You are here

The Twelve Steps for Recording Tapes into your computer

All about the tools and techniques involved in capturing sound, in the studio or on location.

The Twelve Steps for Recording Tapes into your computer

Postby Tomás Mulcahy » Sun Sep 06, 2020 1:10 pm

This topic comes up often on the Sound On Sound forum, and I am grateful to Hugh Robjohns and James Perrett there for the advice and discussion over the years. Along with my own practice, I’ve whittled the process down to 12 steps.

http://madtheory.com/site/the-twelve-st ... -computer/

I hope this is helpful for some of you!
User avatar
Tomás Mulcahy
Frequent Poster
Posts: 1799
Joined: Wed Apr 25, 2001 12:00 am
Location: Cork, Ireland.

Re: The Twelve Steps for Recording Tapes into your computer

Postby James Perrett » Sun Sep 06, 2020 2:40 pm

That's a very useful link and I've just downloaded the A1 Stereo Tool you mention. It looks a bit like Waves S1 so it will be interesting to give it a try.

Just a few suggestions - it might be good to go into a little more detail in the cleaning section because it seems to jump into the middle of the process without mentioning the need for swabs and IPA (rather than cleaning tapes) and where to apply them.

It also might be worth mentioning the need to check pressure pads - for me this is the most common reason for transferring tapes to a new shell.

I think the advice on speed correction may also change if RX8's wow and flutter module lives up to expectations.

I'd also recommend Stillwell's Exciter JS Plugin for recreating high frequencies which is available in the Reaplugs package for those who don't use Reaper. If you can't afford RX then Reaplugs also includes ReaFIR which can be used to reduce noise.

I like the Yamaha cassette deck at the top of the article - it kind of symbolises the transition from top loading to front loading cassette decks that was happening at the time it came out.
User avatar
James Perrett
Moderator
Posts: 9679
Joined: Mon Sep 10, 2001 12:00 am
Location: The wilds of Hampshire
JRP Music - Audio Mastering and Restoration. JRP Music Facebook Page

Re: The Twelve Steps for Recording Tapes into your computer

Postby Hugh Robjohns » Sun Sep 06, 2020 4:46 pm

Tomás Mulcahy wrote: I’ve whittled the process down to 12 steps.

Looks good, although I've only skimmed it so far.

One thing I noticed, though was this:

If it was a mono recording, do still record L and R, and sum them. You get less tape hiss (works the same way as a balanced line) and slightly more bandwidth

I'm not convinced you'd get more bandwidth (and if the azimuth is off you'd definitely get less!). But the BS-bell went off on the balanced line comment.

You do get less tape noise, of course, but I fear the balanced line reference is likely to confuse more than it helps.

The simple reason is that the tape noise is random and thus incoherent between the two channels, so summing will raise the level by only about 3dB. In contrast, the wanted signal is the same on both tracks and thus coherent over the two channels, and so summing increases the level by 6dB.

Hence there's an overall 3dB improvement (6-3dB) in the signal-noise ratio.
User avatar
Hugh Robjohns
Moderator
Posts: 28330
Joined: Fri Jul 25, 2003 12:00 am
Location: Worcestershire, UK
Technical Editor, Sound On Sound

Re: The Twelve Steps for Recording Tapes into your computer

Postby forumuser840717 » Sun Sep 06, 2020 5:01 pm

Useful article, Thomas.

A few other things to add to James's comments.

To reduce pitch wobbles, remove the reels from the shell and put them into something really good like a Maxell XL-IIS shell.
I too find that it's a good idea to check pressure pads as they're inclined to be an issue on older cassettes.
If there 's a mechanical problem with the shell, I find that ceramic shells such as the Sony Metal Master or TDK MA-XG are a usefull step up in stability with most cassette tapes. The older TDK MA-R die cast shells with plastic sides can also work well. The other ones that can be useful with tapes that are a bit sticky are the BASF shells with their Cassette Security Mechanism (identified by the"SM" logo on the shell).

