Digitizing tapes - Format and other recommendations, please!

[Ashville.Guru]

Hi everyone,
I'd like to digitize my dad's sizeable audio tape collection. I have a T.C. Desktop Konnekt 6 firewire interface, and am planning to use Audacity. I've gone through Dan Lavry's whitepapers on sampling, and get the gist (though not all of it!).

I'd like to archive in good quality, and then convert to mp3s for iPod listening. For archiving, which would be the ideal lossless compression format? Audacity gives a lot of options, such as Ogg Vorbis, Flac, etc.

I understand that a sampling rate more than 96Khz isn't a good idea, but what about bitrates? Audacity lets you choose between 16, 24 and 32-bit floating point.

Any other comments/suggestions/caveats are welcome, even if not directly related to my questions!
Thanks in advance.
Cheers,

Guru

 

[bouldersoundguy]

Digitize and store in wave.

Word length (bit depth) relates to noise floor, more bits equals lower noise. Given the high noise floor of tape (assuming you mean cassette) 16bit would be adequate and 32bit float would be gross overkill. Using 24bit you would have plenty of headroom.

It's a similar issue for sample rate. The HF extension of cassette is somewhat limited so it would be gross overkill to digitize them at 96kHz. All you'd be doing is capturing inaudible noise. 44.1kHz has more than enough HF to capture anything you want from a cassette, and 48kHz gives you a little extra "just in case". Some will say to digitize in the sample rate (or integer multiple) of your final product, but I think sample rate conversion is good enough that 48kHz is the best choice even if the files will be delivered at 44.1kHz.

And you have to consider file size. A 96kHz file will be twice the size of a 48kHz file of the same word length. Increasing word length also increases file size. Larger files not only take up more disk space but take more time to load, move, copy etc.

Expect to have to process the files after digitizing. Leave lots of headroom during the capture process. You can do all sorts of things to prepare the files for their final delivery format so don't try to capture them in a ready-to-listen state. Just get them in cleanly with plenty of headroom.

VERY IMPORTANT: For optimum audio quality you will need to adjust the head alignment for every tape. Then you will need to use the same noise reduction settings the tape was recorded with.

 

[ecc83]

You will also need a bucket full of ISOPROP alcohol and a mountain of cotton buds because you MUST clean the heads, guides etc after every tape run. You should also demagg before you start and at the start of every day. However, if the tape machine has a combined record/play head simply put it into then out of record mode, the decaying bias will do a fair de-magg job. (were you UK I would gladly loan you my trusty Ferrograph defluxer!).

Yes, leave lots of headroom and yes, 16bits is good enough*, even the best tape was barely 11 "bits"! I would use 44.1kHz because although there might in theory be a smidge more HF at 48 k I reckon most systems leave the same filter in!
Now, since you have the economy of 16/44.1 use linear .wav as storage as suggested, can't see there ever being a time when you cannot decode .wav?

*But 24 bits for vinyl so that you can run at -22dBFS-ish leaving a vast headroom for those pesky spikes!
Lastly, when you have a goodly batch "in the can" download the 30 day trial of Sony Soundforge. Super editor and great noise supression apps.

Dave.

 

[Tim Gillett]

Yes, 44/16 is more than adequate. Most people dont understand just how good it is. You do well to read Dan Lavry! The weak link will be the cassette recordings.

A good quality stereo cassette deck, in top condition, is the most important hardware item, to make the best of the recordings and it's where you should place your attention. (Your TC interface is excellent for the purpose.)

As Bouldersoundguy said, custom adjusting azimuth for each tape, even each tape side, is very important, and sadly is the step most people skip because of lack of knowledge and/or because it's a bother to do.

Be careful about decoding Dolby tapes. They will only decode well if everything is lined up and working to perfection. You only need one misalignment and it can sound terrible. Sometimes better to just leave it undecoded. You'll have to make your own call on that.

Good luck!
Tim

 

[Ashville.Guru]

Thanks, bouldersoundguy, ecc83 and Tim Gillett, for the replies!

Wow, this is far more useful and encouraging than I expected, actually. 44.1/16 solves a lot of issues.
Another quick question - can someone elaborate on head azimuth alignment and stuff? I hadn't heard of it before. If you could point me in the direction of some resources, that would be appreciated!
Cheers,

Guru

 

[ecc83]

Thanks, bouldersoundguy, ecc83 and Tim Gillett, for the replies!

Wow, this is far more useful and encouraging than I expected, actually. 44.1/16 solves a lot of issues.
Another quick question - can someone elaborate on head azimuth alignment and stuff? I hadn't heard of it before. If you could point me in the direction of some resources, that would be appreciated!
Cheers,

Guru

Whilst I completely understand the comments regarding the adjustment of azimuth for tapes made on another cassette machine this is something the non-tech should approach with great caution.
The VERY first step is to make a test tape of the machins as is. Since I am guessing you don't have a 'scope (!) the best bet is to record a whole C90, both sides with white noise (most DAWS will generate it, certainly Audacity) at around -20VU, Dolby off (and keep it off). This is your benchmark/default tape, keep it safe.

