less important than two new metrics 1) absolute stability of the noise floor & 2) the amount of "temporal blur" in the digital system.
I don't know yet how effective MQA will prove to be to our auditory perception - I wait to hear a consensus about the sound when using this technology. But I'm fairly well convinced by my own various experiments that point 1 - "absolute stability of the noise floor" - is crucially important. What strikes me is that this noise floor stability is one area which seldom gets measured using suitable test signals & measuring techniques.
Some other things that struck - Figure 2 "impulse response of typical 192/24 Vs MQA reminds me of an interesting thread on DIYAudio which investigated various iterations of digital filters for the Soekris DAC - it's a long thread but gets interesting around the time that the patent paper from Michael Pflaumer (of Pacific Microsonics & Berkeley Audio Design fame) is discovered http://www.diyaudio.com/forums/digi...ilter-brewing-soekris-r2r-60.html#post4284615
After that some interesting Impulse Response graphs are investigated for each of the filters designed & it guides the design for quite a while. However, by the end of the thread, it seems other criteria become more important in the audible differences between filters. An interesting thread!!
So, IMO, Stuart may be on the right track & it remains to be seen if MQA will prove to be a sufficiently audible improvement to become mainstream?
Last edited:
Good info, Tony.
I'm curious if these so-called digital screw ups are discovered or realized by analyzing the digital stream or audibly?
On another note, I find it interesting that Stuart claims, "MQA is focused on the fact that converting sound to digital is an unnatural act in the first place."
Seems to me that if there was any truth to this statement and if MQA was a legitimate solution regarding performance improvements, Stuart's solution (MQA) or any solution would have to occur outside of the digital realm.
Logically, I just don't see how 2 unnatural acts can equal a natural solution.
Last edited:
A lot are questioning it here too but you are arguing with them nevertheless. As for AVS, be careful to not get confused by a few hall monitors who roam that forum to fight against anything above mass market audio.
Didn't we discuss this earlier?
The unnatural aspect comes down to the filter implementations not being perfect and causing timing-phase-ripple/ringing, our discussions expanded to show this is one area digital active speakers also correct and importantly I showed how very similar language is used by Grimm Audio (which I quoted) involving engineers respected by others here (even you commented about Bruno seeming to be ok and a smart engineer - and he was heavily involved in that speaker).
How important is this in the scheme of things, well its great to resolve this if one also has a solution that also alleviates the human errors involved in upsampling/downsampling/etc that do happen in real life in the music industry (which is also part of MQA).
My biggest gripe beyond any of this is that it would mean part of what we like about DACs would be sidestepped as there would in theory be one correct filter implementation at the client end rather than say being able to select slow-minimum phase/linear/steep, and to me this is one aspect that helps us with our long term preferences.
Another technical aspect that has me wondering, is the interpolation filter and how the client MQA integrates its functionality 'within' that of the DAC whole architecture; at what point in the DAC (will be mostly network streamers rather than traditional DAC IMO it is implemented in) does the MQA encoding function.
Cheers
Orb
It seems you are a very good sentence parser, but you seem to consistently miss the forest while you analyze individual trees out of context. And, your obvious anti-digital bias plays a big role throughout.
It could easily be said that analog recording, mastering, pressing and playing vinyl is just as unnatural an act. Mechanical grooves on LP or variable magnetic fields on analog tape are also man made and done by technological processes, hence "unnatural". In fact, all recorded music is unnatural. Digital is just the newest and most sophisticated way of accomplishing the difficult task of recording, for better or for worse, like it or hate it.
What Stuart was alluding to is the conceptual difficulty of recording and playback, in this case by digital means, not that it is bad or good. Digital recording is newer, still evolving at a higher rate, much more complex and therefore less well understood by ordinary folk, including audiophiles.
Yes, since all recordings, analog and even digital ones, are eventually converted to analog pressure waves in the room so we can hear them, the audible result of his process will ultimately have to be judged by what we can hear. If that is all you are saying, fine. But, there is nothing whatsoever in the MQA process that needs to or actually does take place in the analog domain.
>>And, your obvious anti-digital bias plays a big role throughout...It could easily be said that analog recording, mastering, pressing and playing vinyl is just as unnatural an act.<<
Unless Mr.John Stehno can "manage" to prove me wrong ... he has NO concept of hi-end vinyl capabilities, past shows, reading reviews, or fabrications ... but because it is Mr.Stehno your addressing, I understand the confusion.
