Sped up recordings

[revanyoda777]

Are there any famous examples of sped up classical recordings? I am not too experienced in the science of recording piano, and the techniques that can make sped up performances happen; how can one tell if a recording is actually edited in this way?

My favorite interpretation of the Liszt etudes is done by Yukio Yokoyama, though I recently read some youtube comments stating it's likely sped up, though they could just be trying to disparage his technique. What do you think?

Here's a Yokoyama recording of Feux Follets

[louispodesta]

Are there any famous examples of sped up classical recordings? I am not too experienced in the science of recording piano, and the techniques that can make sped up performances happen; how can one tell if a recording is actually edited in this way?

My favorite interpretation of the Liszt etudes is done by Yukio Yokoyama, though I recently read some youtube comments stating it's likely sped up, though they could just be trying to disparage his technique. What do you think?

Here's a Yokoyama recording of Feux Follets

Starting with the rock n roll group Led Zeppelin in the late 1960's (and then later in the 1990's with classical recording artists), I have spent the better part of my adult life researching this subject.  I really do not know where to start.

In regards your recorded example, of course it is doctored, but the important part is something that never ever gets talked about on websites like this.  Therefore, if you want to contact me by PM, then maybe we can further this discussion.

I will leave you with a paraphrase from Alicia de Larrocha who stated the following when asked about her preference, in regards the difference between recorded and actual performance:

She said, in so few words, of course I prefer the live performance.  The recorded is fakery.

[quantum]

May I suggest trying it out in Audacity and making observations.  
https://www.audacityteam.org/

A good musician will know that to change the tempo of a piece also requires changes in elements such as phrasing, articulation, pedaling, and breathing to name a few.  Some of these may not translate well in an altered recording.  This is just one of many points that can be used to scrutinize a recording.  

***

It is not necessary to double post the same topic in multiple boards.  People here tend to look around. 

[lostinidlewonder]

Lol, well I think that Liszt you posted from Yokohama is sped up, there are some acrobatics he is doing which i think is just humanly impossible to make as fast and clear as he is doing (like the RH double notes start of page 2, and a part which Arrau once listened to from a student who played it very fast and got so depressed he practiced and practice to try to match it). The speed is ugly imo, you miss out on a lot of beautiful details because it just flies right past us, distasteful.

[ted]

As well as blurring the beautiful details, speed fundamentally alters the perception of rhythm. This fact still hasn't sunk in with recent ragtime and stride players, who seem bent on obliterating all off-beat sublety in a desperate ninety miles an hour competition. It produces nothing but a uniform stream of notes, effectively removing syncopation altogether.

[ahinton]

Lol, well I think that Liszt you posted from Yokohama is sped up, there are some acrobatics he is doing which i think is just humanly impossible to make as fast and clear as he is doing (like the RH double notes start of page 2, and a part which Arrau once listened to from a student who played it very fast and got so depressed he practiced and practice to try to match it). The speed is ugly imo, you miss out on a lot of beautiful details because it just flies right past us, distasteful.

I'm afraid that I find it almost impossible to disgaree with this; even if it weren't sped up, the first question to ask is "what's the point?". I'm not saying that what I hear is necessarily humanly impossible but it is certainly humanly improbable; whilst the articulation is generally quite good, if the subtle harmonic and textural changes and contrasts are not brought sufficiently into relief beacuse the performance insists on velocity being the order of the day (as here), then the end result is inevitably severely compromised. This is a most wonderful piece, both as a study for pianists and as a concert work, but what I hear here does it less favours than it deserves; the requisite attempts at gossamer delicacy are not ignored so much as self-defeated and, without that kind of approach, the piece just doesn't work.

If it's actually played at the speed that one hears, it's undoubtedly a remarkable achievement in itself, but it's supposed to be a piece of music, not a mere achievement!

I once heard a young pianist give an almost note-perfect rendition (one might say extraordinary rendition!) of Chopin's Op. 10/1 at a mere whisker under crotchet = 240 and, whilst the dynamics made good sense, pedalling was sensitive and never overdone and every conceivable attempt at good phrasing was made, the sheer speed simply ruined it and indeed even managed to make the piece sound less spectacular than it is.

Best,

Alistair

[Petter]

Case in point by jazz pianist Lennie Tristano

I think it enhances the music, but he too was ostracised.

[abel2]

I know a lot of Leslie Howard's recordings from his complete Liszt set are horribly sped up. The example of the original Liszt tenth etude that I wanted to use was banned so you'll just have to trust me.....

[louispodesta]

I know a lot of Leslie Howard's recordings from his complete Liszt set are horribly sped up. The example of the original Liszt tenth etude that I wanted to use was banned so you'll just have to trust me.....