Clean the heads of the deck.
Don't forget to clean the rest of the tape path too - pinch rollers and capstan. The metal capstan can be cleaned with alohol but only use clean - preferably de-ionised or at the very least boiled (to remove chemical residues left on drying) - water on the pinch roller(s)

Don’t use alcohol on the rubber parts of the transport it will dissolve them eventually.
It can do that but I've never had it happen to me on a quality cassette deck. The main problem I've seen with alcohol in cassette decks is that it 'cures' the rubber parts, seemingly accelerating ageing and causing them to harden and go shiny (and eventually brittleness and cracking/falling apart) so things like pinch rollers stop pinching properly and skid along causing instability in the tape path against the capstan. Apart from being really irritating on the speed stability front, it can be particularly problematic on a closed loop dual capstan deck where the tape can become unpleasantly tangled if the take up roller is slipping badly whilst the feed roller drives normally.


Adjust the head azimuth
When it comes to azimuth adjustments, the Nakamichi Dragon mentioned above has it pretty much covered with automatic azimuth adjustment which could track and adjust azimuth as the tape played* but some other decks help with this by giving either an external control knob/buttons for tweaking azimuth, or holes for easy access for trimmer tools to azimuth adjustment points without needing to open up the deck or remove the transport door/whatever. (*The only other full sized deck to do this (though not as well as the Dragon) was the even rarer Marantz SD930 but Nakamichi also fitted it to their high end in-car cassette player, the TD1200.)

As well as azimuth adjustments, it's helpful to get the speed adjusted correctly to the recording. Again, some decks offer varispeed controls for this, otherwise it's usually a matter of finding the speed adjustment trimmer on the circuit board or wherever. Even on a properly calibrated deck, the replay speed will probably need some tweaking to match the speed at which the tape was originally recorded as there was often a rather wide tolerance on manmufacturing parameters of, particularly, domestic cassette decks. Getting the replay speed right at the transfer stage is really a better idea then trying to fix it later with software tools.


Even if Dolby was used, consider leaving it off.
The main problem with Dolby (or any other encode/decode noise reduction) is that, as you describe, to work properly it requires a fairly tight tolerance in the mechanical and electronic alignment of the record and replay deck(s). It usually works ok on a reasonably well set up, good quality, deck for playing back the recordings made on the deck but is prone to decoding errors when playing things recorded on other decks. This is simply because of alignment differences.

Regardless of the age of the tape, the chances of domestic (or pro) recording decks being anywhere near the same alignment, mechanical or electronic as the replay deck are pretty slim. Even where the record and replay decks are the same, not all (most) manufacturers didn't pay sufficient attention, particularly on mass market products, to making sure that the deck was properly aligned before it left the factory and 'within manufacturing tolerances' didn't always mean within Dolby tolerances. Given that it was random enough getting alignment right within the same deck, from one deck to another of a different make or model the chances of reasonably accurate alignment between them were pretty small. Very few decks let the user adjust the levels in/out of the NR circuitry without taking off the lid and, even then, with the tolerances/coarseness of the alignment controls in many cheaper decks, it's very easy to misadjust and really screw up the NR system. Running without decoding Dolby is usually the simplest solution. Some decks have the send and return loop connections for connecting external noise reduction processors

It's worth mentioning that noise reduction types found on cassette aren't limited to Dolby, though that (Dolby-B) is by far the most common. A few decks offered dBx II or Telefunken High Com/Nakamichi (Telefunken) High-Com II and there were other companding NR systems like BNE (Burwen Noise Eliminator), JVC Super ANRS, Sanyo Super D, and Toshiba ADRES. Also, "Dolby" could refer to Dolby Types B, C or S, all of which have different profiles and alignments; Dolby A and SR were really for professional use and aren't found fitted to cassette decks but it's just about vaguely possible that some clever dicky could've used them externally to the cassette deck itself. In the past I've been asked to transfer cassette tapes (and DAT cassettes) with audio encoded with A or SR.

