Home Theater Forum and Systems banner
1 - 13 of 13 Posts

·
Registered
Joined
·
192 Posts
Discussion Starter · #1 ·
Hi all,

I was recently given an old cassette tape to digitize containing a "sentimental value" recording. While monitoring, I could hear that the tape had stretched over time, so the pitching was off a little. The client didn't want any restoration done, just digitization along with burning to CD, so I didn't do anything about it (apart from archiving the file in case he changes his mind!), but I've been wondering how I would go about restoring it if necessary. I'd imagine the tape hasn't stretched uniformly, so applying a global timestretch/pitch shift probably wouldn't be too successful. Is there any easy way of going about this kind of work? It's a solo vocal with no backing whatsoever, which could simplify things a bit...

What would you do? :dontknow:
 

·
Registered
Joined
·
241 Posts
You would need to "warp" the track. Basically you manually set markers for where the beats and bars are, and then the warping process will pull everything to the grid. It's also called elastic audio in pro tools......
 

·
Registered
Joined
·
192 Posts
Discussion Starter · #3 ·
Hadn't thought of that - thanks! I'm using Wavelab, and I wasn't thinking in terms of bars/beats because this seems to have been done a tempo. Also, warping wouldn't correct the pitch simultaneously, would it? I mean, that's the whole point AFAIK - timestretching without pitchshifting, right?
 

·
Registered
Joined
·
241 Posts
You can do the warping and pitch independently - in Ableton for example, you can choose re-pitch mode, this turns off the pitch correction part, and any bits you shorten or lengthen will adjust pitch accordingly.

And even if the tempo is rubato - where the performer slows down and speeds up expressively, there are still bars and beats....they're not at a strict tempo, so the term "grid" doesn't really apply, but the process is still the same.

So you can listen carefully, place markers around the bits you want to correct, and do it by ear. Or you can make a groove template to match the rubato bar and beat locations, and use that.
 

·
Registered
Joined
·
192 Posts
Discussion Starter · #5 ·
As I feared then - a fair bit of work however I do it :scratch:

It just means I won't volunteer to do it, and that if he requests it I'll let him know it won't be particularly cheap...:no:
 

·
Registered
Joined
·
18 Posts
the best plug i've seen so far for this has been Seratio Pitch ' Time. i think they offer a free demo of their plugin. . . but i'm not sure. also, not sure they sell a version that works with wavelab.

and no, it won't be easy. . .

- nick
 

·
Banned
Joined
·
1 Posts
I use Nuendo, but if your pitch correction tool is similar, you can create an envelope pitch. Define the area to be "unstretched" and divide it into its own clip/region/segment. Open the pitch process for this clip and choose the envelope tool. Your waveform should be super imposed over the octave markings. Insert grab points and "pull" the pitch line upwards at either end. Audition it (in loop mode) to get the settings right. You must not let your pitch correction tool correct timing or it will come out wrong. This is a trial and error method, as there is no real magic tool to completely fix it.

Neil Kesterson
Dynamix Productions
Lexington, KY
 

·
Premium Member
Joined
·
1 Posts
I think in the case of the solo vocal, it might be hard to know when the right pitch correction is reached. You would have to "earball it," so to speak.
But it is an interesting problem because to restore it you would need to adjust the pitch and the time to the same amount to get to a correct restoration. I'm wondering if there is a way to modify a tape machine to output the bias frequency along with the audio tracks. By capturing that signal too, it would give you a very clear reference signal to go by when adjusting the pitch and the time.

Tim
 

·
Registered
Joined
·
192 Posts
Discussion Starter · #9 ·
Some great ideas here... thanks! :clap:

@Nick - thanks for the tip - I'll check it out.

@Neil - I use Cubase, so no great difference there! As you say, no magic tool...

@Tim - "you would need to adjust the pitch and the time to the same amount to get to a correct restoration" - that's why I'd been hoping there was some "automatic" method ;) The tape in question is a consumer cassette job, so no bias frequencies or anything like that (not that I'd know what to do with a pro tape machine if I ever got the opportunity to use one :dumbcrazy:)

Cheers all :T
 

·
Registered
Joined
·
110 Posts
Somebody invented a system where the biasing frequency was detected on playback of the tape, and that was used as a reference to pitch the whole thing back to original speed automatically. Don't know if any system has filtered down to an affordable level from the hallowed halls of the restoration agencies (eg CEDAR), but if you recorded a piece at 24/96, it would be interesting to see if you can detect anything in the 18-19kHz and upwards range on a spectrum analyser plug-in. Not an expert on biasing frequencies for different systems, so can't be of any help beyond that.

>

PS 8.3kHz for ferric cassette, 14.3kHz for chrome cassette, 19kHz for 15 IPS reel from memory, research needed for anything else...

>

Edit: Scrub that, I've got it completely round my neck -it's 30kHz upwards (some are 150kHz) not much audio frequency recording equipment will be capable of recording that, so it will be serious restoration setups that can do it...

>
 

·
Registered
Joined
·
18 Posts
Edit: Scrub that, I've got it completely round my neck -it's 30kHz upwards (some are 150kHz) not much audio frequency recording equipment will be capable of recording that, so it will be serious restoration setups that can do it...
yeah, the studer i just restored has a bias frequency of 150kHz (+/- 3kHz is within spec). i think you'd be hard pressed to get that recorded digitally. on a side note, the bias oscillator had a very 'unique' circuit design with a set of PNP transistors and NPN transistors and a variable transformer. . . the only variable transformer i've seen. . .

- nick
 

·
Registered
Joined
·
110 Posts
That's why it puts this methodogy firmly in the realms of restoration agancies, as it needs specialised, purpose-built detection equipment which can then modulate the recording or pitch-control process.

Tends to be used for uneven wow rather than consistent tape stretch, I'd guess, but a wonderful piece of lateral thinking and development on somone's part...

>
 
1 - 13 of 13 Posts
Top