this is just awesome..big thanks yet again! I am using lodepaint now to modify a few building site plans I have as pdf files and it works great will try out the save as jpg option when I get home tonight ..
also how do you clear the screen ?? i.e when I do alot of erasing and need to refresh the screen .
@kas1e Just tryed it..........WOW! major speed boost! it is actually quite usable here compared to first versions is night and day!
I was wondering (now that it's quite usable he he), is the "Magic Wand" supposed to work? (I select it, click inside the window as I do in PS but no go, only rectangle selection works).
Also, is there/will there ever be support for layers?
I was wondering (now that it's quite usable he he), is the "Magic Wand" supposed to work? (I select it, click inside the window as I do in PS but no go, only rectangle selection works).
I noticed that too today. Looks like it not works at all (will write to main author about).
Quote:
Also, is there/will there ever be support for layers?
Yes, it will be implemented (just not right now)
Quote:
Finally will there ever be ASL saving as well?
Yes of course. Rigth now i just think how will be better: wait while Rigo will implement need it feature to ASL, or just make gui selector myself (or with help of Trixie) and use it while ASL will not have need it improvments.
@DAX For example on my pegasos2 with 1ghz , when i works with lodepaint i have maximum cpu loading at 85%. On sam800 someone say that is 96%. So we can expect from 1800mhz 50-60% of cpu loading. Cant say exactly will it any speed up LodePaint or not (because it even for me 80%, so, CPU not fully loaded).
I more think that problem exactly in our slow realisation of OpenGL. It can, and should works faster in 2 times. Maybe P96 somehow slow-down it too. Dunno.
For now i think that CPU will be not so important problem (just because my 80% only). Imho faster bandwich will help more, and of course, opengl speeduped in twice (but something make me think, that x1000 will not have any new opengl for begining). It will be cool, if it will just have warp3d drivers for new gfx card (because very possible that not, and only 2d for begining).
In other words - cant say how x1000 will be done, need to test :) But on winxp on celeron 1.1ghz and radeon7000, it works absolutly fast, smooth, and FPS all the time are in 2 - 2.5 more than for me on peg2 (and it faster because of everything - faster hardware itself (not cpu itself) , better opengl implementation, etc)
In general i hope of course, than on x1000 it will just fly, with low cpu-loading and so on. IT also interesting to test lodepaint on new upcoming sam (where as i can see will be fast hardware itself). Never check hardware details of it deeply, but hope it will be faster than peg2 by HW.
ps. I write to author about non working magic wind already.
ps2. Btw, you can check speed by FPS counnter too (number in the bottom-right area of lodepaint), to see make CPU differences or not.
Edited by kas1e on 2010/7/8 16:51:59 Edited by kas1e on 2010/7/8 16:52:36 Edited by kas1e on 2010/7/8 19:58:32
Lets test LodePaint on different machines, to see which differences have place. For testing lets use 16 bit wb and 32 bit wb, for 3 resolutions. Resolutions are changes in file "settings.txt" (in root of latest LodePaint directory). Line 21 and 22 mean X/Y:
FPS counter are at bottom/right area of LodePaint.
So, for my peg2 / 1ghz / radeon 9250 / aos4.1 update 2 :
Quote:
------------------------ 640 x 480 x 16: 56 FPS - when load lodepaint and do nothing while FPS counter are stay stable. 19 FPS - when choice "Pixel Brush" with default size 8, and maniacally draw on screen in loop while fps counter not start to be stable
640 x 480 x 32: 41 FPS - when load lodepaint and do nothing while FPS counter are stay stable. 17 FPS - when choice "Pixel Brush" with default size 8, and maniacally draw on screen in loop while fps counter not start to be stable -------------------------- 800 x 600 x 16: 45 FPS - when load lodepaint and do nothing while FPS counter are stay stable. 17 FPS - when choice "Pixel Brush" with default size 8, and maniacally draw on screen in loop while fps counter not start to be stable
800 x 600 x 32: 31 FPS - when load lodepaint and do nothing while FPS counter are stay stable. 15 FPS - when choice "Pixel Brush" with default size 8, and maniacally draw on screen in loop while fps counter not start to be stable -------------------------- 1024 x 768 x 16: 36 FPS - when load lodepaint and do nothing while FPS counter are stay stable. 16 FPS - when choice "Pixel Brush" with default size 8, and maniacally draw on screen in loop while fps counter not start to be stable
1024 x 768 x 32: 18 FPS - when load lodepaint and do nothing while FPS counter are stay stable. 11 FPS - when choice "Pixel Brush" with default size 8, and maniacally draw on screen in loop while fps counter not start to be stable -------------------------
Will be cool if someone can do exactly the same on their hardware (SAM, A1). Then we can understand better what we can expect from new HW. And even its just interesting to see difference beetwen current HW on real live example.
