MobileRead Forums - View Single Post

NiLuJe · 12-21-2018, 12:58 PM

@jackie_w:

Great, that seems to concur with my findings on the H2O

.
And yeah, without dithering on our end, JPG does the job okay for B&W, and the difference in file-size vs. PNG is minimal, and can swing both ways depending on the content. (I think I mentioned that I messed up my previous batch of JPG tests by starting with a completely mangled over-compressed JPG? That heavily biased my previous results in favor of PNG. Not this time ;p.).

Yeah, to be honest, I only caught the Forma thing by doing the tests (well, I had noticed the size inconsistency when updating the patch, but that's it), before delving into it, things generally looked okay-ish

. But it does favor doing as much processing ourselves for now (especially since I couldn't get it to do the dithering on its own

).
You'll see that I tried to feed it an A1-sized image, and the results are not much better, so I went against it, as it mangled dithering a tiny bit more.

P.S.: Ahaha, I'd completely forgot about the button thing, I'll add that the to TODO list

.

EDIT: Updated the patch after another review pass

. Made things slightly simpler, and the new settings hopefully fool-proof

.

12-21-2018, 12:58 PM	#284
NiLuJe BLAM! Posts: 13,478 Karma: 26012494 Join Date: Jun 2010 Location: Paris, France Device: Kindle 2i, 3g, 4, 5w, PW, PW2, PW5; Kobo H2O, Forma, Elipsa, Sage, C2E	@jackie_w: Great, that seems to concur with my findings on the H2O . And yeah, without dithering on our end, JPG does the job okay for B&W, and the difference in file-size vs. PNG is minimal, and can swing both ways depending on the content. (I think I mentioned that I messed up my previous batch of JPG tests by starting with a completely mangled over-compressed JPG? That heavily biased my previous results in favor of PNG. Not this time ;p.). Yeah, to be honest, I only caught the Forma thing by doing the tests (well, I had noticed the size inconsistency when updating the patch, but that's it), before delving into it, things generally looked okay-ish . But it does favor doing as much processing ourselves for now (especially since I couldn't get it to do the dithering on its own ). You'll see that I tried to feed it an A1-sized image, and the results are not much better, so I went against it, as it mangled dithering a tiny bit more. P.S.: Ahaha, I'd completely forgot about the button thing, I'll add that the to TODO list . EDIT: Updated the patch after another review pass . Made things slightly simpler, and the new settings hopefully fool-proof . Last edited by NiLuJe; 12-21-2018 at 04:57 PM.