Dolby processing in AA

The recent discussion about squeaky cassettes proved that still many people are interested in playback of this material. A common problem is Dolby processed tape. In most cases this is Dolby B or C as the professional Dolby A is much more complicated and rare. I would therefore like to have the possibility to do the Dolby playback processing in AA for type B and C.

At the first instance this seems to be stupid idea as most players that have survived have switches you may set accordingly. But Dolby works fine only when adjusted carefully to the brand of tape used and within the machine used for recording AND playback. This calls especially for careful level alignment that has to be equal in recording and playback. In most cases neither the playback machine is the same as for recording, nor is the alignment correct for the actual tape. Even then minor deviations in track width or track position may cause an offset of several dB’s and impair the process.

Processing in AA would therefore allow for different settings of the levels. Although there is no objective indication for the correct level, the possibility of shifting up and down could help to improve by experiment. Does anyone know about such a SW-tool that can be used for that purpose ? If not this might be an urgent wish for a new edition of AA ! Please concentrate the discussion on a technical level first as I know pretty well that this would eventually implicate negotiations with the Dolby labs.

For playback, only the level needs to be aligned; on most decks it's not that hard to check and adjust the alignment with a test tape.  The tape type affects recording in more ways (e.g. bias and equalisation, as well as level).

Paul

Please concentrate the discussion on a technical level first as I know pretty well that this would eventually implicate negotiations with the Dolby labs.

Can't do the technical stuff first, because Dolby beat everybody to it decades ago, by the very simple device of flatly refusing any permission for this at all. It was discussed quite extensively in at least one thread in the forum archives, but nothing's changed. They regard the processing algorithms as proprietary, and would sue anybody successfully implemeting a decode and releasing it in public.

For playback, only the level needs to be aligned; on most decks it's not that hard to check and adjust the alignment with a test tape.  The tape type affects recording in more ways (e.g. bias and equalisation, as well as level).

I know pretty well how to align a tape deck for Dolby processing both on the recording as on the playback side. However when I intend to transfer a number of (my own) cassettes to CD it is extremely impractical to run the alignment process for each tape individually. Nobody in practice would take the pain to dismantle the deck and run the test tape again and again. The result is that you rather start tweaking the frequency response in a wild manner.

They regard the processing algorithms as proprietary, and would sue anybody successfully implemeting a decode and releasing it in public.

I remember processings named DLPF, DNR, NR-Exp or similar to circumvent the  Dolby playback patent by a number of companies such as ITT, Philips and others. These algorithms have been published, and I am not sure to posses at least descriptions of the basic parameters governing the Dolby processing as well. Beyond that, Dolby should be aware that they are on a sinking ship in the digital age.

it is extremely impractical to run the alignment process for each tape individually

But there is only one correct alignment for playback, which should be stable in normal use.  Or are you suggesting mis-aligning for each tape played to match a presumed mis-alignment during recording?  In which case a test tape is irrelevant, except to provide a reference for the mis-adjustment; and judging the mis-alignment would be wildly difficult, I would expect.

Paul

I remember processings named DLPF, DNR, NR-Exp or similar to circumvent the  Dolby playback patent by a number of companies such as ITT, Philips and others. These algorithms have been published, and I am not sure to posses at least descriptions of the basic parameters governing the Dolby processing as well. Beyond that, Dolby should be aware that they are on a sinking ship in the digital age.

I don't think you remember this correctly at all - all of those companies produced ICs under licence from Dolby, mainly to produce domestic NR chips for cassette recorders. All of the people who tried to circumvent Dolby were actually trying to improve on what Ray (and Henry) had done originally - with no conspicuous success. In fact the algorithms aren't difficult (except for C, which never really caught on) - the difficulty was impementing them in hardware reliably. I have everything Dolby published at the time about this - it contains enough information to create whatever's needed, but you'd need a good plugin builder at a bare minimum...

It doesn't bother me, because I have standalone hardware versions of all of the processors that matter (A, B, SR) and a cassette recorder that does C with adjustable setup, and cassette recorders is all that was ever used for.

Thanks Paul and Steve for the comments.

