But potentially pretty lucrative. They have been eyeing Dolby's success for years.
I'm still lost as to how the "processing" makes things sound better than the original content. They claim it is removing some "dispersion" in the ADC/DAC. I would really like to understand this one better. Its related to the filtering I believe but all the stuff is really murky. I have not been able to measure anything like that so far.
I wonder if any of the HDCD (PM) related patents are still around and valid? I fail to see how MQA could not infringe on them. At best, MQA seems like HDCD-type features encoded into the bitstream paired with a lossy codec.
Just like HDCD, it's a solution to a problem that doesn't exist. HDCD at least provided some measurable benefit to the standard playback chain because the PMD-100 and PMD-200 filters were better than many other filter ICs of that era.
Just like HDCD, it's a solution to a problem that doesn't exist. HDCD at least provided some measurable benefit to the standard playback chain because the PMD-100 and PMD-200 filters were better than many other filter ICs of that era.
Last edited:
Not sure if anyone has linked this, but I think it sums up their business model pretty well:
https://www.linn.co.uk/blog/mqa-is-bad-for-music
https://www.linn.co.uk/blog/mqa-is-bad-for-music
HDCD patent is expired
try google search?
https://www.google.com/patents/US5479168
HDCD disks revealed - General Hi-Fi Discussion - StereoNET
try google search?
https://www.google.com/patents/US5479168
HDCD disks revealed - General Hi-Fi Discussion - StereoNET
I could also ask Keith about it but its not important since Microsoft isn't interested. They wanted the patents for the watermarking. The reference danced right past the fiter issues and HDCD encode and decode does synchronize the filters. Keith, Phlash and Mike spent months tuning the filters by ear for maximum transparency. I'll check on the claim that the Pacific Microsonics gives the engineer the options on the encoding. I had not heard that before. I know it does watermark the LSB even at 24/192 but doesn't engage any of the other features.
Yes that is the thing. I am sure that pretty much everybody in that demo with you did clearly hear it! That is how perception and psycho-acoustics work.
And that is also why these reports are worthless.
All the studies included in the meta-study I mentioned, IIRC about two dozen, were controlled in some fashion, and thus scientifically sound and relatively trustworthy. The meta-study on these studies was clear: well-trained listeners have a 60% chance of hearing a difference. This is important for Joe Consumer. What you as an individual think you hear in a setting heavily skewed towards the agenda of the presenters is really irrelevant for the rest of us.
Jan
Last edited:
Now when talking about future of the industry I feel that the youngest member of the Stereophile staff shows the most integrity. Not only has she published her audiology results but actually bothered to get her ears properly cleaned before attending a hifi show (ok she was also getting ear moulds done, but still...)
https://www.stereophile.com/content/janas-first-morning-la-show
OK still silly priced stuff that I will never even consider buying she is writing about there, but a small start... Sadly I think the next youngest stereophile contributor is 30 years older than her.
https://www.stereophile.com/content/janas-first-morning-la-show
OK still silly priced stuff that I will never even consider buying she is writing about there, but a small start... Sadly I think the next youngest stereophile contributor is 30 years older than her.
HD
Although I am no longer young and my hearing is not quite as good as it used to be despite having taken pains to protect it, I have and continue to have experience with both live music (just saw Bent Knee live again...amazing young people!) and hi-res files vs. CDs, that obsolete format of which I had a hand in making close to a billion (no typo). With this caveat here is my opinion, as if any of you need another...
The importance of the job of transferring from original sources to the replication or distribution format has been woefully under-appreciated since the early LP days. As an example, in the 1980's we received safety masters from Atlantic of the Led Zeppelin catalog for replication and I was so looking forward to hearing them..and they SUCKED. No way to say it nicely. I brought in some of my VG condx original vinyl discs to compare and the newer 1630 digital masters contained a large increase in distortion, clipping due to too high a transfer level, and the noise floor was very dirty.
Having had years of interaction with record labels, I can tell you most transfers of catalog are performed by low-paid "engineers" if you can call them that, who are given a list of masters to pull and a schedule for release. The equipment they use is unsophisticated and pre-set for average transfer levels with limiting and compression and in some cases equalization depending on intended final delivery media. I have watched as these titles are transferred and in many cases the "engineer" present was reading a book or magazine while the transfer was taking place, seemingly having no interest in audio quality.
