Schiit Audio boycott MQA

[Jim Audiomisc]

I have never actually listened to R3.

Ironically the thing that probably most rekindled my desire to revisit digital this time was my newfound interest in classical music, particularly Shostakovich and while I can tolerate vinyl's surface noise with most music - dance tends not to have all that many quiet passages - I can't do it with classical. It destroys it.

(snip)

So where I am is that I'm waiting for MQA or hi-res to become successful for digital audio to become "accessible" to me. My next foray into digital will most likely be a much higher-end (than the Meridian USB thing) DAC with MQA.

Right, I have some bedtime reading to do on sampling theory.

Its a shame you haven't spent time listening to R3 FM as that would have been interesting. The reason being that R3 FM has often been cited as superb 'analogue' by those who complain about the awful 'digital' failings of Audio CD. Yet it has been distributed since the 1980s via a digital (NICAM) system that has fewer bits per sample and a lower sample rate than Audio CD.

The point here is that people can easily blame the bottle for what it contains.

Radio 1 will probably be processed to death. And also I suspect often starts with source material that has been processed to death. So you've listening to crapped-up audio content, not what the delivery engineering *can* do if used with care.

Here's a suggestion. Keep your eye on Radio Times or the iplayer. Install get-iplayer. Then when you see a live concert (e.g one of the approaching Proms) download it as a 320k aac file. Play that carefully.

It is 48k based on a 48k/24 input feed. (That is the BBC standard for internal distribution) It is a mere 320kbps aac. Yet it can sound superb. Often better than many Classical CDs. You or I might wish for a 96k/24 flac version. But FWIW when I've compared source LPCM files given to me by the iplayer people for tests I've found it hard to tell them apart from the 320k received versions.

If you like Shostakovich, though, I'd suggest you find a CD of the Philips label version of the cello cons by Schiff , Maxim Shostakovich and the Bayern Radio Orch. Or any of the Denon JVC series of the symphones Rosdestvensky and USSR CSSO. See what you think.

The Rosdestvensky versions seem to have been recorded by the JVC/Denon engineers setting the levels to not peak clip, and then sitting with their arms folded though a performance. Risky, but a quite remarkable result.

BTW If you like Britten, the remasting of the War Requiem sounds a lot better to me that either the original SET LPs or the first CD issue. (Which sounded worse than the LPs.) The 96/24 may be best, but I'm not sure. Quite amazing for a 1960s recording of a vast work that covers such a wide range.

 

[davidsrsb]

Its a shame you haven't spent time listening to R3 FM as that would have been interesting. The reason being that R3 FM has often been cited as superb 'analogue' by those who complain about the awful 'digital' failings of Audio CD. Yet it has been distributed since the 1980s via a digital (NICAM) system that has fewer bits per sample and a lower sample rate than Audio CD.

I used to live in Reading and R3 was marred by multipath nasties, I was never aware of "digital sheen" or other CD effects

 

[darrenyeats]

R1 always sounded rubbish and it still does. And IMO they've always played crap music too under the pretence of being cutting edge (hint: how niche is Radio ONE really?) Horrible station.

Mark, how have you done these comparisons between 16/44 and hi-res? Have you got a hi-res file and down-sampled it with something good like SoX?

Several (I'd guess most) of my few past SACD/DualDisc purchases appear to have rather different masters on each layer/side. Often audible instantly, similar to what you describe. Yet when I down-sample the hi-res master myself, the difference seems to disappear.

E.g. http://dr.loudness-war.info/album/list?artist=Jamie+Cullum&album=dualdisc

 

[Jim Audiomisc]

Multipath was one of the real snags for FM. Curiously, very few surveys were done of its severity. The BBC did one published in WW and I think the IBA once did one. A directional antenna could sometimes help, but it was a matter of where you lived.

The other snag with stereo FM was one which, again, rarely seemed to be mentioned. In general, THD measurements on tuners used a sig get that wasn't bandwidth limited, and did the measurement using 30% mod *mono* at 300Hz or some other low frequency. That fails to show just how much more distortion you get for stereo with significant HF. In part down to the need to bandwidth limit to avoid interference.

Again, R3 tends to use lower modulation levels than other stations, so tends to avoid having such high added distortion from the FM link.

 

[sls4321]

If you like Shostakovich, though, I'd suggest you find a CD of the Philips label version of the cello cons by Schiff , Maxim Shostakovich and the Bayern Radio Orch. Or any of the Denon JVC series of the symphones Rosdestvensky and USSR CSSO. See what you think.

