My simple analogy is what speaker correction s/w does, correcting known anomalies for a specific design on the fly, implemented for a known Dac
Anyone heard about Meridian's new project called MQA
- Thread starter thedudeabides
- Start date
The impression I get is that Bob Stuart, Robert Harley, John Atkinson, and others are intentionally all over the map trying to make MQA all things to all people and IMO coming up with some of the most outlandish and hyped claims in an industry already known for its hype. Almost as if to make it so overwhelmingly impossible or exhausting to debunk the technology as a whole.
Personally, I think it's perhaps the greatest charade ever pulled on this industry.
For those still living in the 1990's with limited bandwidth and 1GB hard drives, it's the answer to their prayers.
For those who've fallen in love with current hi-rez formats claiming it's night and day difference over the supposedly inferior Redbook format, it's higher than high rez.
For those who want to hear what was heard in the recording studio, Bob guarantees you will hear what was heard in the recording studio.
For those struggling with inventories of different formats, those days are over.
For those who consider psychoacoustics and neuroscience crucial to achieving true high fidelity, MQA is there.
For those who have noise floors no lower than Bob, your music should not suffer loss.
For those who think Bob's research in neuroscience is accurate and trustworthy, you may be in for a special Bob Stuart DSP treat.
For those who may not care about Bob or MQA, you'll never see the cute little green light and you'll get PCM in it's supposedly original context.
For those who play MQA without an MQA decoder, you'll hear better than CD quality - According to Harley.
For those who play MQA with an MQA decoder, you'll hear EXACTLY what the engineers heard in the studio. - according to Harley.
In his interview with Harley, Stuart said, "When we listen to music, it's analog...." "... converting the sound to digital is an unnatural act." Yet, here's Bob doing a wholesale sellout in the digital realm. I wonder if Bob thinks amplification is a natural act.
I could go about the unnatural gibberish, but it really is rather overwhelming. John Atkinson, who recently exclaimed at an audio show that the Vandersteen Model 7A's "made the hairs on the back of my neck stand up" and were "Musically perfect..... across the board." who apparently experienced musical perfection is on board MQA's superior sound, thereby implying better than perfect is obtainable.
About the only thing I see coming from this MQA propaganda wave is that there'll be an entirely new avenue of new component reviews on both the pro and consumer sides to potentially inject some life-blood into a dying corpse, and a whole lotta' money going into Meridan's and Stuart's pockets (and perhaps others) via product licensing and royalty fees while potentially compromising the level of fidelity of today's recordings.
Personally, I think it's perhaps the greatest charade ever pulled on this industry.
For those still living in the 1990's with limited bandwidth and 1GB hard drives, it's the answer to their prayers.
For those who've fallen in love with current hi-rez formats claiming it's night and day difference over the supposedly inferior Redbook format, it's higher than high rez.
For those who want to hear what was heard in the recording studio, Bob guarantees you will hear what was heard in the recording studio.
For those struggling with inventories of different formats, those days are over.
For those who consider psychoacoustics and neuroscience crucial to achieving true high fidelity, MQA is there.
For those who have noise floors no lower than Bob, your music should not suffer loss.
For those who think Bob's research in neuroscience is accurate and trustworthy, you may be in for a special Bob Stuart DSP treat.
For those who may not care about Bob or MQA, you'll never see the cute little green light and you'll get PCM in it's supposedly original context.
For those who play MQA without an MQA decoder, you'll hear better than CD quality - According to Harley.
For those who play MQA with an MQA decoder, you'll hear EXACTLY what the engineers heard in the studio. - according to Harley.
In his interview with Harley, Stuart said, "When we listen to music, it's analog...." "... converting the sound to digital is an unnatural act." Yet, here's Bob doing a wholesale sellout in the digital realm. I wonder if Bob thinks amplification is a natural act.
I could go about the unnatural gibberish, but it really is rather overwhelming. John Atkinson, who recently exclaimed at an audio show that the Vandersteen Model 7A's "made the hairs on the back of my neck stand up" and were "Musically perfect..... across the board." who apparently experienced musical perfection is on board MQA's superior sound, thereby implying better than perfect is obtainable.
