Karajan Mahler 6 - exposition repeat = cut-and-paste?

[5against4]

i was listening last night to the Karajan recording of Mahler's Sixth Symphony, and though i was loving it (and still do love it), i began to suspect during the first movement that the exposition repeat wasn't merely a performed repeat but a LITERAL repeat, i.e. cut-and-pasted in the recording studio. So i've just spent time this morning going carefully through the waveform, and i don't think there can be any doubt about it - i've particularly paid attention to a number of tiny tics, mannerisms and accidentally off-beat notes (it was an early horn note that first tipped me off) that are replicated EXACTLY in the repeat.

Questions: First, i'm assuming i'm not the first person to have spotted this? Second: considering Karajan usually didn't perform expo repeats in live performances, is it therefore possible he took this same approach in the studio too, just performing it once and doubling it up in the edit? Third: perhaps i'm naive, but assuming this is what was done in this recording, is it (or was it) common studio practice to do this with expo repeats in standard repertoire?

If this is all widely known and previously-discussed, then of course i apologise, but i've never seen this mentioned anywhere, and i don't recall ever hearing any other recording of Mahler's Sixth (believe me, i have lots) that even remotely hinted that the expo repeat was a LITERAL cut-and-paste job. But now it's got me wondering...! And what other pieces too.....?

[Roasted Swan]

i was listening last night to the Karajan recording of Mahler's Sixth Symphony, and though i was loving it (and still do love it), i began to suspect during the first movement that the exposition repeat wasn't merely a performed repeat but a LITERAL repeat, i.e. cut-and-pasted in the recording studio. So i've just spent time this morning going carefully through the waveform, and i don't think there can be any doubt about it - i've particularly paid attention to a number of tiny tics, mannerisms and accidentally off-beat notes (it was an early horn note that first tipped me off) that are replicated EXACTLY in the repeat.

Questions: First, i'm assuming i'm not the first person to have spotted this? Second: considering Karajan usually didn't perform expo repeats in live performances, is it therefore possible he took this same approach in the studio too, just performing it once and doubling it up in the edit? Third: perhaps i'm naive, but assuming this is what was done in this recording, is it (or was it) common studio practice to do this with expo repeats in standard repertoire?

If this is all widely known and previously-discussed, then of course i apologise, but i've never seen this mentioned anywhere, and i don't recall ever hearing any other recording of Mahler's Sixth (believe me, i have lots) that even remotely hinted that the expo repeat was a LITERAL cut-and-paste job. But now it's got me wondering...! And what other pieces too.....?

Famously in Karajan's recording of Heldenleben, when a split note was pointed out to the great man, he was rather scathing about not believing people would notice and/or be bothered.  But of course people did notice....  One hundred percent Karajan would have sanctioned a recording containing the repeat but whether or not he again thought the us lesser mortals wouldn't notice a cut and paste I don't know.  Or perhaps, they did record the repeat but there were (relative) flaws in it which made the editor opt for a cut and paste.  Of course back in the day of analogue masters this would be a considerably trickier process since there would be an immediate degradation in the master if a copy had to be made to be inserted.  Another possibility I guess is than at the point of creating a digital master - post Karajan's death - a cut and paste version was inserted for reasons unknown.  It would be interesting to go back to an LP original and see what happens there....

Of course ALL recording - studio or live - is a Black Art designed to create the illusion of a continuous/coherent performance.  I am sure there are all kinds of audio sleights of hand in many famous recordings along exactly these lines

[5against4]

Of course ALL recording - studio or live - is a Black Art designed to create the illusion of a continuous/coherent performance.  I am sure there are all kinds of audio sleights of hand in many famous recordings along exactly these lines

Absolutely, and i don't have any issues with that - except when the edits are so blunderingly obvious that they ruin the experience (which isn't the case with this Karajan).

It's interesting regarding the analogue aspect: this was apparently recorded in 1975/1977, so therefore before Karajan's digital era, and this perhaps explains why the repeat doesn't seem to have the EXACT same waveform as the first time, since the CD is a digitisation of two different analogue renditions (albeit of the same thing). As a consequence, the repeat also lasts a fraction of a second shorter, presumably for the same reason. But there's no noticeable reduction in sound quality - so perhaps there has been some jiggery pokery going on in the digital remastering stage, though this would surely involve completely re-compositing the audio, in effect creating an all-new version rather than what was originally issued. Hmm, going to have to try and find a rip of the original vinyl to compare...