But there is only one correct alignment for playback, which should be stable in normal use

That is absolutely correct. In theory. For playback you have to present the correct level to the decoding processor and unless it has errors by itself, everything is OK. In practice there are several points to consider that may cause difficulties:

1. The basic definition of levels on tape refers to magnetic flux. If you are lucky to have a really professional reference tape, it will be a full track type making sure that the gap of the playback head is fully overlapped on both edges. I don't know if there are such tapes available for CC. Mine most probably are NOT. So you hope your reference setting in playback is correct under the given assumptions which you cannot check. You rely on faith.

2. Subsequent alignment of the rest of the recorder (in the recording part) relies on point 1 which hopefully is fine. A look at the track arrangement given by regulation DIN 45516 shows that there are at least tolerances of 10 % in track width as well as in track position. There are also different specifications for prerecorded tapes and for recording and playback in home recording. Just these differences alone account for errors in the order of 2 dB excluding those that may stem from point 1 unless you have the same machine for recording and playback. Although recorders with separate recording and playback heads are superior types, in this respect they may suffer from more problems than those having a single head. Furthermore you have no indication for the correct level on a third party machine.

3. None of the machines as delivered new from the factory was really well aligned. The procedure simply is too much time consuming to be carefully made in a production line.

I have currently 3 machines as choice, all aligned carefully ( 2 x SONY WM6DC prof. Walkman, 1xDenon DRM 800 dual head dual capstan) but none of the tapes produced of any of the recorders mentioned plays correctly on any of the others. I had made "reference tapes" on a very stable TEAC also that has gone to trash (skip ?!) some years ago. Same with these tapes.

All the effects mentioned go into direction of a Dolby reference level that is too low. The result in playback is a reproduction that is rather dull despite careful azimuth correction.

I have everything Dolby published at the time about this - it contains enough information to create whatever's needed, but you'd need a good plugin builder at a bare minimum...

I consider calling one of my active colleagues at my former school - it might be a good exercise for a student !

But there is only one correct alignment for playback, which should be stable in normal use.  Or are you suggesting mis-aligning for each tape played to match a presumed mis-alignment during recording?

It doesn't bother me, because I have standalone hardware versions of all of the processors that matter (A, B, SR) and a cassette recorder that does C with adjustable setup, and cassette recorders is all that was ever used for.

But suppose:
- you don't have the hardware like Steve
- you don't have a calibration tape
- the one deck you have doesn't have the right Dolby playback
- the tape you want to restore was made in a deck that wasn't correctly aligned

What can you do with software? There are no plugins for political reasons so we have to improvise. I think this is a valid idea somehow. What would be a workflow that let you get the most out of that tape?

Look at it like doing in software some 78rpm EQ. I fear it will have the same limitations but there should be something possible.

I consider calling one of my active colleagues at my former school - it might be a good exercise for a student !

Er, I don't think so - unless you can find an exceptionally clever one. You have to develop eight bilinear compressors (for stereo use) used in their expansion mode, and these have to have carefully controlled overshoot characteristics. Also for each decoder, you need to reproduce accurately the four band splitting filters. If you start out by considering that any of this is a trivial process suitable for students, you will fail. You have to bear in mind that one of the significant difficulties with this is that you have to match the characteristics of the hardware pretty closely, or it simply won't work properly.

If you want to find out what's basically involved, then the Dolby website has a good overview technical paper about this on it. It's not for the faint-hearted though.

Er, I don't think so - unless you can find an exceptionally clever one. You have to develop eight bilinear compressors (for stereo use) used in their expansion mode, and these have to have carefully controlled overshoot characteristics. Also for each decoder, you need to reproduce accurately the four band splitting filters.

Steve you disregard that I am speaking about type B and C which don't have the complicated band splitting as Dolby A. I agree that the Dolby A case would be rather tough, but it could serve as a diploma or master work. For the time being I would be quite content to have Dolby B and C treated.

Steve you disregard that I am speaking about type B and C which don't have the complicated band splitting as Dolby A. I agree that the Dolby A case would be rather tough, but it could serve as a diploma or master work. For the time being I would be quite content to have Dolby B and C treated.

I think that you'll find that by quoting the requirements for A, that in fact I've given you the easy option - B and C use sliding filter characteristics and are way more complicated to code, especially C. There may be more devices invoked in A, but inherently they are the simpler ones.

I'm serious - don't kid yourself that there's anything easy about this at all. None of it is, which is another reason that nobody's made any serious attempt to do it.

But look at the bright side.  If you DO manage to pull off the technical aspects (and I agree it's a long way from simple) you'll win the delights of a lawsuit from Dolby labs if you try to distribute your success.