[Cin] Plugins in integer or in float

[Andrew Randrianasulu]

пн, 6 нояб. 2023 г., 01:53 Andrea paz <gamberucci.andrea at gmail.com>:

> > sorry for absence of easy answer ...
> You are right: I cannot understand the operation by superficial
> reading of the code. Each plugin has all the strings that I thought
> could discriminate their operation.
> Trying to read the code of histogram.C I found:
>
> CLAMP(output, 0, 1.0);
> case BC_RGBA_FLOAT
> // uint8 rgb is 8 bit, all others are converted to 16 bit RGB
>
> The first CLAMP might indicate the normalized range (0, 1.0); The
> comment seems to indicate that these are integer calculations (8 or 16
> bits); etc. Plus in the mailing-list (later reported in the manual) GG
> was talking about 8- or 10-bit calculations, not 16-bit.
>

well, does histogram beizer works better?

do you have simple test file we can load, add plugin in rgba_float), output
and easily see that it clipped?



> In short, I don't have the ability to discriminate how plugins work.
> I thank you for the explanations and the tests you have pointed out.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cinelerra-gg.org/pipermail/cin/attachments/20231106/46528d5b/attachment-0001.htm>