Next you will need a tool to adjust the head. Most often this is a very wee crosshead screw, best bet is a jewellers screwdiver set. Note, the screw is likely to be locked by a compound of some sort, use a very fine flat blade driver to clean out the head...BUT you must get the tools demaganetized before you touch the heads!!! If you magg up the head you will kill the very HF you are trying to recover by adjusting azimuth!
If you cannot beg or borrow a de-fluxer, a local watch repairer might demagg them for you for a small fee. Then don't bang them about!

Been a LONG time since I don'ed it but you should not need to move the screw more than 1/2 a turn either way to recover HF. Go too far and you will lose the screw and possible a tiny spring.


Dave.

 

[PRHunt]

The VERY first step is to make a test tape of the machins as is. Since I am guessing you don't have a 'scope (!) the best bet is to record a whole C90, both sides with white noise (most DAWS will generate it, certainly Audacity) at around -20VU, Dolby off (and keep it off). This is your benchmark/default tape, keep it safe.

Is this in order to restore the machine to it's default azimuth?

If one did have a scope, what would you be looking for?

Paul

 

[ecc83]

Is this in order to restore the machine to it's default azimuth?

If one did have a scope, what would you be looking for?

Paul

Well, there are(were?!) various ways to case the azi-moggy but one very accurate way was using a full track width recorded tape at 10kHz and initially sum the track outputs.

The head is then adjusted for maximum output (actually an AC meter is better for this than a scope). Next the two channels are connected to a dual beam scope and VERY slight adjustments made to get the two waveforms bang in phase. If you only had a single beam scope you set it up for Lissajou figures.

Ah yes! I miss those days of hum and noise......Like a bum tooth!

Dave.

 

[bouldersoundguy]

Whilst I completely understand the comments regarding the adjustment of azimuth for tapes made on another cassette machine this is something the non-tech should approach with great caution.

My decks are now playback only so it doesn't bother me much that they're not correct in an absolute sense. Only the playback alignment matters to me. But both my current decks are upper range Sony dual decks with unaltered azimuth on the record side.

 

[Tim Gillett]

Whilst I completely understand the comments regarding the adjustment of azimuth for tapes made on another cassette machine this is something the non-tech should approach with great caution.
.

I have to agree with this. Even though adjusting azimuth is the best way to do it, and I do it every time myself in my work, it's not without its pitfalls for the inexperienced.

For a start, adjusting azimuth on an old machine with a worn head can cause damage to tapes, as well as actually make them sound worse in some cases.

Also if you adjust the wrong head screw, or adjust the right one too far, you end up with big problems with the head. If you're not confident about this azimuth area just leave it and play the tapes as best you can.

The simplest way to go these days may be a small USB interface cassette player. At least you can be sure the player is brand new.

Cheers Tim

 

[Ashville.Guru]

Thanks, everyone. The tape head thing appears something I shouldn't be experimenting with (though it does look very interesting, especially dig the white noise idea).

Another related question - are headphone outs in any way less ideal than dedicated line outs - more noisy, perhaps? I understand that there is a gain/level difference, but not sure why that would be important. Any inputs would be appreciated!

 

[ecc83]

Thanks, everyone. The tape head thing appears something I shouldn't be experimenting with (though it does look very interesting, especially dig the white noise idea).

Another related question - are headphone outs in any way less ideal than dedicated line outs - more noisy, perhaps? I understand that there is a gain/level difference, but not sure why that would be important. Any inputs would be appreciated!

You are correct, headphone outputs are not ideal but, depending on the quality of the desgn, probably only marginally so compare to line (RCA) outputs and as ever with audio, there are "it depends"!!

They should certainly NOT be noisy. The sensitive nature of headphones demands very low noise but some older amps as are likely to be found in cassettte machines could be poor. The biggest problem is likely to be distortion. They are after all mini-power amps and can produce more distortion than low power line amps the most isidious of which is crossover distortion. This will be worse if the phone amp is not loaded, the 10k+ of most line inputs is effectively no load at all and if HP outs must be used, load each channel with a 100 Ohm resistor. Crossover disrortion is not always immediately obvious but causes a "roughness" to low level sounds not unlike a speaker with a "rubbing" cone. Solo piano shows it up better than most other signals.

All that said (phew!) headphones outs CAN be very useful to get a feed down a very long wire. Their low Z nature (especially if loaded) and almost total stability means 100mtrs+ of mic cable will cause them no problems at all.

Dave.

 

[Ashville.Guru]

Thanks, Dave. So ideally, it's a question of proper gain-staging, correct? If I were to ensure a very low volume out from headphones, and increase input gain on my T.C. Desktop Konnekt 6, would distortion effects be reduced/eliminated?

More generally, how do I find the optimum setting of volume+input gain?