[Biffo]

Not quite on the same scale, Barbirolli's recording of Mahler 5 was found to have a missing note in the horn solo in the scherzo. I am not sure when this was noticed but it was after the recording was released. Years later, when digital jiggery-pokery was available, EMI took Nicholas Busch back to the original venue and recorded the missing note (or possibly the passage in error). The note was then inserted into the recording and appears in later reissues of the disc.

[aukhawk]

i was listening last night to the Karajan recording of Mahler's Sixth Symphony, and though i was loving it (and still do love it), i began to suspect during the first movement that the exposition repeat wasn't merely a performed repeat but a LITERAL repeat, i.e. cut-and-pasted in the recording studio. So i've just spent time this morning going carefully through the waveform, and i don't think there can be any doubt about it - i've particularly paid attention to a number of tiny tics, mannerisms and accidentally off-beat notes (it was an early horn note that first tipped me off) that are replicated EXACTLY in the repeat.

The first and obvious point is that Karajan's Mahler 6 was an analogue recording and the technology to copy-and-paste (as we now use that expression) simply did not exist then.

There was of course the option to copy a passage of music onto another tape recorder, then cut that and splice it into the master.  Or, to make the original recording onto two (or more) tape machines running simultaneously (no doubt DG did this) and cut the passage out of the backup recording and into the master.  These and other options had their pros and cons but were all inferior (technically) to the digital copy-and-paste that is routine today in word processors, forum posts and audio workstations alike. 

If on close inspection the micro dynamics (relation between leading edges of various instruments playing together) seem the same then undoubtedly this (analogue copy) is what was done.  Closely inspected, the waveforms will never be the same because each tape recorder has its own fingerprint - frequency response, distortion, noise - and however you do it in the analogue age, the process requires more than one tape machine.  I don't know that Karajan would necessarily have been consulted on the matter - it's a production decision - but if he was and there was a good case for it, of course he would have agreed to it.  'A good case' would typically be on economic and logistical grounds - recording time with an orchestra is not an unlimited thing.  Decca record producer John Culshaw describes similar 'shenanagans' in his book Ring Resounding about the recording of Solti and the Wagner Ring - which pre-dates Karajan's Mahler by a few years.  He also goes into a lot of detail about the economic and time constraints during that major recording project - it's a fascinating read although he (Culshaw) comes across as rather fawning and sycophantic towards his major artist(e)s - part of the job no doubt.

Come the digital era, and of course the ability to losslessly copy-and-paste was a revolution in audio production terms, and this and a general approach of short takes spliced together (also described in detail by John Culshaw in his book) became ever more routine.  The relatively recent trend towards 'Live' recordings (where 'Live' confers some kind of extra virtue) is a reaction to this perceived process of 'artificial' performances and recordings.  Of course even 'Live' recordings are now subject to the same artificial processes, but presumably on a lesser scale.

[Spotted Horses]

I think in the 1970's it would have been unusual to perform a long orchestral movement in one take and it was common practice to use basic tape editing techniques splice on an exposition repeat. In the earlier days editing would have been done with a razor blade and adhesive tape. Perhaps by the 70's they had a more refined bag of technical tricks. I find this regrettable. We are hearing the same music twice and it seems an opportunity is missed to vary the interpretation the second time around.

[aukhawk]

A composer - be it Josef Haydn or Philip Glass - may enclose a passage of music in "repeat" marks.  From a composer's point of view, this is done for reasons of economy, and I suppose to reduce page-turning for the performer.  From an interpretive point of view, the most literal meaning is simply "play the same again" and it would be difficult to argue against doing just that.

(As a retired sound engineer whose career spanned the last 2 decades of analogue and the first 2 decades of digital, I don't recall any advance on the razor blade and sticky tape.  Which worked very well, but there was no way to edit with an overlap - or crossfade - without dropping a generation, which was itself a big deal and to be avoided wherever possible.  Also things got difficult towards the end (of analogue) when multitrack tape machines used 1" or even 2" tape - that stuff was very difficult (and expensive) to edit physically.
Digital edits are not only non-destructive and so very easy even in unskilled hands, but routinely use overlaps to smooth the join - not possible with tape.)

