Audio Restoration: Old Buddhist Teachings

[Rubber Lizard]

The tapes are from as long ago as 1978. There are boxes and boxes of them. The recordings are of Tibetan Buddhist teachings, many of these old Tibetan Lamas are long reincarnated. The recordings will cover the entire range, from unintelligible to pretty good. Someone is donating money for equipment and software to facilitate getting these teachings into a digital format. I can buy what I need, within reason. That's where I would love to hear from any who have experience in the restoration of such audio. I have some specific questions as well.

 

Step one will be digitization. I can buy a new cassette player. Probbly can't afford a Nakamichi Dragon (dang). Would a current Denon or Tascam be "gentle" on the poor old tapes, or is there a better approach? And is there some kind of built-in noise reduction that would help clarify the sound at this stage?

 

I will probably recommend a 2-channel M-Audio sound card and their $99 monitors. Remember, these are nasty old cassette tapes, and I'll be training a few volunteers to do the hours and hours of work (though I'll be involved as well).

 

Then there will be digital processing. I'm thinking Sound Forge and some noise reduction plugins. Sound soap?

 

Do you think that autotune plugins like Roland's V-Vocal might help with wow and flutter?

 

Okay, that's enough to see if anyone is interested in helping out.

 

Until next time, Vajra Satva Hung, Store Beer in a Cool, Dark, place. Over and Out.

 

-mark lacoste

Rubber Lizard Studio

Ashland, Oregon

[russrags]

Hey Rubber Lizard,

 

 

Sounds fun. I do a lot of audio archiving of old tapes, but not that many cassettes, but can offer a few suggestions. Surly your familiar with the saying "less .. is more" it certainly applies here, so I'd forget the autotune notion and if it were me, in order to prevent disaster, I'd leave the volunteers out as well.

 

A good cassette deck straight into Sound Forge is enough as your biggest concern are the tapes themselves, although 1978 isn't that old, it's still long enough for damage if not stored properly.

 

Nakamichi has their own EQ curve and unless the tapes were recorded on a Nakamichi, you might not get as good of results as your expecting. Denon or Tascam will probably be fine. Pick a deck with a very gentle transport and even then, you'll want to avoid using the fast forward and rewind buttons. Cleanliness is going to be critical, regular alcohol you probably already have in your medicine cabinet is not what you want to use on your cassette deck, get head cleaner or 99% denatured alcohol .. plenty of Q tips (No Plastic sticks as they will melt) and get a head demagnetizer and learn how to use it properly.

 

Important: If the tapes are recorded on High Bias tape ... put the deck in the high bias position. If the tapes are recorded on Normal bias tape ... put the deck in the normal bias position. It all has to do with a built in EQ curves for each position.

Look for any information you might find written down about dolby or not, but yes Sound Forge has a noise reduction plug in which will be very useful and is non destructive.

 

Important: POP all the tabs off of the tapes before you begin to avoid accidental erasing.

 

Hopefully these tapes have been kept in their boxes, but loose tapes will cause the tape to be loose as well. Carefully tighten the tape with a pencil to start with ... if the tape is in the middle, play it on the B-side until it stops, then turn it over to the A side to begin archiving.

 

Start with the least important tapes until you get a handle on it. Your going to want to keep a real close eye on the cassette decks heads after each pass. If you get a tape with a tremendous build up of gunk on the heads let me know.

 

Sound Forge: Upon making a recording look at the sound waves and determine if Left and Right look even or close enough to it. If not it may be alignment issues with the deck the tape was originally recorded on .. or, you might determine MONO is better for voice only. Record and label a batch of tapes, but be sure to BACK UP at this stage before you start manipulating the sound with noise reduction and EQ.

 

I'll leave you with this much for now as I need to get my day going here and I let my coffee get cold.

 

Good luck,

 

Russ

Nashville

[MikeRivers]

I can't suggest a specific model cassette deck, but look with your own eyes before you buy. Remove the front cover of the cassette well and see if you can access the head alignment screw without further disassembly. It would be a good ideal to learn about azimuth alignment (and teach your interns) and touch it up before playing each tape. It'll probably change through the tape, but at least you can get it reasonably matched to the condition of whatever recorded it.

 

Do you know if the tapes were recorded with Dolby noise reduction? Can you tell by listening? (it's really difficult on speech, so don't be embarassed to say you can't) Given the content, it would probably be best if they had been recorded with Dolby B, and played back without it. That way you'd have the advantage of a little compression and high end boost, which would probably be an advantage.

[groovetube66]

I did something similar to this for work recently for a series of workshops about habitat restoration in Washington state. The tapes weren't old but motor noise presented a big problem. They used the built-in mic and the recorder had AGC, with the recorder set directly on a table top. This took lots of playing with EQ, compression etc. in SoundForge. If you are lucky they used a remote mic.

[Lee Knight]

I highly recommend the Waves DeNoise plugin (I think it's called) from the Sound Restoration Bundle. Not the cheapest, but very good.

[boosh]

What would be utterly cool is find the new incarnations of these Lamas and let them sing second voice to the vocals of their former incarnations,....

[philbo]

I've done some work in this vein, though it was reel-to-reel rather than cassettes.

 

A couple of ideas that helped for me:

1. If you can get a 4 track unit, you can record tracks 1&3 as one stereo pair, and 2&4 as another stereo pair, all in one pass. Then simply reverse the recorded data from tracks 2&4. This will cut your time in half.

 

2. If it's just speech, and the 4-track unit has double tape speed, use it. By recording the tracks in at 88.2KHz, then retag the files to play back at 44.1KHz, you cut your time in half again.