I will never forget the comment made years ago by a local wannabe classical radio station presenter.  The subject at hand was Maurizio Pollini, and his response was that there is nothing better than Pollini "live."

I have Pollini's recording of the Chopin Etudes, and there is no way this is an un-doctored straight recording.  However, in reference to a later point in this discourse (regarding drugs), I have heard Pollini play the Schumann Piano Concerto live on radio wherein he played it so fast that he actually played through the melody.

Returning to the OP's premise, just like the reference I made before to rock n roll, studio recordings for a very long time have been doctored.  Glenn Gould had a professional state of the art recording studio at his Estate, and Vladimir Horowitz (Horvitz) had Columbia Records construct one in his spacious New York apartment for his personal use.

With Horowitz, any time he felt like "running some tape," they (Columbia) would put the producer Thomas Frost in a cab and then send him over to record whatever HRH Horowitz felt like recording.  After that, if he liked it, then the two of them would argue/fight over the mixing/splicing of the final product.

Next, the untold story is of the Russians/Ukranians, in Contest, juicing a particular performances with methamphetamines.  After that, in recital, all of a sudden they do not have all of this magnificent technique.

I tell you what:  why don't you compare the recent Piano Street photographs of their favorite female pianist, with her promotional photographs on her website?  The point being is that regular users of "Speed" have a certain facial look about them, especially under the eyes.

Go ahead, take a look.  In one set of pictures, she looks like a "Babe."  And the rest:  you decide.

Hey, do you think Rock and Rollers (and "Jazzers" before them) invented this stuff.  The fakery, and everything coupled with it, all started with classical performers.

[ahoffmann]

I listened to the beginning of it and I'm not convinced that it's sped up to be honest. I also remember a live video of a pianist playing at that speed. It's very difficult but not impossible.

Actually I think he leaves out notes though. At least I can't hear both voices at the really hard part.  That is a more common form of cheating these days. When listening to various instances of the chromatic etude, I noticed this happening quite a lot.

The worst kind of cheating I've seen so far is what Paul Barton does. Among others (probably also speed alterations) he changes the volume after the recording. Apart from being an absolutely shameless way to try and appear like a better pianist, it's also terribly artificial because all individual piano notes are an organic decrescendo. So when you change the volume of the entire track, it's immediately noticeable.

Honestly I have little respect for this sort of thing. Aside from being pathetic, it gives many of us who try to surmount difficulties the wrong impression of where we stand. Especially young aspiring pianists might be discouraged or lead to believe they have some natural deficiency because everyone on youtube seems to be able to do it perfectly.

I did a studio recording ages ago and I was going to attempt editing for the first time we couldn't because I play differently every time and the sound technician said it's such a different sound, it's just not going to work. So we just resorted to playing through two or three times and picking the less messy version. And that's what I've been doing on youtube as well. So my flaws are and always will be out there for everyone to see.

[klavieronin]

I have a little experience with sound editing and it hard to imagine you could speed up a piano performance all that much while keeping the correct pitch and the tone of the piano. I know there are ways to do it but the more you alter an audio recording the more obvious it becomes so I imagine any speeding up would have to be quite subtle.

That said I don't like idea of sped up recordings. It seems a little dishonest to me. The only way you might be able to get away with it is if the musical results justified it (which I'm not sure they do in the case of that Liszt recording) and if the pianist was upfront about it.

[revanyoda777]

Thanks for the discussion everyone. I really enjoyed all the responses and didn't really mean to post to two different boards. I don't know how to delete a topic and had to leave it I guess :/

It was mentioned that speeding up a performance has the great potential of losing the fine details such as clarity and voicing. I agree with this. I prefer to play a little slower than the metranome marking most of the time because I love juicing out all the colors and bringing out the harmonies; however I don't think this is mandatory all the time.

This performance in particular has me stunned; not just because of the technique, but the total vision for the piece as a whole. Yes, some sections weren't really dwelled on for too long to appreciate, but then again maybe that's how Liszt wanted it? A torrent, a whirlwind, that's how I imagined this etude in the beginning and I really enjoy the interpretation. That being said I completely agree that speed for speed's sake is not real art

I think Glenn Gould was one of the first classical pianists to focus heavily on recording technology. I don't doubt that he did all sorts of mechanical wizardry to his recordings to enhance them. What about all the live recordings that artists release though? I heard that even these so called "sacred" performances are doctored to some extent. I don't think that Horowitz did too much to his live recordings since there were quite a few mistakes in his Carnegie hall recording.