How is this relevant to our discussion? The end listener, and this applies to almost everyone in this forum, has no way of determining where the sources came from, what their history was in terms of transfers, and yet many compare them as if they are finding out truths regarding the storage medium on which they are contained.
IMHO, the very best thing about linear HD files is the fact that the best ones are the result of a skilled transfer engineer going as many generations back into the vault as they can, and transferring them without spectral or dynamic limiting of any kind. This gives us, the listener, the best listen into what the original source may have sounded like. Since some of these recordings when not dynamically limited exceed 16 bits of dynamic range a 24 bit delivery format is justified. Even when it is not from a pure dynamic range perspective, it eases the constraints of setting the transfer levels perfectly to avoid clipping or excessive noise.
In the best transfers, the end results can be amazing compared to earlier transfers, but I would have trouble arguing the difference is the 24 bit or 96/192 sample rate format, but I really don't care, I just want the music. Storage has become stupidly cheap so storing these files is just not much of a problem. Efforts at reducing file size are still appropriate for streaming and smaller storage media, but then we are not talking about a function which optimizes quality over other factors.
Just my <$0.02 worth...
Howie
Although I am no longer young and my hearing is not quite as good as it used to be despite having taken pains to protect it, I have and continue to have experience with both live music (just saw Bent Knee live again...amazing young people!) and hi-res files vs. CDs, that obsolete format of which I had a hand in making close to a billion (no typo). With this caveat here is my opinion, as if any of you need another...
The importance of the job of transferring from original sources to the replication or distribution format has been woefully under-appreciated since the early LP days. As an example, in the 1980's we received safety masters from Atlantic of the Led Zeppelin catalog for replication and I was so looking forward to hearing them..and they SUCKED. No way to say it nicely. I brought in some of my VG condx original vinyl discs to compare and the newer 1630 digital masters contained a large increase in distortion, clipping due to too high a transfer level, and the noise floor was very dirty.
Having had years of interaction with record labels, I can tell you most transfers of catalog are performed by low-paid "engineers" if you can call them that, who are given a list of masters to pull and a schedule for release. The equipment they use is unsophisticated and pre-set for average transfer levels with limiting and compression and in some cases equalization depending on intended final delivery media. I have watched as these titles are transferred and in many cases the "engineer" present was reading a book or magazine while the transfer was taking place, seemingly having no interest in audio quality.
How is this relevant to our discussion? The end listener, and this applies to almost everyone in this forum, has no way of determining where the sources came from, what their history was in terms of transfers, and yet many compare them as if they are finding out truths regarding the storage medium on which they are contained.
IMHO, the very best thing about linear HD files is the fact that the best ones are the result of a skilled transfer engineer going as many generations back into the vault as they can, and transferring them without spectral or dynamic limiting of any kind. This gives us, the listener, the best listen into what the original source may have sounded like. Since some of these recordings when not dynamically limited exceed 16 bits of dynamic range a 24 bit delivery format is justified. Even when it is not from a pure dynamic range perspective, it eases the constraints of setting the transfer levels perfectly to avoid clipping or excessive noise.
In the best transfers, the end results can be amazing compared to earlier transfers, but I would have trouble arguing the difference is the 24 bit or 96/192 sample rate format, but I really don't care, I just want the music. Storage has become stupidly cheap so storing these files is just not much of a problem. Efforts at reducing file size are still appropriate for streaming and smaller storage media, but then we are not talking about a function which optimizes quality over other factors.
Just my <$0.02 worth...
Howie
Thanks, Hoyt, nice post.
Indeed, you've hit the proverbial nail on the head about why people seek first pressing... it's a shame isn't it? And to think Apple wants to release HD content at some point... Is it even a good idea if the sources are all from "safety" masters that are already like MP3's?
Indeed, you've hit the proverbial nail on the head about why people seek first pressing... it's a shame isn't it? And to think Apple wants to release HD content at some point... Is it even a good idea if the sources are all from "safety" masters that are already like MP3's?
Howard
Your contribution worth some thousands but who is going to pay them ?
Thank you again !
George
Well, what can I say? I heard a difference, it was a small difference, but it was a positive difference in any case. Was I prompted? No. Did I discuss my listening experience with the principals at the time? No. Do I have MQA in my system at the moment? No. However, does the MQA concept and my listening experience make me feel that MQA is a worthwhile improvement? Yes.
