Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Been thinking about 'enhacement' to DolbyA decoder (Read 2464 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Been thinking about 'enhacement' to DolbyA decoder

The idea of this enhancement goes beyond that of decoding DolbyA, and if I make the interface general enough, it might be made to be useful with a potential future DBX, DolbyB/C or even DolbySR decoder.

The enhancement would be another layer of noise reduction, but this would be SINGLE ENDED.  Philosophically, it would be similar to the kind that takes the stats, and then filters the audio (probably FFTed or better yet complex MDCTed) based upon the noise stats.  However, some assumptions can be made about the noise stats considering the kind of material being processed.  Basically, it might have several modes, one for each tape EQ/SPEED combination used in pro situations -- and maybe a generic one for consumer and/or cassette tape (in the case of DolbyB/C or S -- but I am probably never going to waste time on DolbyS.)

So, it would go like this -- the NR system sits in front of the DolbyA or whatever, and it passes through a raw copy of the signal, and a NR processed copy of the signal.   These two signals would be used as input to the modified DolbyA decoder.   The DolbyA decoder would take advantage of the un-NRed signal to get more accurate version of the signal level, and the NRed signal after additional DolbyA processing would be the output.

The NR knows what kind of noise that is expected in the signal, and can do whatever 'magic' necessary to clean up the signal that is eventually provided as output.  So, the stats of the original are maintained for very accurate DolbyA decoding (or whatever NR system.)

The big advantage is that the NR system in front of (and in combo with) the DolbyA can get at the noise before the DolbyA does, and remove as much as reasonable -- so that when the DolbyA sees the cleaned  up signal, its NR will seem to be more effective.  The accuracy will not be lost because the DolbyA actually gets the real (original) unmodified signal to do the level detection/etc.

This seems like the ideal of all worlds, and would work in real-time (if the computers are fast enough) because the signal stats are already known to be 'tape noise.'   The goal would NOT be to remove noise beyond what is caused by the recording mechanism, but enhancements might be able to provide more NR than the initial goal also.

I am going to listen to several DolbyA recordings to see how much noise is normally left -- it is possible that it isn't really all that necessary.
Oh well...

John