Andreas Koch on MQA

[davkj]

Ouch.

http://positive-feedback.com/audio-discourse/questions-answers-mqa-interview-andreas-koch/

 

[amirm]

Thanks for the link. Fair amount of hypocrisy and sour grapes there. Andrea while at Sony was behind SACD: Sony and Philips' attempt to renew the CD cash cow that had gone away with patents running out of time. It also went all out creating a format war with DVD-A at precisely the wrong time: when MP3 and conveniences of it was coming to market. And he talks about DRM? SACD's #1 claim to fame that Sony heavily advertised to other labels was that it did not allow playback on PCs! It had strict copy protection technology down to the drive. Yes you were not by license even allowed to play it let alone rip it. Now all of a sudden he gets religion over DRM, proprietary formats, royalties, etc. This was all good for him while helping Sony but is bad if Meridian tries it?

There is also of course bad blood going back to those format days with Meridian providing MLP lossless compression for DVD-A (and now to Blu-ray's Dolby TrueHD) and PCM format versus them pushing DSD.

The issues he raises may be real but he should be the very last person to write about them.

P.S. I am dubious that labels are paying for MQA. They are the kingmakers here. If there is any fee to be paid on the encoding side, they would stick it to the content distributors.

 

[Fitzcaraldo215]

Thanks for the link. Fair amount of hypocrisy and sour grapes there. Andrea while at Sony was behind SACD: Sony and Philips' attempt to renew the CD cash cow that had gone away with patents running out of time. It also went all out creating a format war with DVD-A at precisely the wrong time: when MP3 and conveniences of it was coming to market. And he talks about DRM? SACD's #1 claim to fame that Sony heavily advertised to other labels was that it did not allow playback on PCs! It had strict copy protection technology down to the drive. Yes you were not by license even allowed to play it let alone rip it. Now all of a sudden he gets religion over DRM, proprietary formats, royalties, etc. This was all good for him while helping Sony but is bad if Meridian tries it?

There is also of course bad blood going back to those format days with Meridian providing MLP lossless compression for DVD-A (and now to Blu-ray's Dolby TrueHD) and PCM format versus them pushing DSD.

The issues he raises may be real but he should be the very last person to write about them.

P.S. I am dubious that labels are paying for MQA. They are the kingmakers here. If there is any fee to be paid on the encoding side, they would stick it to the content distributors.

You know way more of the ins and outs of the overblown DSD/SACD vs. PCM wars than do I. And, there were no doubt many deeper agendas. But, yeah, there is a lot of behind the scenes garbage. Personally, though revered in some circles, I am not all that impressed by Koch and his accomplishments. Stuart is not perfect either, but he has more credibility to me. Of course, that does nothing to assure the commercial success of MQA.

And, yes, it is true that SACD's DRM was really, really tough to crack. But, as you know, I have tons of SACDs on my NAS thanks to exciting developments over the past 5 years or so.

 

[Jinjuku]

It's not lost on me the Koch wrote his own 'Let me interview my-self' and yes there may be sour grapes but it is informed with the past kings clothes.

MQA is no more solving a consumer problem than DSD did.

 

[Fitzcaraldo215]

It's not lost on me the Koch wrote his own 'Let me interview my-self' and yes there may be sour grapes but it is informed with the past kings clothes.

MQA is no more solving a consumer problem than DSD did.

Yeah, well. The consumers do not think they have have a "problem". Most are deliouously happy with lossless formats. So, only idiots, like audiophiles, could possibly want more.

My own advocacy is for hi rez over RBCD. I am not a DSD advocate. If your point is 44k/16 RBCD was sufficient, I am afraid I must disagree. OTOH, I am probably the last to argue that hi rez DSD is somehow vastly or even in any way superior to hi rez PCM, though some argue that. There may be sonic differences, but either is superior in my experience to RBCD in carefully done native recordings at those higher sampling/bit depth rates. Most hybrid SACDs will let you experience this from the same master, or even via Mch, if you are so inclined. To me, with my choice of recordings, the results are rather clear.

You are certainly entitled to your own opinion, as am I. Hi rez offers dimishing returns, no doubt. But, it still represents a sonic improvement in my view, more importantly to my ears. I do not think I care one way or another about MQA, unless it advances audio quality one way or another. It might. Until I see definitive evidence to the contrary, I am keeping an open and watchful eye on the technology.

 

[dallasjustice]

Don't streaming services pay substantial fees directly to the ISPs? IIRC, for example, Netflix has to pay up to various ISPs so their service doesn't get throttled by the ISP. If that's also true for Spotify and other music streaming services, wouldn't the bit rate matter to their bottom line too?

 

[amirm]

Cost of bandwidth has come way down but yes, it is still not free. And CDNs (caching networks in the middle that push the content to the edge) will charge for bandwidth *and* storage costs. So all else being equal, MQA is a big win for them relative to streaming uncompressed high-res.

 

[RayDunzl]

Cost of bandwidth has come way down

There's an understatement...

Remember how excited we were with a mere $0.10 cents a minute long distance calls in the 80's?

At that cost per bit (64kbit or 56kbit if you want to get pedantic) a gigabyte download would cost $12,500.00 to $14285.00

 

[Blumlein 88]

While I too thought the self interview and content were hypocritical, he also has been in a position to know why things really get done. So I think he wrote a good op-ed whatever his motivation.

 

[RayDunzl]

I had to look him up to see who he is...

 

[Blumlein 88]

I had to look him up to see who he is...