Unless Mr.John Stehno can "manage" to prove me wrong ... he has NO concept of hi-end vinyl capabilities, past shows, reading reviews, or fabrications ... but because it is Mr.Stehno your addressing, I understand the confusion.
Here is the problem with your bolded statement: Stuart called digital musical recording an unnatural act because music is analog in nature-it's not digital. Analog may and does have its warts, but at the end of the day, it's analog. It used to be against the law in many countries (and still is in many countries) to commit an unnatural act and if we still had archaic laws on our books in the U.S., everyone who converted analog to digital would be serving time in jail where they would be further subjected to other acts that could be considered unnatural by those who stand in judgment of such things. It's a complicated time in a complicated world, but analog is analog and music is analog in nature.
If you want to be pedantic, "analog" recording on magnetic tape is both sampled and quantized, and poorly at that.
It is sampled because the magnetic flux on the tape is set by the sum of the signal and bias currents. The bias current varies from min to max over the period of one cycle of the bias frequency. Worse, the timing of the setting of the flux level depends on where on the bias waveform curve the sum of the currents occurs. (In other words, jitter.)
It is quantized because the flux level is determined by the number of magnetic domains flipped by the magnetizing field. There are a finite number of domains per unit length of tape, so a finite number of possible flux levels. The number of domains varies somewhat along the tape length, it's not a constant, therefore the total flux varies, which causes noise when read. (Like a digital system where the number of bits for each sample varies randomly.)
You want me to respect Bruno whom I do not know? I can do that. It's usually the ones I do know or those whom I may know something about that I struggle with the respect thing.
But along that same respect vein, just because somebody has a certain level of correct intellect about a given technology, by no means does that imply that ultimately they know what they are doing or talking about. Especially if they start to stray from the core of their area of expertise. Let's say there's a well-respected amplifier designer who routinely produces a superior amplifier. What happens all too often with this "respect" thing is that since he's somewhat of an expert at designing amplifiers, instantly he's well-respected for being an expert when he tries his luck at designing a line conditioner, even if his line conditioner is a POS. IOW, we should always be on guard and apprehensive when even well-respected people speak.
You speak of this unnatural aspect of filter implementations not being perfect, causing timing-phase-ripple/ringing, etc. I may or may not agree with that but I get that. My question to you is, are these unnatural aspects discovered by listening to output or by viewing readouts?
Also, when Stuart talks about noise floor, which noise floor is he speaking of? I have to assume every processing aspect of a signal regardless of type must have a potential noise floor. The only noise floor I've ever cared about the one generated by my playback system. But again, when Stuart or others talk about MQA and noise floors, which noise floor are they speaking of?
Last edited:
My anti-digital bias? Not sure where you're getting this from. I was quoting Stuart verbatim and then tried to think logically from his MQA solution occurring in the digital realm perspective.
Was Stuart stating that the conversion itself was an unnatural act? Or was he implying any music residing in the digital realm was an unnatural act.
If the former, then is converting back to analog from digital equally unnatural? And does the second sin (DAC) undo the harm induced by the first sin (ADC)? Or have two unnatural acts just been committed?
If the latter, (digital itself is unnatural). why has Stuart spent so much of his time and resources within the digital realm? And why is most of Meridian's gear digital? Moreover, aren't most of the SOTA-level cameras these day digital? I'm not aware of photographers complaining their cameras are committing unnatural acts when converting analog to digital?
In either case, why would MQA take place entirely within the digital realm?
For that matter, is amplified music an unnatural act? How about speaker drivers? Or driving an automobile or using a credit card?
Based on what I perceive to be a small mountain of seemingly nonsensical, illogical, contradictory, and yes impossible statements and claims, my suspicion is Stuart made that potentially meaningless statement (converting music to digital being an unnatural act) with the hope that some analog lovers would quickly read over it without much thought and perhaps jump on the MQA bandwagon.
If it were just one or 2 odd statements, no problem, but when I start to rope them all together and then toss in what appear to be extremely controlled demonstrations, I see so many red flags I can no longer count them. But that's me.
I remember that theory being debunked in a thread a year or so ago. Do a search.
Maybe it's the other way round; digital is bad analog tape.
Last edited:
BTW, that's vinyl limitations, not capabilities. Which in its current state seems just a tad below digital but it's pretty close and neither are very impressive. That is, unless one didn't know what they were talking about. Perhaps a bit like yourself?
But you, being a smart feller, realize that one needs to also look at an objects potential as well as its actual.
Mod: post edited. Do not get personal. Stay on topic.
Similar threads
