I have mulling over how I might be able to process a recording of a LP with off centre hole.
The first challenge is to detect the magnitude, phase and exact frequency of the wow.
On a test record, life is easy and we just process the ~3kHz reference tone.
On music, much harder.
An offset of 1mm means a maximum speed variation of +/- 1/58 on a 58mm inner groove. This must also cause both FM and velocity variation of the same magnitude. The velocity variation then should cause amplitude modulation at 0.55Hz. I am thinking that this could be filtered out to determine magnitude and phase?
The first challenge is to detect the magnitude, phase and exact frequency of the wow.
On a test record, life is easy and we just process the ~3kHz reference tone.
On music, much harder.
An offset of 1mm means a maximum speed variation of +/- 1/58 on a 58mm inner groove. This must also cause both FM and velocity variation of the same magnitude. The velocity variation then should cause amplitude modulation at 0.55Hz. I am thinking that this could be filtered out to determine magnitude and phase?
The late Michael Gerzon used the residue of the HF bias of the mastering tape recorder for that, but that only works when there is an analogue tape recorder involved in the recording process. Then again, digital recorders may have clock residues.
Post processing
I feel that the amplitude modulation will be tiny and require averaging acrooss a side to be measured accurately
Clever idea, but I would need a MC cartridge to stand a chance of recording the bias leakage
Last edited:
Sensing the tone-arm movement ought to be possible in various ways.
Correcting in the digital domain requires buffering and interpolation to a resampled rate clock.
The FM needs to be corrected, the AM you can live with, its well below thresholds of perception, and rather complex to correct as you have to inverse RIAA, correct and redo the RIAA.
Another distortion off-centre pressing does is to move the RIAA curve around slightly, again not worth worrying about.
A practical detail is that the correction mechanism won't necessarily be very accurate at the start of a track when the lead-in groove ends, as the tone arm motion is more complicated when this happens.
If you can measure the turntable speed accurately too the same system could cancel speed variations too.
Correcting in the digital domain requires buffering and interpolation to a resampled rate clock.
The FM needs to be corrected, the AM you can live with, its well below thresholds of perception, and rather complex to correct as you have to inverse RIAA, correct and redo the RIAA.
Another distortion off-centre pressing does is to move the RIAA curve around slightly, again not worth worrying about.
A practical detail is that the correction mechanism won't necessarily be very accurate at the start of a track when the lead-in groove ends, as the tone arm motion is more complicated when this happens.
If you can measure the turntable speed accurately too the same system could cancel speed variations too.
Last edited:
Interesting thought - the DAC broadband spectrum can be quite revealing, Fs/2 and Fs peaks may be recoverable
You could sense the pickup position by using some kind of optical position sensor. Then frequency modulate the ADC clock with this signal. The following DAC should give a corrected signal.
Hello Mark,
Yes, it is. I have developed a method to do that ~ two years ago and I filed an application for US patent late last Friday night. The idea involves a head-shell that oscillates along the proverbial "straight line" a distance slightly greater than the worst possible total runout of an LP.
Sincerely,
Ralf