About the only thing I see coming from this MQA propaganda wave is that there'll be an entirely new avenue of new component reviews on both the pro and consumer sides to potentially inject some life-blood into a dying corpse, and a whole lotta' money going into Meridan's and Stuart's pockets (and perhaps others) via product licensing and royalty fees while potentially compromising the level of fidelity of today's recordings.
IMO, there's so much wrong with the logic and words coming out about MQA that I can't help but think it utter nonsense.
If for no other reason than the problem every last one of our playback systems face has almost nothing whatsoever to do with the resolution format. The MQA format can no more solve this problem than a reasonable component or cable upgrade, just like all the other already established high-resolution formats have demonstrated.
All higher-rez promoters are all barking up the wrong technology tree. That is, if one is seeking anything more than relatively minor performance improvements.
What nobody seems get or at least admit to, is that listening to even the most well-engineered high-rez recording on a well-thought-out $1million SOTA-level playback system, the listener is not audibly hearing anything more than roughly the equivalent music info capacity of a well-engineered MP3 recording.
That's the intellectual truth of the matter. For example:
o Robert Harley, editor-in-chief at The Absolute Sound said in the Mar/Apr, 2009 issue, "I believe that the primary reason reproduced music doesn't sound like live music is some kind of catastrophic loss that occurs at the microphone diaphragm."
o Jonathan Valin, senior editor at TAS said about one year earlier, "We are lucky if even our very best playback systems can capture even 15% of the magic of the live performance."
o John Atkinson from the Sep, 2009 Stereophile issue, "I'm starting to feel that it is something that is never captured by recordings at all that ultimately defines the difference between live and recorded sound."
When it comes to performance, I don't put much stock into anything these guys say. Nevertheless, even a broken clock is right twice a day and their statements here are relatively accurate. Even Atkinson's.
The cause of this catastrophic problem (to use Robert Harley's words) is not inferior resolution. Rather it's an overwhelming amount of universal distortions that utterly destroy or cripple our sensitive component's precision and accuracy so that the noise floor is raised so high as to render much of the music embedded in the recording below the much raised noise floor and therefore remain inaudible.
I would attest that perhaps every last reasonably thought-out playback system is reading and processing roughly 99.999% of all the bits embedded in a given recording. No better or worse than downloading a 100MB flac file or burning that same flac file to your $40 CD burner that came with your computer. Better yet, how about backing up your payroll system to tape, byte for byte, experiencing a hard disk crash, and then restoring your payroll system (byte for byte) back to a new disk drive. IOW, the integrity and fidelity of reading and processing data has been extremely high for decades now and that is proven most every time a restore takes place.
All the detail captured by the recording mic's is being read and processed, only just like a severely blurred photo of a beautiful red Ferrari where the noise floor of the photo is so high, you can barely even tell it's a Ferrari. Is all the detail of the Ferrari stored in the photo? I say yes. But much of the signal is so distorted, it remains invisible. A performance-limiting governor if you will, that plagues every last playback system. This is why even well-thought-out playback systems all sound more alike than they do different and not one of them comes even remotely close to the absolute sound aka the live performance.
Bear in mind that the most severe distortions are not even audible per se. These are inaudible distortions that quietly go about their business of catapulting a given playback system's noise floor to very high levels. And it's not just low level detail as many speculate. The harm is induced pretty much equally between low and high level detail.
And just like with a 3.5MP camera or a 13MP camera, the problem with the blurred photo is regardless of the camera's resolution capability or the number of pixels stored per photo graph. An unblurred photo of a beautiful red Ferrari taken with a 1.5MP camera is far more beautiful than a blurred photo of the same Ferrari taken with a 16MP camera.
And just like the camera analogy, it's not the number of music info bits that remain inaudible, it's the percentage of music info bits that remain inaudible due to the much raised noise floor.
