Weller LvB: A question of sound

[visualnickmos]

Just a long shot, this one:

Years ago I acquired the LvB symphony set, Weller/CBSO (MDC Strand sale, I think.) However it is not the 'original' Chandos pressing, but a pressing under the label of Musical Heritage Society. It does carry all due acknowledgement to Chandos as being the producer, as well as listing the engineer, dates, venue etc. Nothing to suggest a boot-leg copy and I'm pretty sure it's not, so that is not my query.

The thing is, that (and I am NO expert in music, pitch, sound/recording techinques, etc, etc) when I listen to it it sounds to my inexpert and totally untrained ears, ever so, ever so slightly lower-pitched than others I have. Now, this may well be my imagination (probably is, in fact!) and that it is the 'sound' of the recording, (which I have to say is excellent) produced in this venue.....??? which happens to be Birmingham Town Hall. I'd really like to know if any of you experts and better-informed amongst you may be able to shed any light on this little perplexing question. Thanks a million, in advance.

Also, I cannot imagine for one minute that Chandos would approve and sign off something like this (which carries their name) if it were not perfect.....

Nick

I have just read, and am very saddened by the news of our dear Salymap. Her stories, anecdotes and humour were a real splash of happiness and colour, as well as fascinatingly informative. She was clearly a lovely lady.
Thoughts with her family and friends.
We'll miss you,

[ferneyhoughgeliebte]

I can only second whole-heartedly your sentiments about sals, visnick.


As to your main point - when you say "lower-pitched", do you mean it sounds as if the pieces are in a lower key than the other recordings you have, or "quieter", less dynamic power?

[jayne lee wilson]

Yes, I do come across comments about faulty, wrongly-pitched transfers in reviews occasionally. Some reviewers noted that the 1982 Mravinsky DSCH 8th was issued and reissued on Phillips a semitone sharp, and when the Kancheli 3rd appeared on Olympia, David Fanning reported in Gramophone for 9/1990 that it was a whole tone sharp. He commented that the "Georgian technology must have had an off-day" and "thrilling top Gs from the horns and a stratospheric top D from the piano are all well and good, but I would rather have heard the music as played."

[visualnickmos]

I can only second whole-heartedly your sentiments about sals, visnick.


As to your main point - when you say "lower-pitched", do you mean it sounds as if the pieces are in a lower key than the other recordings you have, or "quieter", less dynamic power?

Hello Ferney. Thank you. Not quieter; in fact they are excellent recordings, - the volume is fine, but yes, it's as if the whole thing might, and I stress just MIGHT be a tiny bit lower keyed, but as I wrote, it may be just (and probably is) me. I really just don't know...

I suppose the only sure way to check is to have the original Chandos pressings and make an instant comparison; that would do the trick, but the price of the 'original' set is a bit on the high side, to say the least!

Baffled of Béziers!

[visualnickmos]

Yes, I do come across comments about faulty, wrongly-pitched transfers in reviews occasionally. Some reviewers noted that the 1982 Mravinsky DSCH 8th was issued and reissued on Phillips a semitone sharp, and when the Kancheli 3rd appeared on Olympia, David Fanning reported in Gramophone for 9/1990 that it was a whole tone sharp. He commented that the "Georgian technology must have had an off-day" and "thrilling top Gs from the horns and a stratospheric top D from the piano are all well and good, but I would rather have heard the music as played."

I wonder how this can possibly happen? It does seem strange, given that it's top-end engineering and leading audio experts and so on, who produce transfers etc. Surely there must be some form of quality check before final 'sign off?'

[Cockney Sparrow]

Also, I cannot imagine for one minute that Chandos would approve and sign off something like this (which carries their name) if it were not perfect.....

You might be able to compare the sound with the short preview of the tracks available of the movements of the Chandos set on The Classical Shop (Chandos operated) download site:http://www.theclassicalshop.net/Deta...r=CHAN%208712M

[richardfinegold]

The question of pitch in transfers is always raised with historical material, particularly with transfers from 78s. Regarding material that was originally digitally recorded, it is hard to see how that may occur

[visualnickmos]

You might be able to compare the sound with the short preview of the tracks available of the movements of the Chandos set on The Classical Shop (Chandos operated) download site:http://www.theclassicalshop.net/Deta...r=CHAN%208712M

Good afternoon CS

Thank you for the link. That has proved very useful. Brilliant, as I've just done a test, as it were, and they are exactly the same. In fact I'll carryon later tonight and listen to some more of this rather splendid set. I kinda thought Chandos - especially- wouldn't let something less than perfect out of the door with their name associated with it.

[Once Was 4]

Yes, I do come across comments about faulty, wrongly-pitched transfers in reviews occasionally. Some reviewers noted that the 1982 Mravinsky DSCH 8th was issued and reissued on Phillips a semitone sharp, and when the Kancheli 3rd appeared on Olympia, David Fanning reported in Gramophone for 9/1990 that it was a whole tone sharp. He commented that the "Georgian technology must have had an off-day" and "thrilling top Gs from the horns and a stratospheric top D from the piano are all well and good, but I would rather have heard the music as played."

