track separation on a 1974 audio cassette

[MyIdenTT]

I am very new to all this, so please forgive me if this post is not quite in the right place.

I would be glad of some advice concerning the separation of tracks on a home-recorded audio cassette that dates back to about 1974. I have been able to play most of the contents into Audacity in order to produce a digital version, but there is one segment which plays two dissociated recordings at double-speed, one of these in reverse.
I can deal with the speed and direction by setting Audacity to play at 50% speed, but I have had no success with separating the tracks. I have tried playing on a stereo machine and muting one channel. My sony tape player is an auto-reverse machine. I gather that playing with the azimuth adjustment on these machines is not advisable, so I dug out an old WM FX141 Walkman. Despite taking the adjustment to its extreme in both directions, I was unable to usefully separate the tracks. Am I dealing with one track recorded over another? Does the unusual speed give any clues as to what sort of machine this was recorded on?
The content is only speech....youngsters at play...so quality is not paramount, but I would like be able to preserve it digitally...any suggestions gratefully received.

[Rare Bird]

is it an home recorded tape! if so your probably hearing what was on the tape before you taped over with what you have now, does this sound right? theres a proper way to wipe a tape for second use & it's not taping over the original programme.

The symptoms of bleed through is a problem. this can be worsened by thinner tapes.

[Beechwoods]

Incomplete erasure of the previous recording is definite possibility, and if that's the case you're pretty stuck as far as 'removing' the 'background' recording.

If the cause of the issue is 'extreme' azimuth error, forensic audio engineers with custom heads could identify and isolate a good 'stripe' of the recording, albeit at reduced volume, avoiding the horizontal stripe of tape that has the old recording on it. The trouble with just adjusting azimuth with standard heads is the head width is fixed, and so at a certain point it will just start picking up the other side of the tape, the other channels or the the incompletely erased segment underneath the track you want.

If the content is just speech, you may find you can identify and 'sample' the background sound (before the voices you want to preserve start up) and use this as a 'soundprint' to subtract the background noise. This is by no means a perfect or high-fidelity approach but it might give you a better result in your situation, if the background noise is fairly constant.

http://wiki.audacityteam.org/index.php?title=Noise_Removal

[MyIdenTT]

Yup, it's definitely home-recorded. I think this cassette must have lived in my mother-in-law's handbag when her children were small, and was slammed into any machine they came across when visiting! Perhaps one of them had an erase head not working (not the children)? The conflicting signals are of very similar strengths. The mixture of forward and reverse lead me down the 'tracking' route, but, to be honest, there are various layers of music making and speech going on; a mixture that Stockhausen would be proud of. I will see what can be done with the filters in Audacity, though I fear this could be somewhat laborious. Thanks for the help.
btw I recognized a Harmon Kardon machine from your picture, Andre. I still have a '50+ multi-channel receiver' in use, bought new in 1976 when the BBC were experimenting with quadraphonic broadcasting.

[Rare Bird]

btw I recognized a Harmon Kardon machine from your picture, Andre. I still have a '50+ multi-channel receiver' in use, bought new in 1976 when the BBC were experimenting with quadraphonic broadcasting.

Hi Brian
Na it's a Sansui '2000X'

[MyIdenTT]

mmm....shows how little I know...

[technobear]

The track in question will have been recorded on a 15/16 IPS four track mono cassette recorder. They were mainly available for professional use but were used for talking books at one time.

Edit: standard cassette running at 1 7/8 IPS of course - that's why it plays at double speed.

[MyIdenTT]

thanks, Chris.....any hope of rescuing it?