[octave_revolutionary]

I don't think that Horowitz did too much to his live recordings since there were quite a few mistakes in his Carnegie hall recording.

Actually, you'd probably be surprised at how much Horowitz liked to doctor up recordings. For example, I read somewhere - can't remember where - that he went so far as to significantly edit his Carnegie Hall concerts, to the point where huge sections were substituted for the actual live performance. If you listen to, let's say, his rendition of "Valée d'Oberman" from his 1966 Carnegie Hall recital, you can hear what sounds like a splice at 3 fourths of the way through bar 197, and then, in the very last bar of the piece, it sounds like there is a splice between the two last chords. Now, I have no way of knowing exactly how much of this recording was really spliced, but I can't help wondering if that WHOLE chunk was a substitution, and whatever the case, it does offer some insight as to how much of a perfectionst Horowitz was, despite his admission to the contrary. I can often hear splices in his recordings of other works, not only by Liszt, but by Scarlatti, Chopin, etc. And this was Horowitz, considered by many (including myself) to be one the very, very greatest pianists of the 20th century.

[louispodesta]

Utilizing my prior rock roll reference, I begin with a very famous "live" jazz recording and then go on from there:

1)  Cannonball Adderly Live!  Not - It was a studio recording with the audience clapping added in later.

2)  The very famous KISS Double Live Album - by Gene Simmons own admission, they stunk live, so they recorded the songs in the studio, and then added in the audience track.

3)  Peter Frampton live, not!  Re-recorded by Eddie Kramer (Hendrix), all of the parts were re-done in the studio, with the audience track added in later.

4)  Gould, (I got this from my late teacher Robert Weaver) would re-write the soprano/alto and tenor/bass in his Bach recordings.  That is he would play the top with two hands and then the bottom the same, in order to bring out the separate voices.  Then, he would mix the two together.

5)  A good rule of thumb (Rudolph Serkin) is to listen to the Oboe parts in the Schumann Concerto.  Technically, they cannot modify the sound wave of this instrument in order to make it fit.  To me, it sounds like a chicken living its last moments.  Just listen to any live video recording of this piece, and you will instantly hear the difference.

6)  When the strings section sounds like they are all on methamphetamines, then that is also a clear sign of fake recording.

[louispodesta]

Oops, and I almost forgot the following:

Most classical musicians without recording experience will tell you, as personally related to me by a violist for the San Antonio Symphony:  (paraphrasing) hey, when you listen to all of these Hollywood movie soundtracks, the orchestra is always out of tune.

Well, when his own orchestra recorded a guitar concerto with Pepe Romero, the initial tracks all came back flat and out of tune, even though the orchestra was properly tuned.  I know this because the assistant manager (and producer) of the symphony was the one and only Kenneth Caswell ("Debussy Composer as Pianist," "Ravel, Composer As Pianist," et al).

He told me that it took them forever to get it right because unlike today's digital recording equipment, analog recordings naturally sound flat.  Therefore, in order to get it in tune, you have to speed up the entire process.

The resultant is that, if it is a concerto that is being recorded, then the soloist sounds like Hercules (or Xena) whatever the case may be.

[iansinclair]

All of which is why I still maintain that for classical music, the early LPs in the Mercury Living Presence series, London FFRR series, and very early Deccas LPs have something really special which later ones just don't.  These are all single microphone recordings, all recorded directly to a master disc by a cutting lathe (later ones were still single mike, but recorded on 1" tape and 15 inches or 30 inches per second -- which allowed limited splicing, but not speed variation).

Particularly the ones recorded direct to the master with a lathe -- you can't cheat.

The same thing is true for most 78 rpm (nominal) recordings, but there are several "gotchas" with those.  First place, the "fidelity" of the acoustics is problematic (not overwhelmingly so, but you do need a very good preamp to handle them well)(from the 30s on, the fidelity is astonishingly good, subject to the next gotcha).  Second, the equalization curve (for the electronic recordings) was quite different from the post WWII curves (which are not all the same, but are close enough for all but the really golden eared) and must be adjusted properly on playback.  Third, the 78 rpm was nominal; while it would be constant across any particular set, and indeed by the mid 20s really was constant at 78 rpm, earlier recordings could be anywhere from say 73 or so up to 83 -- and the only reliable way to tell is to adjust the playback speed to match a pitch.  Last (for this go around anyway!) you really must play them through a cartridge wired for or built for monaural; blending at later stages in the signal path (or worse, taking just one of two stereo channels) just won't do.  All that said, I have a copy of the Oscar Levant/Philadelphia Symphony recording of Rhapsody in Blue (Grofe full orchestra scoring)(Columbia Masterworks 251) which is stunning...