This is not hypothesis, this is fact and is demonstrable. But when "scientists" and the science-minded are already halfway up the wrong technology tree, it's pert near impossible to call them down until the entire industry has been forced to buy into it at great expense.
Only for many to wake up 5 or 10 years later realizing they've been doing cartwheels over small incremental improvements when the industry had them convinced the advancement would be life changing.
This is why a well-engineered MP3 recording played back on a SOTA-level system theoretically isolated from the harm induced by such universal distortions should sound at least if not far more naturally musical than a well-engineered 24-bit / 192 kHz high-rez (better yet an MQA) recording on the same SOTA-level system fully susceptible and affected by these universal distortions.
And like the higher rez camera susceptible to distortions, It's simply impossible for higher rez formats to cure the ailments plaguing our systems. Though it can bring about some nice minor improvements.
If for no other reason than the problem every last one of our playback systems face has almost nothing whatsoever to do with the resolution format. The MQA format can no more solve this problem than a reasonable component or cable upgrade, just like all the other already established high-resolution formats have demonstrated.
All higher-rez promoters are all barking up the wrong technology tree. That is, if one is seeking anything more than relatively minor performance improvements.
What nobody seems get or at least admit to, is that listening to even the most well-engineered high-rez recording on a well-thought-out $1million SOTA-level playback system, the listener is not audibly hearing anything more than roughly the equivalent music info capacity of a well-engineered MP3 recording.
That's the intellectual truth of the matter. For example:
o Robert Harley, editor-in-chief at The Absolute Sound said in the Mar/Apr, 2009 issue, "I believe that the primary reason reproduced music doesn't sound like live music is some kind of catastrophic loss that occurs at the microphone diaphragm."
o Jonathan Valin, senior editor at TAS said about one year earlier, "We are lucky if even our very best playback systems can capture even 15% of the magic of the live performance."
o John Atkinson from the Sep, 2009 Stereophile issue, "I'm starting to feel that it is something that is never captured by recordings at all that ultimately defines the difference between live and recorded sound."
When it comes to performance, I don't put much stock into anything these guys say. Nevertheless, even a broken clock is right twice a day and their statements here are relatively accurate. Even Atkinson's.
The cause of this catastrophic problem (to use Robert Harley's words) is not inferior resolution. Rather it's an overwhelming amount of universal distortions that utterly destroy or cripple our sensitive component's precision and accuracy so that the noise floor is raised so high as to render much of the music embedded in the recording below the much raised noise floor and therefore remain inaudible.
I would attest that perhaps every last reasonably thought-out playback system is reading and processing roughly 99.999% of all the bits embedded in a given recording. No better or worse than downloading a 100MB flac file or burning that same flac file to your $40 CD burner that came with your computer. Better yet, how about backing up your payroll system to tape, byte for byte, experiencing a hard disk crash, and then restoring your payroll system (byte for byte) back to a new disk drive. IOW, the integrity and fidelity of reading and processing data has been extremely high for decades now and that is proven most every time a restore takes place.
All the detail captured by the recording mic's is being read and processed, only just like a severely blurred photo of a beautiful red Ferrari where the noise floor of the photo is so high, you can barely even tell it's a Ferrari. Is all the detail of the Ferrari stored in the photo? I say yes. But much of the signal is so distorted, it remains invisible. A performance-limiting governor if you will, that plagues every last playback system. This is why even well-thought-out playback systems all sound more alike than they do different and not one of them comes even remotely close to the absolute sound aka the live performance.
Bear in mind that the most severe distortions are not even audible per se. These are inaudible distortions that quietly go about their business of catapulting a given playback system's noise floor to very high levels. And it's not just low level detail as many speculate. The harm is induced pretty much equally between low and high level detail.
And just like with a 3.5MP camera or a 13MP camera, the problem with the blurred photo is regardless of the camera's resolution capability or the number of pixels stored per photo graph. An unblurred photo of a beautiful red Ferrari taken with a 1.5MP camera is far more beautiful than a blurred photo of the same Ferrari taken with a 16MP camera.
