Dolby B decoding

I have a stack of old cassettes to digitize. In your opinion, would it be best to decode on playback, or do a straight transfer and use post-processing, e.g., a plug-in of some sort? These tapes were stored under ghastly conditions for the better part of 30 years and might be good for one pass only, if that.

I'm not aware of a Dolby-B decode plugin, or anything that would emulate the decode process easily. Would be interesting to know though.

If using a hardware Dolby-B decode, I'd listen out for HF loss, although I suppose this is best corrected with EQ as the code/decode is not a simple boost and cut as I'm sure you know.

Hi,

While 'technically incorrect', I wouldn't be surprised if the result of a direct no-decoding transfer from such cassettes would give the best subjective results.

As we'll all recognize, back then it was pretty common for consumer-grade decks & both 'Ferro' & 'Chrome' tapes to record with dolby-B engaged and not use it for playback to make up for the lost highs. dbx-NR is way more unforgiving here (actually impossible/unlistenable), but for dolby-B, skipping it wasn't too bad... imho

Regards,

  Peter

Unfortunately, there are no Dolby plug-ins available. I'm not sure why; the patents expired long ago. Perhaps nobody's done it because the trademark Dolby B still applies, so they couldn't tell anyone what it did except in general terms -- like "Decodes the most common form of noise reduction on cassettes, if you know what we mean."

Getting maximum quality out of cassettes is about two things: tape path and azimuth. The cassette actually is the tape path, and usually sucks. In old casssettes, the pressure pad that holds the tape against the head is often deteriorated, so you don't have proper contact, and your highs are gone or fluctuating.

There are two possible solutions. The best is to use a 3-head Nakamichi deck, which pushes the pressure pad out of the way and maintains tension using dual-capstans. This also cuts way down on scrape flutter, which is good to do.

The second, if you don't have access to a Nak, is to transplant the tape into a new cassette. This will reap other benefits, including fresh slip sheets (the thin paper or plastic sheets above and below the tape), which will be less likely to squeak, and new roller bearings. I do this fairly frequently when transferring cassettes; I use the shells from Maxell XLII, which is still available at Walgreen's.

The other issue is azimuth. If the playback head's azimuth doesn't exactly match the azimuth of the head the tape was recorded with, your high frequencies disappear, and Dolby decoding makes it worse. Most cassette decks have an adjusting screw next to the head, visible when it's playing; turn the screw (with a non-magnetized screwdriver!) until you get maximum high-frequency response with the Dolby off. The most accurate way to do this is to sum the channels to mono; the point of maximum HF is much easier to find in mono. Just remember to go back to stereo for transferring!

The azimuth may be different on the two sides of the tape. In fact, it may be different on different cuts. If necessary, stop and tweak azimuth several times while transferring the tape.

Do these things, and you should get okay Dolby decoding in hardware. Provided, of course, that the machine the tapes were recorded on was set up properly to begin with. But at least this will let you get the most out of them.

Peace,
Paul

If you have a hardware decoder, you could:

-first digitize in max. resolution possible
-after that, denoise (digitally, sample the noise footprint and remove with a plugin or app. - something like voxengo redunoise etc.) and cleanup
- then, process with the hardware box.

this would seem pretty redundant and failsafe to me.

In theory hell yeah...  in practice, it may not be worth the trouble to mess with it..  Dolby B only buys you 10 dB improvement below the raw cassette noise floor which might be -55 dbu on a good day, and only at HF, so the -65 dBu dolby improved noise floor will still be so far above the digital channel noise to worry about improving that.

YMMV

JR

 

And to Paul's remarks about Nak decks:  oh for the days of a well-maintained Nakamichi Dragon, which had a servoed auto-azimuth function.

I have been doing straight transfers, no Dolby B or C and using EQ and noise reduction plugs afterward, if even that.
I have been doing transfers of cass and LP, and I am glad to have the source material preserved at all!  I figure that when I have the time I can go for digital cleaning.  When I retire, I'll have some things to keep me busy besides golf or mah friggin' jong.  And who knows what will be available even 5 years from now in the digital domain.
I did a Yaz concert last year- BBC rock hour or whatever (FM) recorded to cass in 1986, Dolby B encoded, transferred to PT AIFF no plugs and it really pops.  Yes, there is hiss, but I have heard hiss on the show ever since I recorded it. 
I cleaned a performance track for a friend from 1987 or so, Dolby C, and used some Waves noise reduction to clean it.  The "mastering" time consumed was roughly 7 times the 10 minutes of music.

Anyone else have Peter Hamill's "Loops and Reels", released on cass only?  I am wondering who bought the other one. . .
Mike 

sodderboy said:

I cleaned a performance track for a friend from 1987 or so, Dolby C, and used some Waves noise reduction to clean it.  The "mastering" time consumed was roughly 7 times the 10 minutes of music.

