The Talking Machine Forum — For All Antique Phonographs & Recordings

Does anyone know the actual process involved in over-dubbing earlier records in the nineteen thirties ? An example I am thinking of is the Caruso recording of "The lost chord" which was re-published in 1934 in re-mastered form with a "new" organ background. My question is---how was the original pre WWI organ accompaniment erased in the absence of digital technology and without losing some of Enrico's voice ? I can see how an orchestra or instruments could be added but subtraction/negation of a backing is a very different and, I would have thought, difficult proposition at that time.

There wasn't any organ on Caruso's original, only orchestra. They never recorded him with an organ, although organs were recorded in the acoustic era.

They never did / were able to "erase" anything either, since of course it was disc to disc. They achieved what they achieved purely through volume / level manipulation On some of those records you can actually hear the hiss of the original record fading up and down.\

And, In my opinion, they did lose some of his voice on those re-dubbings, partly though the effort to have the new orchestra drown out the old. Makes his voice appear smaller.

Here are the two versions for comparison. The original backing is poor but I think it is actually organ but might be small orchestral as well ?

http://www.youtube.com/watch?v=O9I1R8TK-yc Original recording 1912.

http://www.youtube.com/watch?v=D3AA_GtjFb8 Nineteen thirties recording.

Maybe they were able to do it because the original backing was so poor ? It is still a lot to negate.

For the "Re-Creations" as they were called the conductor, Nathaniel Shilkret in the USA and Lawrence Collingwood in the UK, wore headphones through which he could hear the original recording being played. He had to simultaneously direct the live orchestra to follow the original, not an easy task!

In any lengthy orchestral passages the original recording would be turned down to reduce surface noise, but the problems of pitching and balance resulted in varying degrees of success. The process required intense concentration by the conductor, the musicians and the recording engineers.

Whatever your views on the results, and opinions vary widely, they kept Caruso's recordings selling well for many years after his death.

http://www.youtube.com/watch?v=CoseRUnNfOI
tHere's the Pathé news reel that is more or less an infomercial for HMV's first attemtps to rerecord the accompaniment. I believe that HMV did this first, the Victor offerings with Shilkret conducting were done a few months later. The results do vary. I think the original issue in North America was on the so-called " curtains" label. Here's the Canadian version. The Bizet is superior to Celeste Aida on the reverse side, on which Caruso sounds fairly anaemic in respect to the orchestra. Neither side really sets the world on fire.
Jim Tennyson.

emgcr wrote:Here are the two versions for comparison. The original backing is poor but I think it is actually organ but might be small orchestral as well ?

http://www.youtube.com/watch?v=O9I1R8TK-yc Original recording 1912.

http://www.youtube.com/watch?v=D3AA_GtjFb8 Nineteen thirties recording.

Maybe they were able to do it because the original backing was so poor ? It is still a lot to negate.

There is no organ, trust me. It's just the Victor Orchestra conducted by Walter B. Rogers, and recorded in Camden in 1912. Though I suppose one could mistake the woodwinds and such for sounding like a poorly recorded organ. But if they had bothered to record one, it would surely be more prominent.


O/T are there any records with pipe organ made in Camden before the Trinity Church studio opened in 1918? Any one know?

I agree, what you are taking to be an organ is the low brass substituting for the low strings. The pipe organ is an extremely hard instrument to record acoustically and was rarely tried. If Victor had successfully used one in this session I am sure they would have promoted it.

Thank you for your interesting replies and, given the early state of the technology, I think it is impressive that they managed to negate the original backing to the extent that the new recording has no trace of the old---at least as far as I am able to hear. If the original was orchestral---and thank you for correcting me---the substitution is even more impressive. Overlaying one orchestra on another might disguise the original quite well but a total change of instrument(s) is quite a triumph. I feel there must be more to this story than we have yet explained ? Were they able to play around with frequencies at that stage, for instance ?

They probably had some relatively crude EQ / filtering apparatus available, but not much, in the early 1930's.

emgcr wrote:Here are the two versions for comparison. The original backing is poor but I think it is actually organ but might be small orchestral as well ?

http://www.youtube.com/watch?v=O9I1R8TK-yc Original recording 1912.

http://www.youtube.com/watch?v=D3AA_GtjFb8 Nineteen thirties recording.

Maybe they were able to do it because the original backing was so poor ? It is still a lot to negate.

I think it is possible to hear the original backing on the new recording, especially when high notes are played.

They would have had normal filters like bass, treble and mid-range available back then (you could find them on home equipment, so why not in a recording studio?), and it would have been possible to create a tuned filter if they'd wanted to. I think they just used a high pass filter to remove rumble and a low pass filter to remove crackle, with (as Wolfe said) what today would be called "gain riding" to subdue the sections without any singing.

Apart from the thinness to the voice, the example you've posted works better than many.

I have another one (I forget which, because I've only played it once) and the "new" orchestra is out of tune with the old one and Caruso himself!

There are several Pathé newsreals that are like adverts for record companies - I wonder how much money changed hands?

Cheers,
David.