[relm1]

A composer - be it Josef Haydn or Philip Glass - may enclose a passage of music in "repeat" marks.  From a composer's point of view, this is done for reasons of economy, and I suppose to reduce page-turning for the performer.  From an interpretive point of view, the most literal meaning is simply "play the same again" and it would be difficult to argue against doing just that.

(As a retired sound engineer whose career spanned the last 2 decades of analogue and the first 2 decades of digital, I don't recall any advance on the razor blade and sticky tape.  Which worked very well, but there was no way to edit with an overlap - or crossfade - without dropping a generation, which was itself a big deal and to be avoided wherever possible.  Also things got difficult towards the end (of analogue) when multitrack tape machines used 1" or even 2" tape - that stuff was very difficult (and expensive) to edit physically.
Digital edits are not only non-destructive and so very easy even in unskilled hands, but routinely use overlaps to smooth the join - not possible with tape.)

You could do crossfades back then, those are in the beatles studio recordings from the 1960's and I'm sure it wasn't invented there.  You would have tape reels mix down and could crossfade with volume faders into a new source.  For the Beatles, having a separate tape reel running allowed for the orchestra to be recorded four times. It was then taped a fifth time, onto track four of the first reel, giving the equivalent of 200 session musicians that were then all mixed down into new source.  Also, editors would just need to find good spots for the splice if they needed to switch between takes.  With digital crossfades, you can experiment much easier and undo over and over to find a good and natural "breath" spot to do the cut and obviously with tape, it would be destructive so you wouldn't do that with the master but a copy and look for places like a natural end of a phrase or something where the transition to a new take would be least noticeable. 

As a reference, I'm using this from 1967, George Martin, who was a classical producer as well, worked with the Beatles on the late recording. A 40-piece orchestra was brought into EMI Studios on Feb. 10, 1967, to bring to life Lennon and McCartney's musical visions, which they mapped out as an improvisational avant-garde buildup. George Martin wrote a score based on those ideas, after insisting that the classically trained musicians wouldn't be able to play a piece like the one they envisioned, and the results were stitched together from four different recordingst, giving the orchestral swells the sense of heightened drama the song's composers wanted.  This was a common technique and was done in film and studio work for decades earlier as well. 

[Roasted Swan]

From an interpretive point of view, the most literal meaning is simply "play the same again" and it would be difficult to argue against doing just that.

Really cannot agree with that.  If you are going to take the repeat then there is every reason to find subtle little differences in phrasing or dynamics.  OK there was (historically) a structural reason for an exposition repeat but to imply that musicians just trot out a carbon copy of anything ever is to fundamentally not appreciate the nuances involved in making music!

[Brian]

Not to interrupt, but just want to say this thread has been fascinating and one of the most informative reads on GMG this year. Thanks, all. Okay, carry on!

[Roasted Swan]

Just done a quick, deeply unscientific, test/comparison of 3 Beethoven 5's 1st movement.  Mainly because the exposition is so short.  No real surprise Immerseel, Karajan and Solti all take the repeat and they are "different" versions of the "same" music.

But then moving onto Dvorak's Symphony No.1 which has one of the longest exposition repeats in the repertoire....  Not many conductors take the repeat but some do.  Lining up Jarvi - the 2 versions are very similar indeed but perhaps JUST enough of a difference to suggest he did play the music twice.  Rather disappointingly the oh so famous Kertesz set - the repeat is 100% a cut and paste of the same recording.  Now referring to the earlier discussion about how tricky this was to do in the days of analogue editing and easy to do digitally it would be interesting to compare an LP version or even the newest CD/Blu-ray remastering.  My CD copy is of the first CD box set release.  Back in the day the competition for a complete cycle with Kertesz was Rowicki on Phillips also with the LSO.  Rowicki also takes the repeat and his is just different enough to make me think its an actual repeat.