[abel2]

The worst kind of cheating I've seen so far is what Paul Barton does. Among others (probably also speed alterations) he changes the volume after the recording. Apart from being an absolutely shameless way to try and appear like a better pianist, it's also terribly artificial because all individual piano notes are an organic decrescendo. So when you change the volume of the entire track, it's immediately noticeable.

I knew there was something up with his. They seemed just a little too perfect....

[jimroof]

Oops, and I almost forgot the following:

Most classical musicians without recording experience will tell you, as personally related to me by a violist for the San Antonio Symphony:  (paraphrasing) hey, when you listen to all of these Hollywood movie soundtracks, the orchestra is always out of tune.

Well, when his own orchestra recorded a guitar concerto with Pepe Romero, the initial tracks all came back flat and out of tune, even though the orchestra was properly tuned.  I know this because the assistant manager (and producer) of the symphony was the one and only Kenneth Caswell ("Debussy Composer as Pianist," "Ravel, Composer As Pianist," et al).

He told me that it took them forever to get it right because unlike today's digital recording equipment, analog recordings naturally sound flat.  Therefore, in order to get it in tune, you have to speed up the entire process.

The resultant is that, if it is a concerto that is being recorded, then the soloist sounds like Hercules (or Xena) whatever the case may be.

Analog recordings naturally sound FLAT?  That makes absolutely no sense at all.  Tape records a waveform.  As long as the tape is running at the same speed in playback as it was when recording the pitch will be spot on.  Perhaps your engineer friend used the term 'flat' to describe some other attribute of the sound quality - perhaps less 'sizzle' or not as heavy on the reproduction of higher frequencies.  I could understand that, as tape saturation is a very real thing that is NOT present in digital recording.

I can conceive of no scientific reasoning for analog tape to record A 440 but play back anything OTHER than A 440, provided the tape speed remains constant, which it most certainly should given the engineer is working with something other than a 1/4" Radio Shack reel to reel.

[klavieronin]

Analog recordings naturally sound FLAT?  That makes absolutely no sense at all.

That's kind of what I was thinking too. And if it was flat (if the tape got stretched or something), just play it back slightly quicker to bring it back up to pitch. I do have one CD though by a well know band from the 60s on which one of the tracks is definitely flat.

[louispodesta]

Analog recordings naturally sound FLAT?  That makes absolutely no sense at all.  Tape records a waveform.  As long as the tape is running at the same speed in playback as it was when recording the pitch will be spot on.  Perhaps your engineer friend used the term 'flat' to describe some other attribute of the sound quality - perhaps less 'sizzle' or not as heavy on the reproduction of higher frequencies.  I could understand that, as tape saturation is a very real thing that is NOT present in digital recording.

I can conceive of no scientific reasoning for analog tape to record A 440 but play back anything OTHER than A 440, provided the tape speed remains constant, which it most certainly should given the engineer is working with something other than a 1/4" Radio Shack reel to reel.

Then, maybe you can explain to me why Eugene Ormandy (who recorded tons of concertos with his hometown boy Rudolph Serkin) always tuned his orchestra at 444?  The officially stated purpose was so it would sound bright.

I don't think so, especially as it relates to the recordings Serkin made late in life with other Orchestras in Europe, where the sound was not "juiced."

[richard black]

I see no reason to think that Yokoyama 'Feux Follets' is artificially speeded up. It's impressively fast but not unplayably so.

Modern software can alter speed without affecting pitch (and vice versa) and the piano works quite well. Voice is much harder to cheat: organ is highly forgiving (I've _trebled_ the speed of an organ recording, for very specific reasons, and it still sounded fine). But Yokoyama's recording was made in 1998 and the software available then was very much less capable.

The main cheat used in recordings, since the advent of tape recording in the 1950s, is simply editing. Record a difficult passage many times and select the best bits. Lots of other things are done (compression of dynamics is very common) for all sorts of reasons. Recordings from the pre-stereo era are typically more honest because the various cheats were harder to do, but they're more exciting mostly because recording was a real event, not just another thing that we all do. Can you imagine what it must have felt like to make the first recording of Beethoven 5?

I've been making recordings since the late 1970s, using analogue and digital technology, so this is a subject I feel I know a little about. I'm also one of the most dishonest editors...

[louispodesta]

Cheat is a word that I have spent my entire adult life studying, mostly as it relates to our worldwide 19th century English Socialist Educational System.  However, in regards the OP, what we have just learned is the meaning of the word "fakery."

That means (except Contest), when you listen to any artists' recording, that is supposed to be a true representation of how they perform live (absent a methamphetamine/steroid arranged quasi-recording session).

