Off-center records can now be corrected digitally

[Nicholas]

I searched for informations about possible software based solutions to correct off-center records, and I found two softwares : Respeed, that is included in Cambridge NoNoise solutions (no price available, but a lot of money for sure...),
and Celemony Capstan :
http://www.celemony.com/cms/index.php?id=capstan

Since it's designed to rescue the most valued classical recordings, it could also be an interesting resource in the process of digitalizing any vinyl or tape source, including our beloved Jamaican made records.
It corrects not only wow and flutter, but bigger speed issues, including off center records, probably also issues with warped records, tapes with varying speed (those sound tapes which are slowing as the batteries are low...), etc.
It's quite expensive (just USD 4458 they say !!!), but it's also for rent at USD 199 for five days.

Now, if someone is about to testing it, I would be interested to hear about the results....

N.

[kalcidis]

To be honest I think people should check out audio programming software such as pure data and then do their own solution to this. An off center recording should be able to salvage with a pitch that varies according to the rotation speed (33 rpm etc) so a sinus curve should do the work. Shouldn't be too difficult and would cost you $0. Probably already is a solution out there for free.

[Werner]

To be honest I think people should check out audio programming software such as pure data and then do their own solution to this. An off center recording should be able to salvage with a pitch that varies according to the rotation speed (33 rpm etc) so a sinus curve should do the work. Shouldn't be too difficult and would cost you $0. Probably already is a solution out there for free.

It's not that easy if you would want to do this as an automated solution, as you would have to do an analysis of the audio to determine how far off-center the record is, and so how strong the pitch modulation should be. And you would also have to automatically determine a reference point where to start the modulation cycle. In theory this is all do-able, especially with a simple signal or if you would have a reference signal, but with music it's not so obvious. What would be fairly easy to do would be to make a periodic pitch correction filter with manually tuneable "pitch modulation depth" that you could tweak until it "fits" to the ear. You would then also first have to manually find a proper reference point within the cycle. But if you think about it, this is basically the exact same procedure you use when you manually re-position a 7" on your turntable

[kalcidis]

Yeah of course it would have to be some kind of manual effort to it. But I'm thinking this is the case for those expensive software solutions too? But as you say there's two or three factors to take into account for: what RPM the record is ripped with, what cycle to use and the amplification of the cycle pitch. The first one is quite easy since I think people either rip them on 33 1/3 or 45. The other two are a lot more difficult to solve in any automated sense. There might be some super nifty solution to it but I haven't gotten too deep into audio theory/programming.

[Werner]

Oh, I'm quite sure it should be possible to make an automated system (and I think that's what the mentioned product is, check out the video in the link that was posted), but it's just not something you would easily implement as a hobby project

[zonard]

Nice one Nicholas, Capstan software seems very powerful. Must be of crucial use for old tapes with varying speed issues that are impossible to anticipate.
Dealing with off centre records is much easier since the sinusoid variations have a fix and known period of 0.55 (33rpm) or 0.75 (45rpm)
A simpler (i mean also cheaper) software might be able to fix off centre records.

Or you can fix it the analogic way for 200 times cheaper than Capstan : www.echotone.org

[jb welda]

most of that sort of software, I think, is meant to correct minor fluctuations. I would like to see it handle, say, a really offcenter Jackie mittoo studio one side. organ and keyboard with sustained notes are the hardest to correct to a decent standard.

reminds me of the time I worked for the R&D dept of this high end speaker manufacturer, and they were branching out into other sound shaping devices like graphic eq, dBx that sort of thing. someone gave us a "vinyl noise eliminator" which was pretty sophisticated for the time (late 70s) and it would continuously sample ahead and if it detected a transient (tick or pop) it would replace it with a nearby sound sample in a very small portion, like 11 thousands of a second, that was virtually undetectable to the human ear. worked pretty great with scattered clicks and pops. then someone, not me, decided to give it a real test and sandpapered one side of some dumb supertramp record and played that. the damn machine nearly had a heart attack because it couldn't find a clean sample to use and basically started smoking and went belly up with Xs over its eyes. funniest thing I think I have ever heard or seen, sound wise. bloody well right then!

one love
jb

[d2domer]

I guess the relative variation in speed would increase as the needle goes from the edge towards the center, especially on a 12" record?