Topic: A few newbie questions re: cassette transfers

[dougt502]

I am new to this world but am interested in transferring several analog (and dat) recordings into my hard drive before I lose it all to time.


Soundcard - internal or external USB?  (budget an issue, can spend around $100)
Azimuth - Do I adjust manually on my Nak Cr4a or playback on the original recording decks (d5 and PMD430)?
Sample rate and bit depth - it seems everyone has an opinion without a consensus   
Recording software - budget still an issue but don't need to cheap out


Any help or suggestions would be greatly appreciated!  Thanks!

[Brian Skalinder]

Analog-to-digital conversion outside the PC is best to avoid noise and artifacts, so look for an external soundcard.
Adjust the azimuth, regardless of which playback deck you use, as the azimuth on the original deck(s) may have changed over time.
16/44 should be sufficient to capture the frequency response and dynamic range of a cassette.
Try Audacity.  It's free, reasonably powerful, if slightly clunky to use at times.

[greenone]

Analog-to-digital conversion outside the PC is best to avoid noise and artifacts, so look for an external soundcard.

Or, alternately, use your DAT deck as your analog-to-digital, and output to a separate device with digital in. I run straight to my field recorder - if you have it, no real need for a separate device.

[morst]

16/44 should be sufficient to capture the frequency response and dynamic range of a cassette.

I disagree. I transferred a lot of cassettes at 16/441 and now I want to go back and redo them. If for no other reason, 24-bit sampling allows much more flexibility in regards to setting input levels. But also, I don't master ANYTHING at 16/441 anymore, cause I don't want to play back at such a low quality. I go 24/48 now, and I'm figuring I'll probably go higher whenever it's feasible.

[Brian Skalinder]

I transferred a lot of cassettes at 16/441 and now I want to go back and redo them.

What do you expect to gain by re-doing the transfers?

If for no other reason, 24-bit sampling allows much more flexibility in regards to setting input levels.

Leeway in settling levels is about the only reason I can think of to transfer at 24-bit, though given cassettes have max dynamic range of ~60 dB you have some leeway even with 16-bit transfers.

[dougt502]

Thanks!  I'll take all your suggestions under consideration and let you know how things turn out!

[morst]

I transferred a lot of cassettes at 16/441 and now I want to go back and redo them.

What do you expect to gain by re-doing the transfers?

FAR BETTER SOUND!!

[Brian Skalinder]

FAR BETTER SOUND!!

If 16-bit / 44.1 kHz is more than sufficient to handle the limited dynamic range and frequency response of the original cassette, why would re-transfers at 24-bit / 48 kHz (or some other bit depth / sample rate higher than 16  /44) sound "far better"?

[runonce]

FAR BETTER SOUND!!

If 16-bit / 44.1 kHz is more than sufficient to handle the limited dynamic range and frequency response of the original cassette, why would re-transfers at 24-bit / 48 kHz (or some other bit depth / sample rate higher than 16  /44) sound "far better"?

Agree = Once the net is big enough to land the fish - a bigger net will not make the fish anymore "caught"...

There probably are some other compelling reasons to record at 24 bit - but better sound is not one of them.

I prefer native, undickered 16 bit samples for this kind of material...YMMV.

[andromedanwarmachine]

hmm...

Can I just say that it's perhaps worth remembering dougt502, to run the tapes without NR...

You can't put back what you've lost after the transfer, where as modern filtering goes along way- I've yet to see a cassette deck regardless of budget that made a good fist of the NR encode decode process...

That's how I'd handle it anyway- and I did ALOT of work on cassettes.

JimP

[JEMS]

hmm...

Can I just say that it's perhaps worth remembering dougt502, to run the tapes without NR...

You can't put back what you've lost after the transfer, where as modern filtering goes along way- I've yet to see a cassette deck regardless of budget that made a good fist of the NR encode decode process...

That's how I'd handle it anyway- and I did ALOT of work on cassettes.

JimP

Wow, surprised to hear this. In fact, I feel quite the opposite based on experience--that transfer is the only moment to get it right. Proper playback azimuth adjustment paired with either outboard Dolby units or a deck that does it well (like my Nakamichi CR-7A) can really make Dolby shine. That being said, would love to hear how you're filtering for Dolby B and C.

[andromedanwarmachine]

everyone keeps mentioning azimuth here- it's really terrifying!

