Support for Fairlight CMI 3 voices

[PythonBlue]

Attempted to work with the codebase enough to implement support for Fairlight CMI Series III voices as a source format! Resources I don't currently have will likely be required to make it available as a destination format, but in the meantime, hopefully any retro sample enthusiasts with the appropriate files can try out.

Notes:

-tuning may be severely off when converted, and my experience is that that depends on the destination format, so that would need looking into as well. This version of the code should have correct tuning for SFZ conversion, though.

-filter cutoff and many other bells-and-whistles parameters of the time are currently unsupported. Silver lining is that the original sampler used no interpolation, thus the filter in it was normally intended to only control the imaging frequencies and aliasing anyway.

-CMI IIx files and earlier are unsupported despite their using the same VC extension.

[git-moss]

Thanks for the contribution. Please provide or point me to some test files. Is there a specification available for the format?

Also please improve the code a bit:

• VCDetectorUI can be removed completely since it does not add any settings (simply use MetadataSettingsUI instead)
• Instead of using several Lists (like List subvoiceID = new ArrayList();) create a class for these parameters and then use a list with instances of this class
• Instead of several times using things like "Byte.toUnsignedInt(subFuncBuff[4]) * 16777216 + Byte.toUnsignedInt(subFuncBuff[5]) * 65536 + Byte.toUnsignedInt(subFuncBuff[6]) * 256 + Byte.toUnsignedInt(subFuncBuff[7]));" use the StreamUtils helper class it provides several methods for reading little- and big endian numbers. E.g. look at readUnsigned32

[PythonBlue]

Thanks for the response, first of all.

The first request is a valid one to ask about. It seems most of the community is more restrictive than necessary over sharing the mere way the sample format works, let alone 40 year old recordings for its factory library, but while I hesitate to put my foot in the sand by sharing such samples myself, I did find this upload to their Facebook group that should help with verifying conversion integrity.

https://www.facebook.com/groups/4943488865/permalink/10160569942663866/

As for the code formatting, fair critiques: I fully admit I only started learning Java when I began working on my fork: beforehand, most of my knowledge was in Python instead. If you think the results are promising as it is with the test files linked above, let me know, and I'll try to find the time to reformat the code to be better up to spec.

[git-moss]

Thanks for the files. I am bit unsure of how useful this format is since the library is already available in several other formats but if you improve the code I will integrate it.