IF you can find one, Dolby made outboard processors for Dolby B, C and S in rackmount form (such as the triple format Model 422 which does all three in 1U but they also did single and dual format units) which make life a little easier by having balanced i/o and all the adjustments on the front panel. Other people made outboard Dolby processors, including the rather useful Nakamichi NR-200 which covered Dolby B and C in outboard form with easily accessible adjustments and front panel controls. There is another software multi-format NR decoder in development which Im told is very good but I haven't yet tried it.


Record all of the tape
Definitely do this - not least because there are occasionally random bits of interesting programme which happen when you think it's all over. But more importantly, transferring blank tape can be very useful in identifying issues with the recording and setup of the original deck, and can provide a useful 'noise fingerprint' for noise removal technologies which work that way. Different tapes have different hiss profiles and the hiss can also change with alignment if the recording deck was itself very noisy so whilst stock noise samples can be handy, being able to get a good noise sample from each unique tape is better. Transferring completely unrecorded tape, where it is available, can help too.


Digitise the tape at 44.1kHz 16 bit.
I'd suggest always digitising at 24 bit as it can help with any digital post processing and, these days, the extra storage and processing power aren't really an issue on stereo files.
I usually encode the audio channels as two mono WAV files rather than stereo WAVs as it can sometimes be desirable to process the channels individually. (Of course it depends how your DAW deals with such things.)
And transfer clean, before doing any electronic or digital processing.
forumuser840717
Regular
Posts: 148
Joined: Thu Jun 16, 2016 5:20 pm

Re: The Twelve Steps for Recording Tapes into your computer

Postby Trevor Johnson » Sun Sep 06, 2020 5:47 pm

Excellent, many thanks Tomás!

I've had a Tandberg TCD-440A from new, since 1981. Interesting history, because it came direct from Tandberg, Leeds. My younger brother Ian, (sadly passed away last year), worked for Arthur Andersen at the time, who audited Tandberg. I bought it at cost, £200, I think retail was £400. And for many years afterwards, the great engineers at Tandberg, Leeds, serviced my Revox reel-to-reel machines.

The Tandberg was, and is, a terrific machine, with a tone generator and adjustable record head azimuth. The only weak link are the long throw slide pots for input gain, but they are very easily cleaned. It still amazes me how much research, time and money, were spent on what started out as a dictaphone format!

(1960s note on dictaphones: my dad used to use Grundig reel-to reel recorders as his dictaphone. One use he made of it was to record the audio from Doctor Who TV episodes if we were out that day. I've sometimes wondered whether they could have been the audio of the 'wiped' tapes, but we'll never know).
Trevor Johnson
Regular
Posts: 325
Joined: Sat May 15, 2010 12:00 am

Re: The Twelve Steps for Recording Tapes into your computer

Postby Tomás Mulcahy » Sun Sep 06, 2020 11:47 pm

Thanks for the kind words Trevor, that Tandberg looks lovely. I am a sucker for a well engineered tape deck. SO far I've resisted the temptation to start a collection ;)

Thanks again James, I did not know about the forthcoming RX module, can't wait to try it. Staying on RX7 standard for a while here though. I’m trying to avoid TLDR, so am assuming people know how to clean the heads. Similarly, just by transferring the reel cures any shell and pressure pad problems in one go. A-1 is just an MS processor (it is great though) so I would think S-1 would still be better for widening?
Hugh Robjohns wrote:I'm not convinced you'd get more bandwidth (and if the azimuth is off you'd definitely get less!). But the BS-bell went off on the balanced line comment.
Woah, bit OTT there Hugh. While I accept it's a white lie, it's not BS. I tried to pitch the article so it's not too technical to enable people to get on with doing the job but maybe have their curiousity piqued. Finally, in fairness I used the word "slightly" which is what your figures show!