Proper quality gear used properly should never lead to a need for adjustment. That's what I'd like to think anyway, but in fact my experience with even half decent gear is that the media misaligns between decks...

I did 5 years with Sony WM D6-Cs and despite my best efforts they were NR incompatible with the Technics, Marantz and Aiwa decks I had in the house. In the end I just ran without NR and got a better top end result- without "pumping".

I know that "C" was an encode/decode process which straight notch filtering would not replicate but there are some very effective spectral-footprint tools available in software that will stand a better chance without the NR engaged for initial digital transfer.

Here's a question- which I could go and answer myself at home later; if you record something with NR on and playback with it off, does the material suffer beyond what would have happened if you didn't use it all...

[morst]

FAR BETTER SOUND!!

If 16-bit / 44.1 kHz is more than sufficient to handle the limited dynamic range and frequency response of the original cassette, why would re-transfers at 24-bit / 48 kHz (or some other bit depth / sample rate higher than 16  /44) sound "far better"?

IF it were sufficient, that might be true. I do all my processing at 24 bits or higher, so it doesn't make sense to rip at 16, then convert to 24 when I can just digitize at 24. I'm not gonna go back to 16 bits, cause I no longer like the sound of it. Remember when CD's first came out and the vinyl purists pooh-poohed them? Now I have a decent enough stereo that I can hear what they meant. And it's not just the early 16-bit digital material, though some of those "mastering techniques" were very crude compared with modern ones. 16 bit just sounds brittle and harsh to me. Not nearly as gutted as MP3, but still lacking that thickness. The more I listen to 24, the less I want to hear 16 and below.

There probably are some other compelling reasons to record at 24 bit - but better sound is not one of them.

Then we shall agree to disagree.
...

meanwhile:

everyone keeps mentioning azimuth here- it's really terrifying!

I know, "luckily" all my cassettes are due for re-transfer, so I'll be able to start fresh and align azimuth for each side of each tape when i do. Big project. "Terrifying" is not entirely inaccurate.

Proper quality gear used properly should never lead to a need for adjustment. That's what I'd like to think anyway, but in fact my experience with even half decent gear is that the media misaligns between decks...

Even the azimuth of a single, high-quality well-cared-for deck could change over time. I have masters made on many different decks, and certainly won't be using any of the mastering decks for the transfers anyhow, so I'll be needing to optimize alignment each time I flip or change a tape. If you are transferring tapes, and have the opportunity to adjust azimuth each time, do that and you will not regret it.

Here's a question- which I could go and answer myself at home later; if you record something with NR on and playback with it off, does the material suffer beyond what would have happened if you didn't use it all...

Hell yeah. The spectrum is all screwed up if you don't decode, and decoding depends so much on the same circuit being used, that NR is not worth the hassle. At least that was the "conventional wisdom" passed down to me by the folks who taught me the finger points of analog cassette recording.

[Brian Skalinder]

IF it were sufficient, that might be true. I do all my processing at 24 bits or higher, so it doesn't make sense to rip at 16, then convert to 24 when I can just digitize at 24.

Yes, 16-bit is sufficient to capture the dynamic range of an analog cassette.

I think you're conflating bit-depth of the analog to digital conversion process with post-processing precision.  For the purposes of capturing an analog cassette, transferring at 24-bit will not provide a superior result over 16-bit.  The higher dynamic range possible with 24-bit simply doesn't exist on an analog cassette, so there's nothing "extra" (i.e. higher dynamic range) to be captured by transferring at 24-bit.

Processing at 24-bits (or 32 bfp or higher) v. 16-bits is a different issue altogether.  Most audio editors I know use 32 bfp precision and maintain that precision throughout the editing process (in some cases, like the old CEP, only if configured to do so).  Using this higher precision delivers post-processing benefits in the form of lower quanitzation noise.  But it's a separate issue from the transfer bit-depth.

One benefits from 32 bfp processing precision regardless of whether the original transfer bit-depth is 16- or 24-bit.  And as long as the original transfer bit-depth is sufficient to capture the dynamic range of the source (in this case cassette), there's nothing gained by using a larger than necessary bit-depth during transfer.

[morst]

oh, I guess I was thinking that since Plangent Processes requires 24/96 or 24/192 at minimum that if you're not digitizing at high bits, then you should be sure to save the masters for a redo using future technology. It's way too expensive to do that now.