You can read his interview not done by himself here. Done by our old friend Michael Lavorgna. Complete with deceptive graphs touting the advantages of DSD over deficiencies of PCM.

http://www.audiostream.com/content/qa-andreas-koch#tUXXhTTzCBVIhfD7.97

 

[Fitzcaraldo215]

Yes. Koch has been and is a staunch DSD advocate. He and other supporters of certain other non-MQA formats, tool sets, equipment, etc. appear to be the most vocal critics of MQA. They might be proven right about the negatives of MQA or the insignificance of its sonic improvement, but guys like Koch are not objective, impartial commentators. They have conflicting and competitive business interests, and MQA represents a commercial threat to them. The same goes for Bob Stuart, of course. He has been vocal for many years in condemning DSD, which I do not believe any Merdian equipment ever supported.

It is kind of fun to watch these technical squabbles, though, whatever the outcome. I do not think it is a life/death matter as far as the future of audio is concerned.

Meanwhile, back to Koch's original self-interview. I agree with Amir. I think deep down Koch reveals he is just jealous that he did not think of MQA himself.

 

[Jinjuku]

Cost of bandwidth has come way down but yes, it is still not free. And CDNs (caching networks in the middle that push the content to the edge) will charge for bandwidth *and* storage costs. So all else being equal, MQA is a big win for them relative to streaming uncompressed high-res.

Nothing that FLAC hasn't solved vs MQA.

 

[amirm]

Nothing that FLAC hasn't solved vs MQA.

Lossless encoding gives you about 2:1 compression or so. MQA gets 4:1 compression. And the baseline CD rate can also be flac compressed although probably not with as much gain.

 

[Jinjuku]

Lossless encoding gives you about 2:1 compression or so. MQA gets 4:1 compression. And the baseline CD rate can also be flac compressed although probably not with as much gain.

At this point we are just talking pro and cons. If your only yardstick is transit costs then MP3 and it's brethren won almost 20 years ago.

 

[Ken Newton]

FWIW - I've stated in other threads my assessment that the primary performance objective of MQA is the establishment of a required time-domain optimized encoding/decoding chain. Time-domain optimized in the sense of minimizing anti-alias filter and anti-image (interpolation) filter transient response. One of the factors required toward achieving that goal is the capture of significant ultrasonic bandwidth to provide the necessary wide transition band for transient optimized (lazy) filters. In it's fullest implementation, MQA's recording and playback filters form an holistically integrated system.

Where Meridian describes these filters as 'triangular', they refer to the shape of their impulse response. It is shaped like a geometric triangle, and has no ringing behavior. When two FIR filters are cascaded their impulse responses are inherently combined via convolution. So, upon playback, the triangular impulse response of the DAC interpolation filter combines with the triangular impulse response of the anti-alias filter pre-encoded in to the music track when it was originally recorded. The net result is an quasi-Gaussian shaped overall total system impulse response. Gaussian type low-pass filters feature 'relatively' rapid frequency roll-off for a filter also featuring a non-ringing impulse response.

 

[Fitzcaraldo215]

FWIW - I've stated in other threads, my assessment that the primary performance objective of MQA is the establishment of a required time-domain optimized encoding/decoding chain. Time-domain optimized in the sense of minimizing anti-alias filter and anti-image (interpolation) filter transient response. One of the factors required to achieving that goal is the capture of significant ultrasonic bandwidth, to provide the necessary wide transition band for those transient optimized (lazy) filters. In it's fullest implementation, MQA's recording and playback filters form an holistically integrated system.

Where Meridian refers to these filters as 'triangular', they refer to the shape of the impulse response. It is shaped like a geometric triangle, having no ringing behavior. This is same impulse response as has a moving average filter. When two FIR filters are cascaded the impulse response of the two are inherently combined via convolution. So, upon playback, the triangular impulse response of the DAC interpolation filter combines with the triangular impulse response pre-encoded in to the music track when it was originally recorded. The net result is an overall quasi-Gaussian shaped overall total system impulse response. Gaussian type low-pass filters feature 'relatively fast' frequency roll-off for a filter which also features a non-ringing impulse response.

Ken - I think the "triangular" reference by MQA also applies to spectral level vs. frequency. This is something Stuart has brought up frequently. See figure 7 in this link:

http://www.soundonsound.com/techniques/mqa-time-domain-accuracy-digital-audio-quality

 

[Ken Newton]

Ken - I think the "triangular" reference by MQA also applies to spectral level vs. frequency. This is something Stuart has brought up frequently. See figure 7 in this link:

http://www.soundonsound.com/techniques/mqa-time-domain-accuracy-digital-audio-quality

Hi, Fitzcaraldo,

I don't think so. While the language used in the article isn't as clear as it might be, it, combined with AES papers I've read, leave me fairly certain that filter impulse response is being referred to as triangular. For example, figure 6 of the article you linked to describes a 'triangular sampling kernel'. This also make the most sense from a digital filter theory perspective.

The only frequency domain related triangular shape I noticed in the article's figures is utilized to plot (brown color line) the particular spectrum profile that Stuart feels potentially could hold useful musical content. Spectrum outside that bounded area appears to be considered as expendable.

 

[Fitzcaraldo215]

Ken - you might be right, and it is probably not worth arguing about. The "triangular" concept is not, after all, mutually exclusive. I first heard it used by Stuart himself in conjunction with MQA as a description of the spectral content of sounds in nature and in music. This triangular distribution is as opposed to the rectangular spectral capabilities of standard PCM as we know it.

I could not quickly find a Stuart quote, but here is Harley's explanation, heavily stoked by information from Stuart:

http://www.theabsolutesound.com/articles/beyond-high-resolution/?page=3

See Figs. 4,5,6, especially the text under Fig. 5.

 

[DonH56]

I thought it was to help restore time-based discrimination information, i.e. time alignment that helps us place sounds in space, rather than just cleaning up the artifacts in the filters? Been a while and I am foggy today (short night) so I may well be wrong. They are interrelated in any event.