I really am sorry but I am afraid that this thread has set off a totally irrelevant anecdote which used to be relayed in bandrooms during my younger days.

A very highly respected, but more than a bit eccentric, horn player – let’s call him Mr Branchstream - auditioned for a prestigious post in London (to which he was appointed). No accompanist was provided and he proceeded to play the solo part of Richard Strauss’s 1st horn concerto flawlessly and with a great deal of panache. The panel was enthralled but a couple of them thought, despite this, that there was something odd about it. The pitch did not seem correct. It wasn’t: he had never bothered to purchase the music but had learned the part from a gramophone record. And his turntable was going round too fast so he had played it a semi-tone too high (which would have made it even harder and his technique even more astonishing!)

[Cockney Sparrow]

I really am sorry but I am afraid that this thread has set off a totally irrelevant anecdote

As the original question is now answered, OT is surely not so frustrating....... I'd be interested to know how the horn player fared after the audition......

And my reason for posting is nothing like as interesting. Its just that, as a teenager, the Grieg "Hall of the Mountain King" was never quite as exciting after I corrected the turntable speed to 33 rpm (from 45, not 78!).

[visualnickmos]

I really am sorry but I am afraid that this thread has set off a totally irrelevant anecdote which used to be relayed in bandrooms during my younger days.

That's the beauty of this forum..... meanderings are always welcome

[Gordon]

Further to #3 and #7 it is hard to see where a record/playback speed error can arise in modern digital recording. Still more of a concern is that it was not picked up at source. A semitone error is about 6% and a tone 12% both quite large in engineering terms. Possible as in #7 for old 78s or even early tape machines, eg from beginning to end of reels with editing implications, but even there better than 1% would be expected.

The Chandos processing could alter the sample rates minutely but surely not by that amount; the specs for sample rate tolerances [ie 44.1 or 96 KHz etc] involve crystal oscillators that are usually good for 1 part per million. However I'd not trust a laptop's clock accuracy [or jitter for that matter] because it is not necessarily spec'd for high quality audio. A mystery then? I've tried googling the tolerance for CD sample rates and found lots about the nominal value but so far not a lot about the tolerance.

Originally the 44.1 came via Sony from a TV related recording system, in the case of European PAL the clock source is the old analogue colour sub-carrier whose tolerance was 0.25ppm!!

[Ferretfancy]

I'm not suggesting that Chandos would do it, but I understand that with digital technology it's possible to vary the pitch without affecting the tempo.

Incidentally I had one of those wonderful gadget sales mags inside my newspaper the other day, you know the sort of thing, electronic dog fluff removers or telescopic window washers. One intriguing item is a watch which is accurate to one second in 200 million years--that I must have!

[Gordon]

Indeed FF they can do that where it is necessary eg to correct bit of flat singing in a cold hall or to get a piece of audio to fit a precise time period [eg movies]. One would expect production staff to know what they are doing.

Looking further at the CD spec [assuming the OP was listening on CD not to a download - he is not specific in #1 - he has it seems checked and they are the same implying that the file is wrong] the tolerance on the track speed [the rate at which data goes onto and comes off the disc] is + or - 0.1m/s in a nominal value of 1.3, about + or - 8%. IOW the spec allows the cutting lathe [not the audio file itself] to run 8% fast [or slow] and also allows the playback system to do the same but opposite. If either end is at the extreme of tolerance then there could be a 16% error. One would expect that cutting lathe systems are professionally built and therefore accurate - but only within the spec.

In practice the CD player and cutter should be locked to an accurate crystal clock but to allow for the mechanical defects in playback [eg W&F] some slop has to exist independently of the audio file being recorded/played. The player will have a buffer store and there is a 75Hz reference in the data on disc so that the servos can get the average rate near to correct entering the buffer store leaving the store to deal with remanent jitter and precise sample rate accuracy. Once that 75 Hz is burned into the disc nothing the player can do to fix it if that 75Hz is wrong. The player will spin the disc to get what it thinks is 75Hz to prevent the store running short of data or indeed getting too full.

If there is a speed error the nominal timing of a track will be wrong too - so check the actual against claimed track durations, a semitone pitch error should show up as a 6% duration error. IOW a 10 minute track that is sounding flat is running slow and should last longer by 36 seconds. If it doesn't then the 75Hz's are near enough the same and the pitch error has another cause; this time the suspects are the source/player sample rates themselves. Most players use the 75Hz to get the duration reported in the machine's display so also check using your watch - like the super accurate one reported by FF. The duration on the disc sleeve note is what the original source audio file says it is and is in theory "correct".

Quite a watch - so how do they know? How is it calibrated? Have they waited 200M years to check? No, it's a projection from some other measurement. Put that watch close to a massive object [like a human body] or take it down a mine shaft or up in an aircraft] and Mr Einstein suggests that it will run slow/fast by an amount that breaks the spec. 1 sec in 200M years is two orders better than Caesium [as in GPS] which is only 1 in 10^14 and GPS has relativistic corrections in the satellites.

[pastoralguy]

Very interesting, Gordon. As indeed all other contributions to this interesting thread have been.