Click to expand...

Hi,

A danger indeed, most of the various stuff I've transferred, cleaned up etc is certainly never listened to again for months - heard it more than enough during the "mastering". 

Bye,

  Peter

   

I transferred all my cassette tapes during the 80's when digital audio first started to be available.
I had access to a Nakamichi deck and spent a good deal of time getting the azimuth etc. all lined up.
However, I found that with the Dolby thing different decks obviously had different reference levels so I found it better to get a decent lift of the track and then do the Dolby processing later.
Fortunately the studio I was working at had many junked cassette decks so I found a couple of Aiwa decks and dug out the Dolby cards inside them. (They were mainly based on a Hitachi chip).
I built a box with a variable input and output level and used that to process the audio.
I've still got it.
I've even used it as an effect on a couple of projects!

As mentioned I was a Dolby licensee back in a former lifetime and while my recollections are fading, there was a standard reference calibration that Dolby machines were "supposed" to meet. Perhaps some cheap consumer machines deviated from the standard some in production but Dolby was very specific, and strict about it as I recall. There were also 3rd Party calibration reference level tape standards involved. I guess using different tape formulations, bias adjustments (or drift), head alignment, etc could impact HF response which would get expanded by the 1:2 playback expander. 

If you are seeing improvement from tweaking the level between playback electronics and decoding, I would suggest considering HF EQ in between those stages also, perhaps before major level tweaks since IIRC the level reference was calibrated at lower frequency (say 400 Hz)  and shouldn't be as affected by the sundry errors in the cassette format.

JR

I wonder if if you used a modeling reverb ala Oxford to shoot a Dolby A or SR module in decode mode and use that as a possible decoding plugin for all sorts of tape transfers.
I always thought it would be smart for Dolby to make decoding plugins to fill the void of all those SR/A racks now in landfills.
Mike

Wouldn't work.

Just as modeling reverbs can't model modulated reverb... albeit for a subtly different reason.

Remember also that there are time constants involved, plus peak detection (as opposed to RMS in dbx units)...

Almost certainly not gonna happen, unless some university kid makes it his project or something similar. -To be sure, you cannot use current reverb modeling methods.

Keith

Here:

http://refinedaudiometrics.com/products-clasnr.shtml

actual dolby B plugin. I don't understand why they have slapped together a BBE maximiser type process with dolby B, but there it is, and in very high quality. The guy who programmed these is a retired NASA DSP guy. A *real* rocket scientist, if that means anything to you people.

I don't think this is an actual Dolby B plugin. i.e. same curves and characteristics etc.
I tried this a while back and wasn't that impressed - maybe it has improved since, I don't know.
I found the PSP Mix Treble thing much better.
I'll download the demo and have a listen.

The description doesn't match my understanding of Dolby B either...

I was under the impression that B wasn't a sliding filter, rather a single-band expander, with a common control for two channels. (unlike 'A', which was a 5-band system with four single-band expanders, and NO inter-channel connection of control voltages.)

Dolby 'B' was designed from the outset for non-decoded signal to be 'tolerable'. Since even factory machines were rarely ever anywhere close to each other in terms of tape interchange, my own favoured method of decoding tapes was to use a stand-alone outboard TEAC encoder/decoder and use static EQ for any requisite brightening.

Keith

I am digitising many of my old cassettes which I used for recording gigs in the 80s and early 90s. I am using a brand new JVC TD-718 which I was kindly given recently. I am replaying with dolby off and using a little hiss reduction in Adobe Audition, and am very pleased with the results. I mostly listen to mp3 versions of these in my car and they sound pretty good to my tired ears.

SSLtech said:

The description doesn't match my understanding of Dolby B either...

I was under the impression that B wasn't a sliding filter, rather a single-band expander, with a common control for two channels. (unlike 'A', which was a 5-band system with four single-band expanders, and NO inter-channel connection of control voltages.)

Dolby 'B' was designed from the outset for non-decoded signal to be 'tolerable'. Since even factory machines were rarely ever anywhere close to each other in terms of tape interchange, my own favoured method of decoding tapes was to use a stand-alone outboard TEAC encoder/decoder and use static EQ for any requisite brightening.

Keith

Click to expand...

You are correct Dolby B is not sliding band.  Fixed band (HF only) below threshold symmetrical compressor/expander.

When properly used it does not alter source noise, just the tape record/playback HF noise.

JR

SSLtech said:

I was under the impression that B wasn't a sliding filter...

Click to expand...

Now that you said it, yes indeed it sounds like an envelope follower controlling a low-pass (2- or 4-pole) filter 3dB point.