Skip to main content
Topic: Slower rotation speed for digitizing to reduce distortion — plausible? (Read 767 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Slower rotation speed for digitizing to reduce distortion — plausible?

This is primarily a theoretical question — I don't have vinyl turntable at my disposal, but I'm interested to know if everybody on the Internet who digitizes vinyls does it in optimal way or it can be improved.

Let's suppose we have a record that was only released on vinyl (so that digitizing it actually makes sense).
The goal is to digitize it with best possible quality (that is, minimize amount of non-linear distortion when compared to the original master which we don't have).

All else being equal, is there any possible benefit of spinning the disk slower than the intended speed? (modding the turntable if needed)
I suppose that with slower rotation speed the needle will follow the groove more accurately since it won't have to react to high frequencies (they become lower proportionally to slowdown). And correcting the speed in software via resampling is a very accurate process as we all know, so it's not a problem if we know the actual speed with enough precision. It's also possible to deal with RIAA curves in software, too.

Am I right? Or are there drawbacks which outweigh the benefits(if any)?

Re: Slower rotation speed for digitizing to reduce distortion — plausible?

Reply #1
I think that this is essentially the same as recording a vinyl record at a high sample rate. For processing there may be an advantage, but for general play back - no.

If you're really interested in making good recording from vinyl, look at purchasing a RCM and playing wet.

Re: Slower rotation speed for digitizing to reduce distortion — plausible?

Reply #2
Although tracking at HF will be improved, tracking at LF will be impaired. Tonearm/cartridge assemblies have one or more LF resonances which are designed to be below the lowest frequencies typically found on an LP. Lowering the rotation speed is going to lower the frequencies of the LF information on the disc, moving them closer to, or into, the resonant range. 
Regards,
   Don Hills
"People hear what they see." - Doris Day

Re: Slower rotation speed for digitizing to reduce distortion — plausible?

Reply #3
But the latter can be fixed by doing another pass with normal speed, and merging the results (using lowpass/highpass filters), I guess?
It probably would be a PITA to do, because then speeds need to be matched exactly, and even small amounts of speed jitter can become a big problem. And if the final benefit is small, I can see why nobody does this.

Re: Slower rotation speed for digitizing to reduce distortion — plausible?

Reply #4
There is no advantage at all doing that.
Most of noise comes from the record itself which will be the same if you play it slower. ;D
The noise coming from the record player itself is much less if it is a good one. It is likely that it is fully below the noise floor of the record so it doesn't really matter.
- I abandoned this account since I didn't find a way to delete it -

Re: Slower rotation speed for digitizing to reduce distortion — plausible?

Reply #5
But the latter can be fixed by doing another pass with normal speed, and merging the results (using lowpass/highpass filters), I guess?
It probably would be a PITA to do, because then speeds need to be matched exactly, and even small amounts of speed jitter can become a big problem. And if the final benefit is small, I can see why nobody does this.

Sounds like a fun project. Try it and see what you can do?

Re: Slower rotation speed for digitizing to reduce distortion — plausible?

Reply #6
This is primarily a theoretical question — I don't have vinyl turntable at my disposal, but I'm interested to know if everybody on the Internet who digitizes vinyls does it in optimal way or it can be improved.

Let's suppose we have a record that was only released on vinyl (so that digitizing it actually makes sense).
The goal is to digitize it with best possible quality (that is, minimize amount of non-linear distortion when compared to the original master which we don't have).

All else being equal, is there any possible benefit of spinning the disk slower than the intended speed? (modding the turntable if needed)
I suppose that with slower rotation speed the needle will follow the groove more accurately since it won't have to react to high frequencies (they become lower proportionally to slowdown). And correcting the speed in software via resampling is a very accurate process as we all know, so it's not a problem if we know the actual speed with enough precision. It's also possible to deal with RIAA curves in software, too.

At one time half-speed cutting was all the rage.    Here is an article promoting it and also listing some disadvantages.

http://www.vinylmeplease.com/magazine/slowing-things-down-art-half-speed-mastering/

Some of the advantages of half-speed cutting will also present themselves while playing back.

For example, the effects of the frequency response of the cartridge and arm will be moved up an octave.  The good news is that the tip resonance will be moved up and possibly outside the audible range. The bad news is that the tone arm resonance will also be doubled, and is more likely to have audible effects.

The playback lateral and vertical tip velocities will be cut in half, which bodes well for trackability.

RIAA equalization would need to be applied carefully - either the corner frequencies cut in half in the realm where the frequencies are as they come off of the half-speed LP, or the normal corner frequencies applied after the flat-playback signal is doubled to make it as it was ioriginally ntended.

Audacity seems to have some features that would help with this.

Execution is pretty simple if you have a turntable with a 16 2/3 rpm speed which was at one time a standard supported speed. The necessary processing is simple enough in the digital domain.

Re: Slower rotation speed for digitizing to reduce distortion — plausible?

Reply #7
Quote
But the latter can be fixed by doing another pass with normal speed, and merging the results (using lowpass/highpass filters), I guess?
It probably would be a PITA to do, because then speeds need to be matched exactly, and even small amounts of speed jitter can become a big problem.
Right...  You're NEVER going to get the waves aligned.   If by the end of the record you're off by 1/2 millisecond you are 180 degrees out-of-phase at 1KHz.   (And, you'd need phase aligned filters.)     

If you simply digitize the same record you won't get the exact-same digital-data.   Even ignoring timing drift & electronic noise, the analog waveform won't be sampled at exactly the same places when you digitize the 2nd time.

With a reasonably good turntable & cartridge the big limitation is the record itself and there's no point in chasing perfection.    IMO - It is  worthwhile cleaning-up the "snap", "crackle", and "pop".

Re: Slower rotation speed for digitizing to reduce distortion — plausible?

Reply #8
It's not that hard to time align two recordings using unsynchronized clocks in Matlab or similar environment. I think if you're doing all the other processing you could manage alignment.

 

Re: Slower rotation speed for digitizing to reduce distortion — plausible?

Reply #9
Quote
But the latter can be fixed by doing another pass with normal speed, and merging the results (using lowpass/highpass filters), I guess?
It probably would be a PITA to do, because then speeds need to be matched exactly, and even small amounts of speed jitter can become a big problem.
Right...  You're NEVER going to get the waves aligned.   If by the end of the record you're off by 1/2 millisecond you are 180 degrees out-of-phase at 1KHz.   (And, you'd need phase aligned filters.)     

If you simply digitize the same record you won't get the exact-same digital-data.   Even ignoring timing drift & electronic noise, the analog waveform won't be sampled at exactly the same places when you digitize the 2nd time.

With a reasonably good turntable & cartridge the big limitation is the record itself and there's no point in chasing perfection.    IMO - It is  worthwhile cleaning-up the "snap", "crackle", and "pop".

Actually, time aligning multiple LP playback sessions can be pretty easy, as impulses (AKA Tics)  are excellent markers for the purpose.  Realistically,  aligning  within 550-100 microseconds should be pretty doable in many cases. That gets your noise reduction working up past 3-4 KHz where the ear is most sensitive. 

However summing samples to obtain noise reduction only works for time-varying noises, IOW noises that are at least a little different every time you play them. That means no help for artifacts like scratches and some tics that are "sticky".

 
SimplePortal 1.0.0 RC1 © 2008-2018