Hi John,
I accept your analysis of the presentation as you stated it. I do wonder what accounts for the difference, however. By their own admission the majority of the MQA process is one of lossy compression and re-purposing the lower 6 bits as subcode data storage area.
The only factor relevant to improving sound quality seems to be their claimed "unique apodizing filter." The rest of the scheme is a music label's dream, allowing one file to be sold at various quality levels. They claim the undecoded file will play back as a 24-bit 48KHz file, but in their tech papers reveal that the rest of the 24/192KHz information is subcoded into the bottom 6 bits (?!). In order to be recognized as a valid file the data would have to be framed conventionally, so what is to keep the bottom 6 bits of data from becoming non-random noise? I'm not smart enough to know how this couldn't happen.
So John, my query would be: what comparison did they do in their demo? An MQA encoded file played back undecoded and compared to a full MQA decode? Or was it an unencoded plain 24/96KHz file compared to their full MQA codec file?
Having worked with Apogee experimenting with different dithers and filters in both A>D and D>A I have indeed heard differences in playback, so I would be interested in hearing this "new" filter without all of the other MQA functions.
I know Charles Hansen has been at the forefront of minimum phase and other filter designs, and I absolutely love the sound of his D>As, so I wonder what his take is on this issue?
Howie
I accept your analysis of the presentation as you stated it. I do wonder what accounts for the difference, however. By their own admission the majority of the MQA process is one of lossy compression and re-purposing the lower 6 bits as subcode data storage area.
The only factor relevant to improving sound quality seems to be their claimed "unique apodizing filter." The rest of the scheme is a music label's dream, allowing one file to be sold at various quality levels. They claim the undecoded file will play back as a 24-bit 48KHz file, but in their tech papers reveal that the rest of the 24/192KHz information is subcoded into the bottom 6 bits (?!). In order to be recognized as a valid file the data would have to be framed conventionally, so what is to keep the bottom 6 bits of data from becoming non-random noise? I'm not smart enough to know how this couldn't happen.
So John, my query would be: what comparison did they do in their demo? An MQA encoded file played back undecoded and compared to a full MQA decode? Or was it an unencoded plain 24/96KHz file compared to their full MQA codec file?
Having worked with Apogee experimenting with different dithers and filters in both A>D and D>A I have indeed heard differences in playback, so I would be interested in hearing this "new" filter without all of the other MQA functions.
I know Charles Hansen has been at the forefront of minimum phase and other filter designs, and I absolutely love the sound of his D>As, so I wonder what his take is on this issue?
Howie
Last edited:
Howie, does Scott's description help give insight into the utility of said apodization filter?
Howie, I was not involved in the demo, but the original recording was made live of Stravinsky's Firebird (I'm pretty sure) in 24-88K and a copy was sent to MQA for reprocessing, then the two digital copies were compared one after the other. The processed copy sounded more realistic and less 'constrained'.
Last edited:
Thank you for the repost, sure I can see that...Does anyone have insight into the claim that use of this filter pre A>D as well as post D>A results in more accurate transient or time response? I'm trying to get at the core of their claim of improved time resolution, because it cannot be improved amplitude accuracy compared to a linear 96k file, nor do they claim that.
Cheers!
Howie
IMO it is impossible. The time response (step response / impulse response) may be "different" with that filter, but not "better". And every difference in the time response would automatically be reflected in the frequency response.
So they claim their processing can improve an already digitized file? Referring to my previous post, if this is the claim the only benefit would come from the reconstruction filter, as the ringing/time distortion they talk about fixing by complimentary filters in both encode and decode would already be encoded into the file by the A>D anti-aliasing filter.
On the other hand, if the file had been anti-aliased by their special filter pre D>A as Scott implies, then use of their complimentary filter as part of the MQA process could sound different than merely sending it through a typical steep brick-wall-type reconstruction filter...maybe that was the comparison?
I have auditioned minimum-phase filters as part of the mastering encode/decode process and the results depended on the particulars...where the corner freq is set, the type of program material to be processed, etc. In one case with loud cymbal material I have heard them accentuate the high freqs. This could have been due to mis-alignment of the filters or out of band aliases summing with the in-band signals, but this was a while ago, so I am most likely out of date here...
Cheers,
Howie
- Status
- Not open for further replies.