And just like the camera analogy, it's not the number of music info bits that remain inaudible, it's the percentage of music info bits that remain inaudible due to the much raised noise floor.
This is not hypothesis, this is fact and is demonstrable. But when "scientists" and the science-minded are already halfway up the wrong technology tree, it's pert near impossible to call them down until the entire industry has been forced to buy into it at great expense.
Only for many to wake up 5 or 10 years later realizing they've been doing cartwheels over small incremental improvements when the industry had them convinced the advancement would be life changing.
This is why a well-engineered MP3 recording played back on a SOTA-level system theoretically isolated from the harm induced by such universal distortions should sound at least if not far more naturally musical than a well-engineered 24-bit / 192 kHz high-rez (better yet an MQA) recording on the same SOTA-level system fully susceptible and affected by these universal distortions.
And like the higher rez camera susceptible to distortions, It's simply impossible for higher rez formats to cure the ailments plaguing our systems. Though it can bring about some nice minor improvements.
Last edited:
He would rightly argue that it is unnatural with regards to separate analogue amplifiers and passive speakers, due to group delay and phase distortion these bring; one area active speakers like the Grimm Audio are superior and technically more natural.
However the audio world has decided to ignore this for decades even though articles highlighting this and some investigating audibility have happened since the 80s (KEF were one of the 1st if I remember to identify and come up with a solution - which did not take off as it did not fit in with the accepted general perception, same could possibly be said about the work done by Gerzon and also with Craven).
Context is important, that example I provided touches a little on what I feel Bob Stuart sees as the issue with digital and filters/ADC/chain implementations, especially in the past and unfortunately how even today it seems many studios make mistakes.
Cheers
Orb
I think it is important, especially in audio, to separate the claims from the underlying science. For some reason after many years in this hobby, I do that instinctively. The unsubstantiated claims go in one ear and out the other for me. But, my impression is that you are hung up on the verbiage of claims and marketing, then you are going through convoluted logic about those claims. However, you are not really understanding the underlying science and its implications. Amir has provided a good explanation of the possibilities of the science here, also avoiding getting bogged down in what are likely to be fictitious or overblown claims.
In all I have read in the actual words of Stuart, Harley and Atkinson, I see plausible explanations from a scientific standpoint of how this MIGHT be able to provide an improvement in sound quality over digital audio as we know it. That potential improvement is not a result of an increased sampling rate beyond hi res audio as we know it, which is a tangent you have pursued.
And, we are all big boys here. If anyone expects this or any other technology to once and for all totally eliminate the gap between live and recorded sound, they are living in fantasy land. But, honestly, I have seen no such claim for MQA in my readings. All that is promised and all that I expect is an improvement, a step in the right direction, that may take us closer to the mike feed and beyond the limitations imposed by the original analog to digital conversion in the recording process.
Per some of your quotes, I agree that the mike itself does not hear live music as we do. Unless and until that can be overcome, if ever, we have to settle for incremental sonic improvements later in the recording chain, such as what MQA purports to deliver. That is all I see Stuart, Harley and Atkinson alluding to.
MQA can not possibly improve sound quality over the original full high resolution input file that it encodes. Anything along these lines is just marketing smoke which makes me suspicious of everything the promoters and their paid shills say.
It is possible that MQA provides a lossy compression method that allows higher quality audio to be delivered at CD data rates. However, all this amounts to is saving a few bits and bits (storage or network transmission) are so cheap that this is a problem that is no longer relevant to audiophiles.
If there is any "benefit" at all to MQA it is in the streaming application where the audio resolution can be automatically downgraded in the event of network congestion. This seems to be of questionable value. For serious listening one will always be worrying if one is going to get the full quality. For casual listening then there seems little need for more than CD quality in the first place, something that can easily be delivered without MQA, for example using FLAC encoding.
Well, objectively (i.e. measurements) it may be able to do that as I explained. Let's create a fictitious example of an A/D converter that drops by 0.5 dB by 20 Khz. If I measure and profile that, then at playback time I can insert an inverse of that restore that droop to a flat line. This would be measureable.