I'd love to know when and by whom the decision was made to make a "mechanical" repeat with Kertesz.  On a slight tangent but on the theme of artistic choices being made on the production desk - in a recent article in the RVW Society newsletter producer Andrew Keener tells a story how he had "cleaned up" a spread chord in a Vernon Handley recording of a RVW Symphony which Handley questioned on hearing the edit as he knew he would not want that chord played that way.  So Keener had to (willingly) admit to his tinkering which he thought made the moment better..... For example Kertesz was notoriously underprepared for these recordings (the story of him learning the scores on the plane to the sessions used to do the rounds).  No idea if that is true BUT it does suggest that he possibly would not be that bothered if the producer touched things up after the event....?

[Brian]

For example Kertesz was notoriously underprepared for these recordings (the story of him learning the scores on the plane to the sessions used to do the rounds).  No idea if that is true BUT it does suggest that he possibly would not be that bothered if the producer touched things up after the event....?

An interesting test would be the Sixth Symphony, a movement in which Kertesz has the repeat in place, and there is an unfakeable transition passage back to the repeat. It would be interesting if they recorded one take through that transition, then recorded the alternative transition which leads from the exposition to the development, relying on editors to splice in the repeat itself. That would come pretty close to proving "intent."

[5against4]

(Apologies if it seemed i bailed on my own thread, but i've had major laptop issues for the last few days...)

It's been interesting to see the various replies and opinions in response to what i noted in the Karajan recording. My own reaction to this kind of cut-and-paste practice has been interesting, mainly as i've not encountered it before (which, considering for how many decades i've been listening, seems surprising!). My gut instinct is that i love the Karajan 6; in contrast to his rather confused Mahler 5, it's robust, muscular, incredibly focused, immaculately performed, etc. etc. - yet the awareness of that literal exposition repeat has become a kind of existential problem for me. On the one hand, that first time through is incredible, but hearing it exactly the same for a second time, in my view, goes entirely against the spirit of what's intended in the score. No composer, or conductor, surely, would intend for the orchestra to play an expo repeat IDENTICALLY to the first time? There are so many opportunities for nuanced differences that it would be rather unthinking simply to "do it again". So this has to a limited but appreciable extent diminished my enjoyment of that movement, which may seem unreasonable considering i regard the first time as immaculate, but there it is.

Thanks again for all the responses so far, it's been really interesting to read your thoughts about this (keep them coming, it's a fascinating area). i'm definitely going to examine a few more expo repeats in both other Mahler recordings and wider repertoire - i'm genuinely curious to what extent this practice is shared in other music. Considering the Karajan is DG, might be interesting to start with that label...!

[Spotted Horses]

Just done a quick, deeply unscientific, test/comparison of 3 Beethoven 5's 1st movement.  Mainly because the exposition is so short.  No real surprise Immerseel, Karajan and Solti all take the repeat and they are "different" versions of the "same" music.

Which of the 4 studio Karajan 5's are you talking about? 55, 62, 77, 83? My impression is that the exposition repeats are identical enough that they could very well be cut and paste jobs, based on 62 and 77, but I have not specifically compared.

My impressions of how widely cut-and-paste was used is mostly based on headphone listening of CD releases from the early days, when they would not usually go back to "session tapes" but would digitize the master used to cut the LP. With LP surface noise and loudspeakers the cuts were typically not noticeable. However, listening to the more pristine CD release on headphones I would sometimes notice obvious edits, typically because the reverberation doesn't quite match up. In later re-masters they would go to the session tapes and re-do the razor and tape edits digitally and they would become difficult to notice. Typically the cuts were made at a point where the music pauses and starts again, and were probably recorded as separate blocks with the intention of splicing.

My understanding is that the most common recording technique today, with dozens if not hundreds of microphones, is to facilitate splicing or even replacing a single orchestral part (i.e., the oboe player squeaks, and they substitute the oboe line from another take). Also, having close microphones isolating instruments and separate microphones to capture reverberation allows them to cut from one take to another, while allowing the reverberation tracks to overlap, obscuring the edit.

[Brian]

My understanding is that the most common recording technique today, with dozens if not hundreds of microphones, is to facilitate splicing or even replacing a single orchestral part (i.e., the oboe player squeaks, and they substitute the oboe line from another take).

Or to record the final rehearsal before recording a weekend of live performances, to patch out coughs and audience shuffling. (The CEO of BIS has a good story about having to edit out a moment in a live recording where he, up in the control room, shouted the f-bomb.) I only mention this because there's a new release coming in New Releases which proudly brags that it is "unedited" live tapes - quite a debatable point of pride, one would think.