There is a little video to show how LodePaint looks like and works on my machine: youtube link
@Daxx Have answer from author about MagicWand:
Quote:
Magic Wand works, but it works on the mask instead of the selection. To see the effect of Magic Wand, enable the mask with the toolbar button labeled "Toggle Mask", then try Magic Want on a picture. I think I'll update the tooltip of it to make that clear :) Non rectangular selections are not supported yet unfortunately, so Magic Wand can only work on the mask.
This is my speed test on Sam Flex 800 with 1 GB Ram + Radeon 9250 (AmigaOS 4.1.2)
------------------------ 640 x 480 x 32: 22 FPS - when load lodepaint and do nothing while FPS counter are stay stable. 14 FPS - when choice "Pixel Brush" with default size 8, and maniacally draw on screen in loop while fps counter not start to be stable ------------------------
------------------------ 800 x 600 x 32: 19 FPS - when load lodepaint and do nothing while FPS counter are stay stable. 12 FPS - when choice "Pixel Brush" with default size 8, and maniacally draw on screen in loop while fps counter not start to be stable ------------------------
------------------------ 1024 x 768 x 32: 14 FPS - when load lodepaint and do nothing while FPS counter are stay stable. 9 FPS - when choice "Pixel Brush" with default size 8, and maniacally draw on screen in loop while fps counter not start to be stable ------------------------
@kas1e I have the same system as Samo above, I can add some more data ("USE" is done by using the 8 pixel Brush moving very fast until FPS counter is stable):
idle 640x480 x 16Bit = 26 FPS USE 640x480 x 16Bit = 13 FPS
idle 800x600 x 16Bit = 20 FPS USE 800x600 x 16Bit = 12 FPS
idle 1280x720 x 32Bit = 14 FPS USE 1280x720 x 32Bit = 10 FPS
I noticed that by choosing the 1 pixel pencil tool even when idle the frame rate increases, also I reckon that minimum frame rate (when using the tool) stays quite "similar" no matter the resolution! Idle frame rate changes way more instead. For example 640x480@16 gives "13" frames, while 1280x720@32 gives 10 frames (it drops just 3 frames even if the resolution and color depth is much higher).
So, for the same OpenGL implementation we have different results beetwen Sam800 and peg2/1000mhz. That is good, because now we can say that new HW will speedup it even with current OpenGL implementation.
Still, wIth having in mind, that AOS4 opengl can be faster on peg2 in compare with morphos one, in 2 times, then, even on 800/1ghz hardware it can works at normal speed if Hyperion will make almost the same opengl by speed as morphos one.
@kas1e I was thinking about this issue: while we see big difference in idle performance between Sam and PegII at several stages, when "drawing" at high resolutions and color depth the difference is negligible (like 1 FPS).
Could this be where the CPU actually becomes irrelevant and our GL problems come in? Or maybe it is a banwidth bottleneck we're hitting (like for example the 133Mhz FSB)?
Yes, difference in USE about 2-3 FPS between the same modes for sam800 / peg2.. Of course 200mhz of difference giving only 3 FPS speedup maximum are not so CPU related..
Dunno about bandwidth bottleneck .. If i remember right, Pegasos2 have better/faster HW in compare with Sam ? If it's indeed really better, then it's opengl problem. If Sam in terms of HW almost the same as Peg2 (all that cashes, speed of PCI and alt) then maybe HW related :)
But it's interesting about almost the same "use" results. Interesting and opengl-on-aos4-suck-in-mind coming :) Maybe someone more skilled in that terms have any ideas about ..