Where it gets tricky is that Bob attempts to correct what he says are the timing distortions created in capture despite the high sampling rates. The audibility of such distortions is not accepted currently. So while objectively we may be able to show this, subjectively, i.e. when we listen, these differences may be well below audibility thresholds.
BTW, it is a shame that if we are going to profile something at recoding stage, is not the sound of that room! What we want to reproduce is what is heard there. Measuring and duplicating that response would be far more valuable than characterizing a tiny component, i.e. A/D signal processing, and undoing that.
stehno
You seem to be getting all excited and hung up on descriptions/write-ups of what MQA will or won't go. I suggest you take the opportunity to hear it before you write another 1000 word essay giving us your opinion as to why it can't work.
The demo I heard may not have knocked me off my chair but I did hear SQ on a couple of tracks that were significantly different/better than any version I have heard in the past.
You seem to be getting all excited and hung up on descriptions/write-ups of what MQA will or won't go. I suggest you take the opportunity to hear it before you write another 1000 word essay giving us your opinion as to why it can't work.
The demo I heard may not have knocked me off my chair but I did hear SQ on a couple of tracks that were significantly different/better than any version I have heard in the past.
It's obvious from reading some of the posts here on this thread that we all either aren't reading the same articles that have been published or we are not comprehending the information in the same way. ESL tried to convince us that Meridian was making no claims for sonic improvement with MQA, it was just merely a more efficient transmission method for streaming. MQA is absolutely not being touted as some small incremental improvement in the recording chain. Meridian is making claims that their MQA files will sound better than existing hi-rez files. Both Stereophile and TAS had glowing things to say about how the demo sounded while others who have heard it are calling it a "game changer."
The problem is that the recording may have been optimized by the mastering engineer to sound the way he and the producer liked based on the particular distortions of the ADC that they used. Second guessing this process amounts to "automatic remastering". Actually, doing this is impossible. The filtering and equalization (or microphone placement for a purist recording) were made based on listening to the final product through a particular playback chain. The filters in the DAC, not to mention the mastering speakers and room, made the recording sound good, but the information about this playback chain is not encoded on the recording.
There is another problem with second guessing the ADC filters. Taking your specific example, that 0.5 dB boost can certainly be applied, but doing so does not just restore the amplitude response to be flat up to 20 kHz. It will have affects on the response at other frequencies as well as the time domain response. These effects may be audible and their evaluation may depend on the specific type of recording and musical genre. Perhaps there will be some future artificial intelligence technology that will automate the mastering engineer function, but I doubt that MQA is such a technololgy.
It is certainly possible that MQA can do a better job of fitting audio into a fixed bitrate than merely adapting a lower PCM format with associated lossless coding. However, I put this in the category of "saving bits" not "better sound" and this is about what is good enough, not what is best. Incidentally, unless MQA comes with a complete specification and open source code for the encoder and decoder, together with free and unrestricted licensing, I have no intention of even investigating the technical details. We have too many formats as it is to justify another one that is propritary. (I never even considered SACD because of its use of DRM and new patentent technology. I believe this contributed heavily to SACD's failure in the marketplace.)
Also I thought the MQA also deals with the real world label-studio-distribution that could do upsampling and/or downsampling (we know it happens), where again it is possible for it to be messed up (definitely happens).
Cheers
Orb
I see your point but the process they going down the path of, takes care of this in that the producers of music will have to a) help profile their ADCs and b) perform quality testing to tell if this is an improvement or not, as to then recreate all of their stereo digital masters for distribution.
Also, the improvements here are not in frequency domain so the timbral aspects won't be changed. Those are the things that are EQ'ed and fixed. What MQA purports to have fixed would not have been things that mastering takes care of ordinarily.
Looks like I created the problem with that analogy
. Per above, there is no frequency response fixing but rather, other processing to better preserve timing resolution. Well, it is better sound if one reads the effect of processing in conversion to long word PCM (i.e. 24 bit) to have impacted time domain/transient response. This is highly disputed as I mentioned but is the nature of what they are trying to do. As they demonstrate in the paper, this can be measured. It remains to be seen if it can be heard.
