MusyX Sample Repacking? by ShrineFox at 4:49 PM EDT on March 23, 2016
Hi, thought I'd make a separate thread for this since it's a whole different undertaking.
I used Nisto's MusyXExtract.py to extract the sound effects from PM:TTYD. I'm in the process of modding that game, so far text and graphics have been edited and all that's left is sound!
I was able to make new .dsp samples that I wish to replace the originals with, and I'm wondering if it would be easy/plausible to write a repacker. Currently, I can use a hex editor to replace the existing sounds in the ISO by pasting a smaller file over the samples. However, the compression to get the files that small requires a significant sacrifice in quality-- so the obvious next step would be to possibly rebuild the samp/sdir files to account for the bigger filesizes.
If it's too much to ask for I apologize, in which case (with permission and credit given ofcourse) I could attempt to reverse the script myself, but Nisto is clearly more experienced with Python and the understanding of the format. Any help/advice would be greatly appreciated!
For reference, I'm not using the official gamecube SDK tool to convert the new samples. I'm using this that I found on gbatemp:
I guess I should also ask if it's an option to increase the filesize of the samp by erasing the original sample data and pasting the data of the bigger dsp's in their place. I could try that but I don't know if it would screw it up by moving the sample offsets. Is that important to the structure of the format?
Repacking samples is definitely doable. We'd mainly need to update the metadata and coefficients in whichever file that is, and then rewrite offsets as we build the .samp file. It's been a while since I looked at/took an indefinite break from MusyX stuff. I still hope we can get pitch bending figured out for the sequence data.
I'll look into it myself since it's not a heavy task. This weekend is when I have time, but I don't mind if someone beats me to it.
Alright, sweet. Thank you! I should mention that I figured out how to keep the quality when I convert WAV to DSP though, needs to be encoded as 32-bit float instead of 16-bit PCM. After that it sounds flawless:
Also, some of the loss in quality had to do with me not also replacing the header that comes with the newly converted DSP.
So even though the sounds are perfect now, it's still a lot of painstaking work to do this via hex editor. I've got like 78 more samples to replace this way...