We must change digital copies from a tape recorder ratio 100:89. The analog tape machine was incorrectly adjusted, ran a little too fast (100 seconds in 89 seconds !!).
We are now wondering if this results in any loss in quality - or if we have the same quality as if the tape had run slower in the 100:89 ratio. Because time and pitch remain in the same relationship. (96kHz 32(float))
I don’t hear any loss.
Am I right in assuming this algorithm (Time-Stretch Elastique Pro - Tape) doesn’t leave any artifacts and has the best quality?
Assuming that this algorithm is really just properly implemented re-sampling (i.e. no pitch-shifting, no time-compression) then there’s not much that can go wrong, I’d say.
I think the approach is not correct here. If the tape was played back at the incorrect speed, then timestretching is not the solution. You would need to play it (the digital recording) back at another speed. This should be done without stretching or time manipulations.
I’m not sure if it is possible in Nuendo though. I always do such operations in Rx. wavelab should be able to do this.
That said: if you hear no difference maybe it’s not a big problem. But in the archiving world a copy should be as close a reproduction as possible. Not sure what your context is.
The “Tape” algo is the equivalent to good old vari-speed. No time-stretching involved, just re-sampling.
1 Like
That’s correct, it should be slowed down and re-recorded.
That gave me another idea. If you calculate a clean recording 44kHz down and back again - down again - and up again (manipulated 4 times), you can listen to the original and the converted at the same time. There is no phase shift when both play at the same time.
Nobody notices the difference in the blind test.
Possibly Nuendo will slow down and re-record.
If at 44kHz the quadruple manipulation is imperceptible, then at 96kHz there are no problems with speech tapes.
1 Like