Transferring Multitrack Cassette Recordings To Computer

[mixsit]

I have done that before, but the first time I did it, I didn't think it through that much. So I just found a clear beat where everything hit at the same time toward the beginning of the song and again near the end of the song. Once I lined those up via time stretching, it was a snap.

One thing I did learn, though, while doing this was that if you have at least two tracks with drums on them, don't transfer them at the same time. Do one drum track with the first set and another drum track with the second set. That way, it'll be easy to align those when you're stretching them and all of the other tracks will go along for the ride. In other words, it's easier to align a left drum track with a right drum track than it is to align a guitar and bass with a drum track.

So, for example, let's say you have 8 tracks as such:

drums left
drums right
bass
elec. guitar
acoustic guitar
keyboard
vocal
harmony vocal

If you were able to transfer four tracks at a time, I would put the drums left track in one batch and the drums right track in the other batch.

I'll defer to you as I haven't had to do this, but I'd think that if there were going to be alignment fixes needed, a stereo kit pair would be the first to show even small drift effects- Haas pan shift comes to mind. It's in full swing in a MS or so.
Guitars, vocals etc, would tend to be much more resilient to a little shift, and they could likely have plenty of spots to be split and slid.

Sweetbeat's suggestion to record a tic' at the head of all the tracks, also solves for tracks that have no handy visual cues.

 

[Farview]

You can't resolve the phase differences because the time stretch algorithm won't stretch two different audio files the exact same way. It treats percussive sounds differently than steady tones. So, picking a landmark at the beginning and another at the end and stretching everything in between doesn't automatically mean that everything in the middle is lined up perfectly with all the other tracks.

Think about it, if a cassette can't play at the same speed twice in a row, it probably isn't playing at a consistent speed throughout the song. The more landmarks you use to line up, the better it will be, but the stuff between landmarks wont be exactly where it was, since nothing can compensate for the inaccuracy of the playback mechanism.

If you dump everything at once, none of this is a problem.

 

[Robus]

You can't resolve the phase differences because the time stretch algorithm won't stretch two different audio files the exact same way. It treats percussive sounds differently than steady tones. So, picking a landmark at the beginning and another at the end and stretching everything in between doesn't automatically mean that everything in the middle is lined up perfectly with all the other tracks.

Think about it, if a cassette can't play at the same speed twice in a row, it probably isn't playing at a consistent speed throughout the song. The more landmarks you use to line up, the better it will be, but the stuff between landmarks wont be exactly where it was, since nothing can compensate for the inaccuracy of the playback mechanism.

If you dump everything at once, none of this is a problem.

Good point about inconsistency within a playback. I've got a bunch of old cassette four tracks laying around, and a Fostex 280. I might try this and see how it goes. I'm pretty sure if the Fostex ever did playback consistent time, 20 year old rubber won't have improved it. What I would try is line up the tracks, divide the song into sections, and try to stretch those to get the tracks in phase. I'm assuming it's only stereo tracks I'm worried about, so the drums.

 

[sweetbeats]

Its not just stereo tracks...any track that has some source material in it the same as another track (like when the drums bleed into guitar cab mic) is at risk for phase distortion/comb filtering. You'll battle a moving target since the analog tape deck will never play back exactly the same twice. Dealing with phase issues in complex audio is not the same as a simple sine wave where two tracks simply arrive at null when you reach 180 degrees out of phase...complex audio will have diminished and augmented bands across the audio spectrum. It sounds like an infant got ahold of the eq section

 

[bouldersoundguy]

Stretching etc. doesn't sound to me like a promising route to a musical result. This is basically the same issue as tracking well to save yourself the hassle of trying to fix it in the mix. Why start the project with a deficiency like shifting time alignment?

But if you can only transfer some tracks at a time pick one track and include it as one of them. Even if you can only transfer two tracks at once put the snare, for example, on the left of every pair and do 7 pairs (not 8 since you've already got the snare 7 times over already). Stretch all the pairs so the snares match. It won't really solve the phase problem but at least it offers a stable reference point, the snare, for aligning everything. I do believe that with the right software (Celemony Capstan) you could very closely match the timing of all the tracks.

 

[Farview]

Good point about inconsistency within a playback. I've got a bunch of old cassette four tracks laying around, and a Fostex 280. I might try this and see how it goes. I'm pretty sure if the Fostex ever did playback consistent time, 20 year old rubber won't have improved it. What I would try is line up the tracks, divide the song into sections, and try to stretch those to get the tracks in phase. I'm assuming it's only stereo tracks I'm worried about, so the drums.

Assuming the drums are just 2 channel stereo, you should record them in the DAW on a stereo track. That way, when you time stretch, it is only one file, so it will be phase coherent with itself. With a separate left and right file, time-stretching them will cause phase problems too. (unless the daw has some way to group multiple tracks together for time stretching)

 

[cosmos021]

Hey everyone - thanks for the great replies! I've learned a lot just by reading this. I can't imagine what I can learn once I dig into other areas of this forum.

The fact that a tape never plays back the same way twice has pretty much swayed me to start to look for an 8+ channel interface. Sounds like there could be some headaches involved if I try to do two tracks at a time.