 

3. Use the de-noiser of your choice. I've always gotten the best results using the hiss reduction and noise reduction in CoolPro/Audition, though. Use with trepidation and great caution - - and keep the (untreated) source files for a month or two just in case.

 

I don't know of anything that does a creditable job on fixing wow and flutter. It'd sure be nice if there was something.

 

Depending on the recording quality, it might be of some benefit to do an S-curve compression - - pushing down the background noise and ambient reverb while simultaneously bringing up the sound of interest.

 

Hope this helps you.

[UstadKhanAli]

I have no suggestions for equipment. I don't know where Kylen is, but he does cassette restoration, so he might be good to ask.

 

Cool project!!!!!

[Rubber Lizard]

Wow, this is great, guys! There's a lot of information here that will help.

 

Russrags, There are boxes and boxes of tapes. Thousands of hours! If I ever hope to play music again I must get these folks to do much of the work themselves. I will ask my Sweetwater guy to suggest a "gentle transport". We will buy a couple of cassette players so that I can do some of the work here at Rubber Lizard Studios while they work in the temple. Rubber Lizard started as a cassette multitrack, so I will dig out the old de-magnitizer and buy some head cleaner. I'll be printing this post of yours to use as a guideline.

 

Mike Rivers, Groovetube66, Lee Knight, and Ustadkhanali, thanks. Your advice will help generate some good karma here. And Booshie; We probably do know some of the recent incarnations... Your idea is weird in the coolest of ways...

 

Philbo, those are some genius ideas. Beyond us for now, but who knows? We may find that time is our biggest problem! And I must investigate this concept of S-curve compression for intelligibility. Yes, Cool Edit Pro did have neat noise reduction. Maybe we can use that. And I may try to find Kylen.

 

Thanks, all! And please chime back in if you get any inspired ideas.

 

-mark lacoste

[spokenward]

I would recommend that you start with a high quality 24bit capture with converters that you personally like and trust and leave "noise reduction" and concerns about wow, flutter, or speed correction aside. You can always revisit those issues, and the technology that you will have at hand will only improve. For your initial review, I like the CoolEdit family (after they released the noise reduction as a plugin) or the Audition releases. The advantage of the Audition product is that it has the ability to ruthlessly slice a window in the frequency domain. This can be helpful for people who are trying to transcribe. The disadvantage of some versions of CEP or Audition are dreadful load times while they open your 24bit file as 32.

 

One thing that has been my experience - when casual listeners describe the tapes of an event as "noisy" they might think that it is because it is on a funky little cassette. More often, it is because the whole thing was mis-miked and/or mis-recorded. Until you can sort that out, remind them that "all is suffering".

[UstadKhanAli]

If using a decent cassette deck, wow and flutter should be negligible.

 

And yes, use the best converters you can (24-bit capture as Spokenward said).

[Lee Knight]

 

I've done some work in this vein, though it was reel-to-reel rather than cassettes. A couple of ideas that helped for me: 1. If you can get a 4 track unit, you can record tracks 1&3 as one stereo pair, and 2&4 as another stereo pair, all in one pass. Then simply reverse the recorded data from tracks 2&4. This will cut your time in half. 2. If it's just speech, and the 4-track unit has double tape speed, use it. By recording the tracks in at 88.2KHz, then retag the files to play back at 44.1KHz, you cut your time in half again.

 

 

Those ideas are so elegantly simple. That's really cool Philbo. I guess you'd want something that doesn't use DBX like the old Tascam 244. Good stuff.

[MorePaul]

Rubber;

 

you may want to give

http://www.historicalvoices.org

 

a shout as digital archiving of spoken word is their line of interest -- they may be able to help you out with more very specific stuff

[joel Oporto]

I don't know if anybody mentioned this, but before I record an old cassette, I usually forward and rewind once (to cover both directions) before playing it back. This is to move the tape prior to actually playing them. Also per batch of tapes you want to record, store in a small box without the cassette cases, and put silica gel dessicant for a whole day or more.

[Rubber Lizard]

Okay, we continue to make progress here in the Colestin valley. My son has completed a CD for testing. All processes were done using Adobe Audition and the built-in-to-XP CD burner.

 

Oh... Some more details for anyone still listening; Preferred delivery format is MP3. See, these Buddhists are VERBOSE AS HELL, and it is best if many hours of instruction can reside on a single disk... "Here is a teaching on 'quieting the mind' in which I will speak at you for 4 hours straight...". Oh, man... that's probably bad karma, making fun of... oh well.

 

See, the teachings are recorded live. There is a Rinpoche (high level teacher dude) and a translator. The Rinpoche speaks for up to 15 minutes, the the translator translates, then the Rimp speaks again, etc.. This lasts for about two or three hours before they take a break (you are sitting cross-legged on the floor for this whole time). Then you have tea, or maybe lunch, and go at it again afterward. This can go on for as long as a month. We will digitize the original recordings, then put split points, queues, markers, whatever, between the Rimp and translator segments. Then we create three versions: Tibetan only, English only, and unedited, to be distributed as MP3 files on CD.

 

BTW, Thanks Joel. I am worried about the old tapes. You recommend ff and rw once... that sounds wise. I hope to also learn about how to find "gentle" cassette transports. Okay, I'm off to bed for now. Thanks, folks!

[UstadKhanAli]

"Rinpoche" is also a Tibetan term for reincarnated lama, and means something to the effect of "Precious Teacher" in Tibetan.

 

I may not be able to beat anyone in Trivial Pursuit, but I do okay at certain subjects... No, seriously, I just figgered you might want to know...