Thanks forumuser840717 glad you found it useful. I will check out that Marantz. Dissolve or cure, I think the distinction is academic in this case and again I’ve tried to pitch it (accidental pun) to be as practical a guide as possible and minimise TLDR ;). I take your point about pitch though! Generally I prefer do varispeed in the DAW where it’s easier to loop a section that has a known pitch. Similarly with the NR, UHE Satin is a worthy substitute for outboard NR boxes, it does several flavours of Dolby and dbx. As I mentioned, IME the Dolby B works really well- better than what you can do on most decks (thanks again to James for that tip). Again you are correct that I’ve glossed over the precise causes of level misalignment, but again my intention was to pique interest. Looks like that was a success :) IME you’ll find with a large archive of cassettes, storage space will be an issue and given the poor dynamic range of cassette, 16 bit is more than enough as I outlined. Interesting you use dual mono in your workflow. Pro Tools lets you treat a stereo source file as two mono tracks if you like, without changing the source file. I don’t know if other DAWs do it that readily, but it's such a useful facility I'd imagine they do.
User avatar
Tomás Mulcahy
Frequent Poster
Posts: 1799
Joined: Wed Apr 25, 2001 12:00 am
Location: Cork, Ireland.

Re: The Twelve Steps for Recording Tapes into your computer

Postby James Perrett » Mon Sep 07, 2020 1:55 am

Tomás Mulcahy wrote:IME you’ll find with a large archive of cassettes, storage space will be an issue and given the poor dynamic range of cassette, 16 bit is more than enough as I outlined.

I'm now using FLAC for quite a bit of my work which can halve your storage requirements. The only time I convert to .wav is when I want to process files in my ancient copy of Adobe Audition or for final delivery.
User avatar
James Perrett
Moderator
Posts: 9679
Joined: Mon Sep 10, 2001 12:00 am
Location: The wilds of Hampshire
JRP Music - Audio Mastering and Restoration. JRP Music Facebook Page

Re: The Twelve Steps for Recording Tapes into your computer

Postby Tim Gillett » Mon Sep 07, 2020 6:40 am

Re software, IME the most useful post tool for mono cassette recordings (transferred in dual mono) is an "Azimuth Corrector" or similarly named tool even though it only time aligns left and right tracks. Cassettes tend to drift in azimuth especially when L and R are summed to mono. (Microcassettes running at 1/2 or even 1/4 cassette speed are far worse, especially if played back in mono. Oh for a forensic JBR playback unit with (ISTR) a 13 track (!) playback head and all for the humble microcassette.

Azimuth often drifts within one cassette side and even from moment to moment. Not only that, a cassette originally made on a simple shoebox recorder but transferred on a high end dual capstan type like a Dragon can actually introduce worse drift errors because the tape tensions across the tape path are now different from when the recording was first made. I saw this pattern repeat itself in our big state digitisation project. Nearly always the azimuth meter in Cube Tec programme drifted slowly the same way (to the right ISTR) at ingest. So much for top quality machines always giving a better result in every aspect.

Speaking of the Dragon, its NAAC function has its definite limitations. First it only works on the right track so is useless on tapes with poor treble on that track. According to Richard Hess it's also very slow in responding due to the way it has to operate. As well it doesnt seem to have the accuracy to time align both tracks reliably without further digital azimuth correction. So the Dragon is rare, very expensive and no magic bullet.

The Azimuth Corrector is definitely not a substitute for physical azimuth adjustment to each cassette of course but if that physical correction is done well the corrector can often correct smaller remaining errors and the summed audio can be clear and solid without comb filtering. As always the best way to reduce tape noise is to transfer the tape's contents optimally in the first place.

As for the A1 Stereo Control I dont understand the value of it with cassettes. Cassettes are far more prone to inconsistencies and dropouts in the highs than the lows.
Added to that, subjectively bass is far more non directional than treble, regardless of playback format issues.

Apart from the modest potential reduction in non coherent noise as Hugh mentioned another reason we try to sum an original mono recording from dual mono back to mono is to allow one track to mask instantaneous tape dropouts on the other, based on the lower probability both tracks will drop out at the same time.
Tim Gillett
Frequent Poster (Level2)
Posts: 2212
Joined: Wed Jan 30, 2013 1:00 am
Location: Perth, Western Australia

Re: The Twelve Steps for Recording Tapes into your computer

Postby ken long » Mon Sep 07, 2020 8:27 am

Cool guide. I would echo James' comments about the pressure pads and also suggest adding a bit about shells and explaining different types. You could also briefly mention baking cassette tape out of its shell or re-lubricating dry tapes that don't play.

