[DiscordArchive] I guess I don’t understand. Could you adapt the above code sample?
[DiscordArchive] I guess I don’t understand. Could you adapt the above code sample?
Archived author: schlumpf • Posted: 2018-06-04T06:39:34.983000+00:00
Original source
I guess I don’t understand. Could you adapt the above code sample?
Archived author: Quantam • Posted: 2018-06-04T06:41:42.436000+00:00
Original source
I actually haven't closely looked at the above code. But in principle a fixed point number is simply an integer mantissa plus a fixed exponent. E.g. for a 12-bit integer part + 4-bit decimal part your encoded value = actual value * 2**4
Archived author: Quantam • Posted: 2018-06-04T06:41:54.931000+00:00
Original source
Assuming you're not using some super weird encoding format where that doesn't hold true
Archived author: schlumpf • Posted: 2018-06-04T06:42:37.226000+00:00
Original source
It might totally be true. I literally don’t know.
Archived author: Quantam • Posted: 2018-06-04T06:43:19.948000+00:00
Original source
You could have weird things like flags or absolute + sign, but I've never encountered a fixed-point format like that
Archived author: schlumpf • Posted: 2018-06-04T06:43:21.978000+00:00
Original source
I also have no disassembly to prove it, making it a complete guesswork for me.
Archived author: schlumpf • Posted: 2018-06-04T06:43:40.905000+00:00
Original source
The format is described on wiki to my full knowledge.
Archived author: Quantam • Posted: 2018-06-04T06:43:49.555000+00:00
Original source
My previous primary project was a DSP-based music synthesizer. It's all fixed point.