 

[ecc83]

Thanks, Dave. So ideally, it's a question of proper gain-staging, correct? If I were to ensure a very low volume out from headphones, and increase input gain on my T.C. Desktop Konnekt 6, would distortion effects be reduced/eliminated?

More generally, how do I find the optimum setting of volume+input gain?

No! You do not want a low level from the cans output! Crossover distortion is most evident at low signal levels, milliwatts in power amp terms, so you want to keep the signal "healthy", if it sounds ok on the headphones, that level will do the TC nicely.

I have checked the AI's specc' and it can accept up to +18dBu, that's over 6volts rms and you will not likely get that from anything but a seriously pro tape deck, especially the H/P out! Remember, this is tape. Noisy, distorted tape. If we allow only 1%thd at Dolby level (Hah! would be nice!) that works out at 0.1% at neg 20VU and the distortion thru' the TC box 20 times would not be that bad!
Even at 16bits the Dynamic Range is going to be close to 90dB. Telefunken, Studer et al would have decimated a small country for 80dB!

But why the cans outs? Why not the RCA connectors?

Dave.

 

[Ashville.Guru]

I was unaware of the relationship between bit depth and dynamic range (as of many other things!)

But why the cans outs? Why not the RCA connectors?

Because my Sony portable player doesn't have RCA outs... . My Dad's ancient Akai deck has developed speed issues - I'm positive everything plays slower on it, so the Sony is the only device I currently have. If the difference in quality is significant, I might invest in a dedicated deck for recording. Headphone outs seem quite ok to my ears, though, so I was interested in knowing how bad an idea it is to record through them.

Caveats/admonishments/suggestions/comments on this issue are welcome!

- Guru

 

[ecc83]

I was unaware of the relationship between bit depth and dynamic range (as of many other things!)


Because my Sony portable player doesn't have RCA outs... . My Dad's ancient Akai deck has developed speed issues - I'm positive everything plays slower on it, so the Sony is the only device I currently have. If the difference in quality is significant, I might invest in a dedicated deck for recording. Headphone outs seem quite ok to my ears, though, so I was interested in knowing how bad an idea it is to record through them.

Caveats/admonishments/suggestions/comments on this issue are welcome!

- Guru

Ah! I see, well you will have to go with the Sony and headphone output. I take it the Sony is a "Walkman", playback only device? As for a new machine (buying a tape recorder second hand is fraught IMHO) this is one of the few I have seen..

W-890R - TEAC - TWIN CASSETTE DECK, AUTO REVERSE | CPC
But I admit I have not done a systematic search.

Re the slow Akai. You could record A 440 on it, almost any DAW will generate that, and then play it back on the Sony and into a guitar tuner app'. That will tell you if the thing is definately out of whack. Next open up the Akai and see if there is a speed adjust pot. Usually a wee hole in the back of the capstan motor or there might be a dedicated motor control PCB. To set the speed tho' you will need a cassette recorded with A 440 on it from a known good deck.

D Range and bit depth (or more properly, word length!). Assuming a perfect system every "bit" gives you about 6dB. Therefore a 16bit system has a DR of 96dB and 24 bits a massive 144dB. We cannot in practice get anywhwere close to the latter figure because analogue ins and outs cannot be made quiet enough. A really good "consumer" grade AI will get to -100dBFS (my NI KA6 e.g.) and top bllx RME/Apogee stuff 110dBFS or a bit better.
In practice there is nothing THAT quiet to feed them and if there were, our rooms are not quiet enough anyway! (An SM57 at max gain into the Ka6, dead 2-3am gets me -75dBFS and that is acoustic, not electronic noise. I live in a leafy suburb, V quiet, googe NN5 5Pletter.)

Dave.

 

[bouldersoundguy]

Because my Sony portable player doesn't have RCA outs... . My Dad's ancient Akai deck has developed speed issues - I'm positive everything plays slower on it, so the Sony is the only device I currently have.

I wouldn't want to commit myself to a portable for anything worth transferring. My two decks cost $35 and $45 used. They are in great shape and it's only the obsolescence that makes them so cheap. Being dual decks I can freely mess with the alignment on the play side while the record side stays stock (though I don't really record cassettes anymore).

 

[Ashville.Guru]

Thanks once again, everyone.

Dave, it's not a walkman, it's more of a boombox kind of thing. But seeing the suggestions here, I'm not too comfortable with the headphone outs anymore. Tape decks are hard to come by in my part of the world, plus the only advantage of a deck vs. a portable would be the azimuth adjustability, correct? I'm not willing to take that risk yet, being a newbie to this stuff.

Meanwhile, I've been doing some scouting, and might be acquiring another portable - the LG LPC-D1000, with dedicated aux outs. Most of the tapes in my collection were recorded from one such portable to the deck (in the late '70s), so I don't see the advantage of using a deck now for digitizing them. However, I'm open to suggestions - if there is a vastly significant advantage, I might consider importing one.

Thanks for all the advice and inputs so far!

- Guru