I'd also add using a phase scope for adjusting azimuth as it also gives you a visual.

I would also agree with the transfer at 24-Bit for post processing statement. Storage costs nothing nowadays.

I'd add the Tascam 122 as a good deck as it has relatively easy access to the azimuth screw and Dolby B and C as well as being a balanced output. Totally agree about leaving Dolby out. "If in doubt, leave it out!!"

Finally, I would suggest any processing be done on a copy of the file so you could always go back to the master should you discover a new, more effective method for restoring the audio.
User avatar
ken long
Frequent Poster (Level2)
Posts: 3378
Joined: Mon Jan 21, 2008 1:00 am
Location: Somers Town
I'm All Ears.

Re: The Twelve Steps for Recording Tapes into your computer

Postby Tomás Mulcahy » Mon Sep 07, 2020 11:26 am

Perhaps I should do a follow up post. Target audience being anyone who is conversant with oscilloscopes and dynamic range. But would that demographic be able to figure out the details themselves anyway? Can anyone recommend links to further reading for more detail? The Rane pro audio reference is behind a paywall now. Would probably make more sense to refer the curious elsewhere.

The very best of the best cassette can get to 70dB dynamic range. Using the rule of thumb 6dB/ bit that gives us 13 bits (rounding up). For the sake of argument, add another 3 bits for your EBU alignment level of -18dBFS that gives us 16 bits.

In practice, a tape that good is going to be extremely rare. The equipment used to record the cassette and/ or the average noisy domestic environment where it was recorded, will trump that noise floor. So I cannot think of an evidence based argument for going to 24 bits for cassette. Surely there is just no information to be recovered from the noise floor?

FLAC is great, I use it myself. But it can get quite time consuming to convert to and from, since AFAIK no DAW can work with it directly. Can Reaper?

I'd recommend using a DAW with non-destructive editing such as Reaper or Pro Tools so you don't have to worry about your original files. Reaper is fantastic, it includes a decent denoiser already.
User avatar
Tomás Mulcahy
Frequent Poster
Posts: 1799
Joined: Wed Apr 25, 2001 12:00 am
Location: Cork, Ireland.

Re: The Twelve Steps for Recording Tapes into your computer

Postby Hugh Robjohns » Mon Sep 07, 2020 12:51 pm

Tomás Mulcahy wrote:Woah, bit OTT there Hugh. While I accept it's a white lie, it's not BS. I tried to pitch the article so it's not too technical...

Sure. I get that... in which case all you really need to say that combining both tracks brings a small signal-to-noise ratio improvement.

No white lie, and no BS.

And if the readers have any understanding of tape recording at all, or if you want to provide a useful handle on why that it, you just need to point in the direction of the fact that a wider track inherently has less noise -- one of the fundamental tenets of magnetic tape recording (and broadly a valid explanation for this specific situation).

However, suggesting that the reason for the improved SNR is related to the way a balanced interface works is quite misleading* IMHO and potentially inaccurate** too so, in my view the analogy is more harmful than helpful.

But that's just my opinion, for what it's worth...


* If readers of limited technical knowledge have an appreciation of the balanced interface at all, it will be a sketchy understanding of its interference-rejecting properties... and so they are likely to assume that this dual-mono track technique has something to do with rejecting noise... which it very much isn't!

** Perhaps the background to your use of the description is based on the notion that an electronic symmetrical balanced interface generates equal levels of random noise on each leg, so the summation of incoherent noise verses coherent signal results in a 3dB overall SNR improvement. ... And that's perfectly true as far as it goes....

But it doesn't apply to the alternate forms of balanced interface... and the comparison crumbles further when it's realised that an unbalanced interface is inherently quieter anyway... And then there's more potential confusion concerning the differential nature of a balanced interface which doesn't apply to the tape situation....

