pops & clicks

[Dan McDonald]

Just another question regarding cleanup when you have a disc recorded slower than it should be played (e.g., 45rpm mastering of a 78).

It seems that you may get an added bonus if you de-click the file at the slow speed as the click should cover less of the sound of the original recording, hence you interpolate a smaller portion of the recording. Is that correct?

Here's what I mean. If you have a 4-minute 78 that you record at half that speed, you end up with a file that is about twice as long as the original recording, or about 6 minutes. Each click that is introduced by the needle hitting a defect or piece of dust,etc. alters the sound for a certain length of time (say a tenth of a second, just for an example). That would cover about a tenth of a second if you were recording at the original speed, but would be something like 1/20th of a second if your recording is at half speed.

It seems you get more of the original music if you record at a slower rpm. Am I thinking correctly?

Dan

[Malcolm]

Re: pops & clicks

Dan, I have a feeling that there have to be some advantages to certain processes taking place at slower than the recorded speed, but I do not think de-clicking would be one of them in the terms of your analysis. If you imagine a speck of dirt in the track, or a scratch of given measure, then the stylus would take twice as long to traverse the disturbance running at half speed, as it would at normal speed. To my thinking, it follows that the ratio of Disturbance Time to Total Time is the same regardless of speed, therefore, no particular advantage. I can see that your idea could perhaps make it easier to find some stubborn defects this way?I?ll give it a try since I have some really elusive ones. Thanks for the suggestion.
Malcolm
P.S. you do not get more of the original music


[This message has been edited by Malcolm (edited 10-09-2002).]

[Dan McDonald]

Re: pops & clicks

I was thinking that the needle gets "knocked around" by the defect (say a speck of dust) and that takes roughly the same amount of time to settle down, whether it gets knocked at 33, 45 or 78, so if the record is travelling at half the speed, the needle has settled down at an earlier point in the record than if it is travelling at the correct speed. If that happens, then it would be able to pick up more music sooner than at the correct speed. That was what I was thinking, anyway.

Dan

[Malcolm]

Re: pops & clicks

Dan, after I wrote my reply , I thought more about it too and reached a similar thought -- at slower speeds, the inertia imparted to the stylus by any disturbance has to be less at the lower speed,, therefore the stylus should recover morequickly and get on with the business of proper tracking. So, your point has validity!
Malcolm

[scc]

Re: pops & clicks

I agree with the inertia theory. The needle should settle down quicker relative to the recorded sound, thus reducing the amount of ringing (if any) present after the click. However, I think the de-clicking would work best after the file is brought up to speed. Reason being that the click will then look more like a click with faster rise and fall times. I'm making the assumption that the de-clicking code is optimized for the characteristics of real-time clicks. I'm also assuming that the re-sampling process doesn't introduce any artifacts that distort the click. Anyway, I always de-click after applying reverse RIAA, re-sampling to the correct speed and applying the appropriate low freq EQ.

---Scott.

[Rick]

Re: pops & clicks

I had not seen this thread before, I am just catching up after being burried in code for a while.

I am not sure on this one, Scott is correct that the click detector is optimized for playback at normal speeds, so there may be some degradation of the detector at 1/2 speed. Most of the difference I think this would come from the longer click lengths you "might" get at 1/2 speed. The more samples you have to interpolate over, the worse the interpolation is.
You would certainly get longer click lengths for clicks caused by surface irregularities such as needle wear and pits or scratches. These would seem to scale linearly with speed.
Dust and dirt may cause shorter disturbances becasue of the slower speed, but I do think that the initial risetime of the clicks would be about the same. My theory is that the recording system bandwidth is actually limiting the click risetime you see. The clicks may actually get shorter if the needle is better controlled at 1/2 speed.
Sounds like this could easilly be verified by exeriment.

Rick