Respectfully, in regards Piano Idol Worship, those millions of you who have admired Rubinstein, Arrau, Horowitz, now have to admit (especially in regards my original thesis) that you have been (euphemistically) mislead!

This regards not only fakery recording by the supposed great concert pianists of our day, it is proof positive that they were sight reading every single recording.  For those who desire further proof of this, I offer you that data by PM.

Thanks.

[jimroof]

Then, maybe you can explain to me why Eugene Ormandy (who recorded tons of concertos with his hometown boy Rudolph Serkin) always tuned his orchestra at 444?  The officially stated purpose was so it would sound bright.

I don't think so, especially as it relates to the recordings Serkin made late in life with other Orchestras in Europe, where the sound was not "juiced."

Tuning to A444 makes the entire orchestra SHARP.  Perhaps Ormandy was looking for an edge over the competition and though a slightly SHARP recording would accomplish that. 

It would have nothing to do with an inherent 'flatness' to analog recordings.  Perhaps Ormandy just wanted to HEAR it a little brighter and a slightly sharp tuning would accomplish that.

You do know that most pianos are purposely not in tune from top to bottom... right?  They are flat at the low end and a hair sharp at the top end to make the harmonics work out properly.  And let's not even get into what happens to even-tempered instruments when played with string musicians who are more likely to play true intervals as guided by their ears.

Something else is going on with relation to the analog thing however.  Tape is fully capable of faithful reproduction and will play back exactly what was recorded without any perceivable variation in pitch.

[ronde_des_sylphes]

I can't hear any particular sign the Feux Follets was sped up. I suspect the poster who commented about judicious omitted notes may have a point. (someone, somewhere, sometime, will produce a totally fake Chopin 10/2 where they have three tracks together, lh, chromatic scale, rh double notes..)

One recording which definitely is sped up is the "Horowitz" Liszt Hexameron on youtube. If you listen carefully you can hear a slightly "swimmy" sound in the slower chordal segments, e.g. the opening. This is a readily recognisable artefact of crude tempo manipulation. (The recording is actually Leslie Howard, sped up about 15% - I don't remember the precise ratio.)

[iansinclair]

I must take a gentle difference with the comment above on tape saturation (or recording lathe saturation, for direct to disc).  This absolutely will not happen -- provided the engineer sets the levels properly to begin with.  Not all do...  There are, however, several phenomena which can produce problems with reproduction.  First, of course, there is high frequency response loss.  Taking just the recording process -- ignoring the rest of the electronics and the microphones! -- is very much a function of the recording and playback machines.  If the speed isn't high enough and the heads are not aligned properly, there can be a significant loss of high frequencies.  Second, there is a potential problem with low frequency reproduction.  This is actually harder to cope with, although it isn't a problem -- except for recording pipe organs, where it most assuredly can be.

Both of those are relatively minor in a properly done recording.  What isn't minor is clipping.  This is a problem if the levels are too high in both digital and analogue recording and to most ears is really noticeable -- and really hurts the recording.  It shouldn't happen.  It often does.  Pianos, incidentally, are among the very worst for clipping, as the initial transient when the string is struck has a lot of amplitude and high frequencies, and splatters all over the place if one isn't very careful!

[richard black]

Pianos, incidentally, are among the very worst for clipping, as the initial transient when the string is struck has a lot of amplitude and high frequencies, and splatters all over the place if one isn't very careful!

Actually piano has always been very easy to meter and, therefore, to record at an appropriate level.The transients aren't particularly violent (trust me, I've spent way too much of my life looking at them).

This regards not only fakery recording by the supposed great concert pianists of our day, it is proof positive that they were sight reading every single recording.  For those who desire further proof of this, I offer you that data by PM.

That's a strong claim and I would love to receive the PM you promise regarding it, Louis. Thanks in advance :-)

[abel2]

I can't hear any particular sign the Feux Follets was sped up. I suspect the poster who commented about judicious omitted notes may have a point. (someone, somewhere, sometime, will produce a totally fake Chopin 10/2 where they have three tracks together, lh, chromatic scale, rh double notes..)

One recording which definitely is sped up is the "Horowitz" Liszt Hexameron on youtube. If you listen carefully you can hear a slightly "swimmy" sound in the slower chordal segments, e.g. the opening. This is a readily recognisable artefact of crude tempo manipulation. (The recording is actually Leslie Howard, sped up about 15% - I don't remember the precise ratio.)

I knew that was fake......no one plays the opening that quickly....

[stevensk]

I knew that was fake......no one plays the opening that quickly....

Yes, sounds like a fake. Not just the tempo. Phrasings, expressions etc