Dotted output with v0.3.0 #3

New Issue

LightArrowsEXE · 2020-06-19T00:52:51+02:00

LightArrowsEXE commented

2020-06-19 00:52:51 +02:00

Ever since updating to v0.3.0, weird green and/or black dots frequently appear on my output when filtering in float bitdepths. I assume this happens for illegal pixel values outside of the limited range colorspace, but I don't know of an easy way to verify it. Here's some example screenshots taken with the following code snippet:

t = depth(src, 32)
a = kgf.adaptive_grain(t, 0.1, luma_scaling=4)
comp = lvf.comparison.stack_compare(t, a)

When limiting it with core.std.Limiter(min=0, max=1, planes=0), the issue is no longer present.

So evidently, there's a range issue still present in the code. However, unlike failing and notifying of the issue with illegal ranges, it now adds this to the output.

In other words,

pls fix

Ever since updating to v0.3.0, weird green and/or black dots frequently appear on my output when filtering in float bitdepths. I assume this happens for illegal pixel values outside of the limited range colorspace, but I don't know of an easy way to verify it. Here's some example screenshots taken with the following code snippet: ```py t = depth(src, 32) a = kgf.adaptive_grain(t, 0.1, luma_scaling=4) comp = lvf.comparison.stack_compare(t, a) ``` ![](https://files.catbox.moe/4njq6c.png) ![](https://files.catbox.moe/ueor86.png) When limiting it with `core.std.Limiter(min=0, max=1, planes=0)`, the issue is no longer present. ![](https://files.catbox.moe/caroxu.png) ![](https://files.catbox.moe/wm9xy3.png) So evidently, there's a range issue still present in the code. However, unlike failing and notifying of the issue with illegal ranges, it now adds this to the output. In other words, pls fix

kageru commented

2020-06-22 13:58:19 +02:00

Fixed in 79596bc4bd.

Fixed in 79596bc4bd.

👍 1 ❤️ 1

kageru closed this issue

2020-06-22 13:58:23 +02:00

Sign in to join this conversation.