So it's really just stirring up a load complications and confusions that inevitably just get in the way of understanding the real points being discussed.
User avatar
Hugh Robjohns
Moderator
Posts: 28330
Joined: Fri Jul 25, 2003 12:00 am
Location: Worcestershire, UK
Technical Editor, Sound On Sound

Re: The Twelve Steps for Recording Tapes into your computer

Postby James Perrett » Mon Sep 07, 2020 1:08 pm

Tomás Mulcahy wrote:FLAC is great, I use it myself. But it can get quite time consuming to convert to and from, since AFAIK no DAW can work with it directly. Can Reaper?

I'd recommend using a DAW with non-destructive editing such as Reaper or Pro Tools so you don't have to worry about your original files. Reaper is fantastic, it includes a decent denoiser already.

Yes, both Reaper and RX can work on FLAC files directly. I nearly always record directly to FLAC unless I'm working with vinyl where I'll want to de-click the file in Audition- my old version doesn't accept FLAC

I always denoise files in RX and save them as a different name as denoising runs at real time or slower on the highest quality settings on my 2nd generation i7 laptop. I may also want to denoise different sections of audio with different settings and I'll often do two or more passes of light denoising rather than one pass of heavy denoising.
User avatar
James Perrett
Moderator
Posts: 9679
Joined: Mon Sep 10, 2001 12:00 am
Location: The wilds of Hampshire
JRP Music - Audio Mastering and Restoration. JRP Music Facebook Page

Re: The Twelve Steps for Recording Tapes into your computer

Postby James Perrett » Mon Sep 07, 2020 1:18 pm

ken long wrote:You could also briefly mention baking cassette tape out of its shell or re-lubricating dry tapes that don't play.

Do you find you need to do this often? I've yet to find a cassette that needs baking although I seem to remember as a youngster having a tape that left gunk all over the heads after only a few months of use.

ken long wrote: I'd add the Tascam 122 as a good deck as it has relatively easy access to the azimuth screw and Dolby B and C as well as being a balanced output. Totally agree about leaving Dolby out. "If in doubt, leave it out!!"

Yes, the Tascam 122 has a good reputation and, if it is anything like the 112, you have direct access to the azimuth screw through a notch in the case under the door. However, I've found that most cassette machines have good access to the azimuth screw if you know where to look. Often you just need to slide off part of the door on front loading decks and most of mine have the door fronts permanently removed.
User avatar
James Perrett
Moderator
Posts: 9679
Joined: Mon Sep 10, 2001 12:00 am
Location: The wilds of Hampshire
JRP Music - Audio Mastering and Restoration. JRP Music Facebook Page

Re: The Twelve Steps for Recording Tapes into your computer

Postby Tim Gillett » Mon Sep 07, 2020 2:28 pm

I've rarely come across cassettes that needed baking. Ampex 20+20, a Denon LX90 and another Ampex 6--? I'm not aware of a definitive list anywhere. ........................

As for azimuth the ideal is a true front panel adjuster. I've custom made them using just hand tools for Tascam 122 mk II and III, Nak classic mech types and for some of my own VHS and Beta machines. I don't know anyone who even bothers to adjust audio azimuth on VHS and Beta and you can tell by the muffled sound. Essential IMO. Saves so much fiddling around too. Poking a screwdriver around always struck me as amateurish and even dangerous to the mech.
Tim Gillett
Frequent Poster (Level2)
Posts: 2212
Joined: Wed Jan 30, 2013 1:00 am
Location: Perth, Western Australia

Re: The Twelve Steps for Recording Tapes into your computer

Postby Tomás Mulcahy » Mon Sep 07, 2020 3:41 pm

Reaper is so much better than Pro Tools in so many ways. I really need to action my transition to it.

Hugh Robjohns wrote:Sure. I get that... in which case all you really need to say that combining both tracks brings a small signal-to-noise ratio improvement.
Duly noted. If only you'd said that first without the BS bit!
User avatar
Tomás Mulcahy
Frequent Poster
Posts: 1799
Joined: Wed Apr 25, 2001 12:00 am
Location: Cork, Ireland.

Next