Page 1 of 2

Removing Surface Cruft in Software?

Posted: Tue Feb 07, 2012 1:27 pm
by Tom Dennehy
I'm digitizing some of my father-in-law's old jazz LPs from the 50s. They clean up pretty well in most cases.

On the current title. I've had to truncate a couple tracks where damage was too severe to correct. It's a shame, because it's a really good album (Anita O' Day, "This Is Anita").

Now I'm trying to preserve a track that has an isolated problem you can hear in the attached file. It sounds like there is foreign matter on the LP surface. What is this characterized as? (E.g. Noise?) And does anyone have any recommendation as to how to remove the scraping sound in Audacity?

Full disclosure: I will be writing about this in a blog I keep about digitizing analog source material and will give credit where credit is due.

Thanks in advance!

Re: Removing Surface Cruft in Software?

Posted: Tue Feb 07, 2012 1:59 pm
by waxcylinder
Try this excellent plugin written by Steve: http://forum.audacityteam.org/viewtopic ... 42&t=62384

You might find this workflow tutorial interesting/useful: http://manual.audacityteam.org/man/Samp ... gitization

WC

Re: Removing Surface Cruft in Software?

Posted: Tue Feb 07, 2012 5:05 pm
by steve
waxcylinder wrote:Try this excellent plugin written by Steve: viewtopic.php?f=42&t=62384
Unfortunately the damage is too severe for that plug-in to help - the "noise" lasts too long. The plug-in is designed to handle relatively short bad spots rather than noise that lasts for several seconds.
Tom Dennehy wrote:What is this characterized as? (E.g. Noise?)
Yes, bad noise.
It sounds similar to very bad surface noise, but looking closely there seems to be something very odd about it. At the point where the noise starts, the left/right channel audio signals go "out of phase" but I don't see how surface damage would account for that. There's also a marked ripple at about 7.6 Hz that kicks in just before the noise starts - it's clearly visible (and reverse phase in Left/Right channels) toward the end of the audio sample that you posted.

If you record the vinyl again, do you get the same noise in exactly the same places?
Is the noise specific to that one album?
Does the noise occur in just one place on the album?
How long does the noise last?

Re: Removing Surface Cruft in Software?

Posted: Tue Feb 07, 2012 5:43 pm
by billw58
@Steve: Wow, good analysis - you need to split to mono, invert one channel then mix and render to see it.

But I think that only the noise is (partially) out of phase - the actual recording seems to remain in phase (and is thus cancelled out, even during the noise burst). If you split to mono then mix and render the noise is still there, so it is not purely out of phase. Oh, and during the appearance of the sub-sonic out-of-phase ripple at the end of the noise burst the audio does come up in volume in the reversed-mix-and-render version.

Just guessing here, but the out of phase sub-sonic noise represents vertical modulation of the groove, so could the stylus have popped out of the groove for 3 revolutions? Is it a pressing fault?

-- Bill

Re: Removing Surface Cruft in Software?

Posted: Tue Feb 07, 2012 5:51 pm
by Tom Dennehy
Pressing fault? I'm guessing my father-in-law bought it c. 1956. And, based on gnarly scratches in a couple other tracks, the album led something of a hard life. :-)

Re: Removing Surface Cruft in Software?

Posted: Tue Feb 07, 2012 5:51 pm
by steve
billw58 wrote: the actual recording seems to remain in phase (and is thus cancelled out, even during the noise burst).
Amplify the track before splitting. You will then notice that the actual recording is cancelled out in the first 0.4 second (or so) and the final 0.3 seconds, but is not cancelled out from just before the sub-sonic out-of-phase ripple kicks in, up to the end of the noise. Weird. Is there a vinyl expert in the house?

Re: Removing Surface Cruft in Software?

Posted: Tue Feb 07, 2012 5:56 pm
by Tom Dennehy
To answer an earlier question, I have only encountered this phenomenon on the album in question, and it only occurs in a < 30 second section of audio. There are two short bursts, then the sustained occurrence.

If I can do so and not run afoul of attachment limits, I'll upload the full section for more perspective.

Re: Removing Surface Cruft in Software?

Posted: Tue Feb 07, 2012 8:42 pm
by waxcylinder
Tom Dennehy wrote:On the current title. I've had to truncate a couple tracks where damage was too severe to correct. It's a shame, because it's a really good album (Anita O' Day, "This Is Anita").
You do know you can get that album on CD from Amazon.com: http://www.amazon.com/This-Anita-ODay/dp/B00000470I

and on Amazon.co.uk: http://www.amazon.co.uk/gp/search/ref=p ... ex=blended

WC

Re: Removing Surface Cruft in Software?

Posted: Tue Feb 07, 2012 8:58 pm
by Tom Dennehy
I didn't have the heart to tell my father-in-law half his LP titles were available on CD.

Sometimes people dont want perfection, they just want music to sound like they remember it.

Re: Removing Surface Cruft in Software?

Posted: Wed Feb 08, 2012 1:05 am
by Tom Dennehy
Attached is a 7-second sequence from the track containing all instances of the phenomenon.


TGD