In 24/48 for £10:

In 24/96 for £13.50:

Stunning recorded sound quality and performances.

This Decca set on CD is also excellently recorded and mastered

It is because of the quality of recordings like these that I have no interest in another format.

 

[music fiend]

I love this recording that includes Shostakovich's cello 1 and 2:

https://www.amazon.co.uk/gp/product/B001A3NG02/?tag=pinkfishmedia-21

Regarding the symphonies I think the Petrenko cycle is great.

 

[mark-hants]

R1 always sounded rubbish and it still does. And IMO they've always played crap music too under the pretence of being cutting edge (hint: how niche is Radio ONE really?) Horrible station.

Mark, how have you done these comparisons between 16/44 and hi-res? Have you got a hi-res file and down-sampled it with something good like SoX?

Several (I'd guess most) of my few past SACD/DualDisc purchases appear to have rather different masters on each layer/side. Often audible instantly, similar to what you describe. Yet when I down-sample the hi-res master myself, the difference seems to disappear.

E.g. http://dr.loudness-war.info/album/list?artist=Jamie+Cullum&album=dualdisc

Comparisons were done with 24/192 source files down-sampled to 16/44 in Audacity and then blind tested (bit perfect e.g. 16/44 as 16/44 not up-sampled, so as it would be on a CD/Tidal) to see if I could spot which one was the hi-res, which proved to be easy.

That said, it wasn't as if the 24/192 blew me away like the best of vinyl can, but it did seem to improve the timing significantly.

A friend into digital tells me that timing failure is connected with band-limiting and ringing. That might be why the 24/192 didn't "set me on edge" like 16/44 always does.

I've gone through so many recordings recommended to me as being good examples of CD only to dismiss them all as crap one after another. It really does seem to be "something to do with 16/44".

I did hear "Graceland", we have the original CD copy from the 1980s, and I don't recall much about it other than thinking it sounded strangely muffled as if the top end had been cut off. As if it was older than it actually was. Will keep in mind as a test album.

I don't have any means of digital playback as I gave the CD player away as it was never used (I never did buy CDs, I only bought a few and then went back to vinyl) and the DAC broke and got returned.

Apart from my PC/laptop which aren't very useful in hi-fi terms despite both being 24/192 capable, just cheap circuit boards, or YouTube on the TV > DAC in the amplifier which is poor.

However I've bookmarked this to seek out those Shostakovich recordings at 32/768 when I get a new DAC And symphony #10 is sublime.

Working my way through the sampling theory now. I have some suspicions already on where this goes wrong, why MQA improves it, and why it needs to operate at both ends.

I can see myself with some analog measuring equipment (oscilloscope?!) and a digital workbench if this carries on. Come along Meridian, get on with it.

 

[Jim Audiomisc]

So far, from reading the WO Patent doc, MQA encoding seems to alter the HF frequency response and add in HF anharmonic aliases. If so: The result thus won't be the same as the source material. But you might like the effect. Decoding can correct (roughly) the response but will need to be good at unscrambling eggs to fully undo the aliasing. However the impression I've got so far from the Patent is that the argument is that aliasing sounds good, anyway.

Personally, having spent decades trying to get sources and systems that give a channel with lower distortion and a flatter response, I can't say I'm encouraged by what little I've read thus far. It would be a shame if a lot of what people buy to play on non MQA kit as 'compatable with LPCM' became effected like this. It might help the music biz to pursuade people MQA decoding sounded 'better' though. 8-]

Still, further reading may reveal a "get out of jail free card" which means no-one need worry about the above. Time and effort will tell...

 

[adamdea]

Working my way through the sampling theory now. I have some suspicions already on where this goes wrong, why MQA improves it, and why it needs to operate at both ends.

I can see myself with some analog measuring equipment (oscilloscope?!) and a digital workbench if this carries on. Come along Meridian, get on with it.

I can only stand in awe of your abilities, but for the sake of humanity I do hope you're saving some time for medical research, famine relief and alleviation or ideally reversal of global warming.

 

[darrenyeats]

Mark, not all down-sampling is equal, I think SoX is supposed to be one of the best. Perhaps you can try that. And also do some blind listening as well as sighted (a mix of both is good IMO).

 

[Jim Audiomisc]

I'd also recommend sox for reprocessing. Audacity may be as good, but I've not checked it.