[Roasted Swan]

Which of the 4 studio Karajan 5's are you talking about? 55, 62, 77, 83? My impression is that the exposition repeats are identical enough that they could very well be cut and paste jobs, based on 62 and 77, but I have not specifically compared.

My impressions of how widely cut-and-paste was used is mostly based on headphone listening of CD releases from the early days, when they would not usually go back to "session tapes" but would digitize the master used to cut the LP. With LP surface noise and loudspeakers the cuts were typically not noticeable. However, listening to the more pristine CD release on headphones I would sometimes notice obvious edits, typically because the reverberation doesn't quite match up. In later re-masters they would go to the session tapes and re-do the razor and tape edits digitally and they would become difficult to notice. Typically the cuts were made at a point where the music pauses and starts again, and were probably recorded as separate blocks with the intention of splicing.

My understanding is that the most common recording technique today, with dozens if not hundreds of microphones, is to facilitate splicing or even replacing a single orchestral part (i.e., the oboe player squeaks, and they substitute the oboe line from another take). Also, having close microphones isolating instruments and separate microphones to capture reverberation allows them to cut from one take to another, while allowing the reverberation tracks to overlap, obscuring the edit.

It was the '77 Beethoven 5 - all I would say with Beethoven is that everyone (well I've never heard a performance that doesn't) does take the repeat and its very short so there is a kind of organic whole.

[Roasted Swan]

An interesting test would be the Sixth Symphony, a movement in which Kertesz has the repeat in place, and there is an unfakeable transition passage back to the repeat. It would be interesting if they recorded one take through that transition, then recorded the alternative transition which leads from the exposition to the development, relying on editors to splice in the repeat itself. That would come pretty close to proving "intent."

The "first times bars" are often quite extended but even when they are short - as in Dvorak 9 - they can't be 'faked' at the desk.  I would not be at all surprised if there have been many occasions where those repeat sections have been recorded just to keep options open.  My feeling from a musical point of view it that they should be taken.  If you stick with a composer such as Dvorak - by the time he came to write his 9th Symphony he was an experienced mature composer who was quite happy to experment with formal and musical aspects of his work.  The fact that he still chose to write repeats meant that they had a function for him that went beyond 'tradition' and performers should respect that....

[Roasted Swan]

Or to record the final rehearsal before recording a weekend of live performances, to patch out coughs and audience shuffling. (The CEO of BIS has a good story about having to edit out a moment in a live recording where he, up in the control room, shouted the f-bomb.) I only mention this because there's a new release coming in New Releases which proudly brags that it is "unedited" live tapes - quite a debatable point of pride, one would think.

The fact that final rehearsals are recorded is different I think from a specific "unedited live concert" scenario.  The rise in the use of performances recorded in concert is a simple economic necessity and is understood to be that.  Most people realise that what you hear is an amalgamation of several performances plus a rehearsal sequence - how else is applause avoided without a sharp cut on the last chord!  Also, the phenomenal quality of orchestral playing is such now that orchestras regualrly play the hardest repertoire with relative ease and can record in concert works that would have had to be recorded painstakingly in the studio.  There is then the added argument that live performances have a frisson that the studio cannot match.  I fnd that last point harder to agree with - the truth is some performers/conductors are able to create electricity more in a studio than others - that's always been the case.  Ultimately any modern performance can now be manipulated to a far greater degree than was ever the case before and for me the danger is that absolutely flawless playing becomes the expectation - even when that "photo-shopped" musical perfection is the result of technical air-brushing rather than superhuman execution.  Of greater worry to me is that personality in performance is often secondary to (perceived) technical perfection.

[Valentino]

I'm not pardoning this thread resurrection at all.

Very interesting and even more so because the Karajan M6 is in the pipeline for 8 track direct to lacquer transfer in the DG The Original Source vinyl series.

Thank you!

[Bachtoven]

Regarding avoiding clipping the last note in order to remove applause, when I attended a San Francisco Symphony concert conducted by MTT ages ago, he came out and announced the concert was being recorded for a CD. He admonished us not to prematurely applaud before the last note stopped reverberating and to wait until he lowered his baton—he had us "practice" this twice! Of course, he asked us to stifle our coughs as much as possible, too.