Great point. Having one's purchased music in this format and having the company discontinue its promotion, could prove highly costly to consumers who would at some have to buy their music all over again.
I, for one, have absolutely zero interest in hearing a new proprietary format. I don't even want to hear files that have been put through such a process, since this is open to all kinds of selection or tweaking. I consider even listening to these files to be a bad idea. The industry needs a new proprietary format like a fish needs a bicycle.
If I am given source code for the CODEC that allows me to experiment with my own choice of source files and which comes with a free and open license for the code and all patented technology, then, and only then, would I be interested in evaluating this format. [I am not singling out Meridian. I went through the same due diligence with FLAC.]
It is too bad that someone from Meridian is not responding to these posts.
As "I" understand the issue, it is not Meridian second guessing what the engineer or the artist wanted. Someone, and I do not know who will be responsible for performing this at each label, digitally signs off that the output is what is supposed to be heard. Then when the file is "unpacked" (my term not theirs) either via hardware or software that can handle MQA, a singal is passed to indicate that the end result is Master Quality Authenticated.
I am not trying to convince you to listen to anything. Too many people here are just commenting when they apparently don't even understand what MQA is or is not.
Do we need a new format? The only format I find most of the music I want is on CD so the answer is no for me, but I am keeping my options open.
Last edited:
Anyone ever wonder why digital audio has so many formats and most consumers don't even know they exist? The average person knows what a CD is and they know they can download music at iTunes (and I'm not even convinced the average person could tell you what format they are downloading at iTunes and they really wouldn't care anyway because the names and numbers don't mean anything to them). I remember many years ago going to a Best Buy store after the Rolling Stones catalog had been released in a dual format of RBCD and SACD. This was back in the day when you could go to a 'music' store and they actually had a large selection of CDs to choose from. I asked the sales clerk if he had the new SACD versions of the Stones. He looked at me like I had 3 eyeballs and asked me what an SACD was. If you asked people where you work tomorrow what they think of SACDs, you would get the same reaction. Never mind asking them about FLAC, ALAC, WAV, 24/48, 24/96, 24/176.4, 24/192, or any of the other 1000 or so PCM based formats.
So now we have one more new digital kid on the block named MQA that is going to head to the local digital saloon and try to elbow his way up to the bar and have a drink and try and get noticed. Once the record labels rubbed the digital lamp and let the digital genie loose, it killed the music business and streaming has buried it for the artists. There are only a handful of artists making money directly off of sales of their music and the rest have to stay on the road and tour in order to make a living. Even if MQA lived up to all its promises, I doubt it would make a splash outside of digital audio circles comprised of people who just love digital audio and chase every new digital format in hopes of getting more perfect digital sound.
So now we have one more new digital kid on the block named MQA that is going to head to the local digital saloon and try to elbow his way up to the bar and have a drink and try and get noticed. Once the record labels rubbed the digital lamp and let the digital genie loose, it killed the music business and streaming has buried it for the artists. There are only a handful of artists making money directly off of sales of their music and the rest have to stay on the road and tour in order to make a living. Even if MQA lived up to all its promises, I doubt it would make a splash outside of digital audio circles comprised of people who just love digital audio and chase every new digital format in hopes of getting more perfect digital sound.
Now here is my audiophile hobby hat on; what I do not like is that I am then forced to use a specific reconstruction filter rather than the mix of minimum-linear phase, and this may be a sticking point for some just like going with an all in digital integrated-DAC amp or active speakers that has limited appeal.
Although maybe most hobbyists will not consider filter selection/specifics but I think they will appreciate switching between DACs (and interpolation/reconstruction filter along with associated internal hardware is an influence) or interested in products reviewed where it goes into detail how a filter option on the product helps.
Now one thing that stands out carrying this on, what happens with regards to the interpolation internal oversampling filter......
The client end solution (hardware and internal software/possibly external software) needs to be pretty slick.
Cheers
Orb
Similar threads
