Hello

This post regards what will basically obsolete the "render" or "export to PNG" function if implemented.



Please try this in inkscape:

Open a new inkscape file
create a total of two additional layers for a total of three
somewhere on the page, place a rectangle on/in the lowest layer of the three
use a stroke of black, fill of white and a stroke width of 10 px

place another rectangle, wholly within the perimeter of the first
place this object within the second layer in your stack
use a stroke of black, fill of red and a stroke width of 10 px

place yet a third rectangle, wholly within the perimeter of the second
place this object within the third layer in your stack
use a stroke of black, fill of blue and a stroke width of 10 px

you should see three rectangles with the red apparently within the white and a blue within the red. All three are on different layers.

select the white rectangle

make bitmap copy

what gets saved? Only the WHITE rectangle!

This needs to be able to "shoot through" all of the layers above it as well as capturing the selected layer. This function will then be similar to what happens in traditional animation when a cell is placed over a setting and" shot through" to capture the image to camera. Doing it this way does not collapse the image. The layers will not need to be compressed. Not everything - including the invisible edges of the marbled ink layer etc - will need to be selected to be shot.

One file with one or two layers - selecting everything is easy
In files with several hundred layers - forget something and it can cost a lot of time to correct it.

This improved camera - make bitmap image - will allow the artist to take a shot, relocate a character on another layer ( or turn off one layer and turn on another ) and take another shot which captures everything. Do this at 24 frames a second for a short animation of 30 seconds - just once - and you will see the extreme benefit of the improvement.

Also, as long as the final bitmap camera is at the same pixel size as the originally selected layer ( 4000 x 4000 pixel layer results in 4000x4000 pixel bitmap ) and if this final image has the same or "similar enough" fidelity as the original vector drawing - rendering will not be necessary. Export to PNG will not be necessary. Hours of our day will be saved.




Tanks
James

Oh gosh I'm taking a huge risk here, because my question will probably lead to answers that I don't understand. But yet, I have to ask (sigh).

Why would one want to do this? What can you do with it? I understand your example of animation, but Inkscape, as yet, has no native animation ability. Are you thinking WAY into the future? Or might there be some current use for such a feature.

We've had a few messages in the last few years, where people were inquiring about camera-like abilities, for one reason or another. And perhaps Inkscape will get there some day. I can see a Camera menu, from my humble, ancient Egyptian mind, lol. Or maybe Camera Mode, or something. I'm as forward thinking as the next gal. But for me, Inkscape is already such a challenge. A feature like that seems almost like science fiction, to me! And my head kind of already aches over all the new technical stuff I would have to learn. I'm really kind of a crayon on paper artist, in some ways.

Not that I don't think it's a great idea, or even a possibility. And I hope Inkscape could someday have a camera type of ability. But for now, that's "out there" dude

Hello

brynn wrote:Why would one want to do this? What can you do with it?

Read the complaints in the other posts. Export to PNG ( render to PNG ) takes too long. It would be a very great time improvement if the system allowed the photographing of the work rather than having to wait for every node to be recalculated.

Save time = save $$$$

brynn wrote:But for now, that's "out there" dude

What is so "out there" about it? Simply improve the < Make bitmap copy > so that it copies everything within a selected region rather than just items on a selected layer.

The artist would:
Create a backer layer - say - 8000x4500 pixels
build his work on his layers above that backer
when time to "make bitmap" from the vectors - he would select the backer layer
the "camera" would then "make bitmap copy" but it would have an option which would allow " copy zone " rather than just copy selected
the "camera" would basically take a shot of the desired zone and save it as a bitmap
( this is similar to the capture screenshot command in the operating systems )

This would eliminate the need to wait for the rendering process.

There is no need for f-stops, and other "camera" functions. It is simply another, more precise, $ saving form of a screenshot command.


Tanks
James

Hi

This functionality - to export layers under a "camera square" is already in place in Inkscape.
( At least on Inkscape 0.48+devel r12322 )

The square on layer "square" will act as a camera when the "selection" is exported;


(Click image for larger version.)

Result:


(Click image for larger version.)

the "camera" would basically take a shot of the desired zone and save it as a bitmap

In order for the "camera" to take a "snapshot" the image will have to be in a raster (bitmap) format somewhere
- either in your display memory or in another "chunk" of memory on your system - to get there it has to be "rendered" - one way or another
- to avoid having to -re-render it when exporting you would need to have a screen with 4000 x 4000 pixels ( I wish ).

Good Luck
RGDS
Ragnar

Hello

If the bitmap must already exist for this to work - then the multi-board render farm to manage the render ( export to PNG ) is the only real option. GPU could enhance the render and print to screen/monitor, but, their codecs and such may limit their appealability when compared to a multi-board render farm option.

Tanks
James

Hi

If the bitmap must already exist for this to work
- then the multi-board render farm to manage the render ( export to PNG ) is the only real option.
GPU could enhance the render and print to screen/monitor, but, their codecs and such may limit their
appealability when compared to a multi-board render farm option.

Is this something you know or do you only assume?
What are the GPU's "Codecs and such"?

Remember that each Motherboard & CPU in a render farm needs; cabinet - power-supply - boot options (hard drive) - memory - LAN connection.

By the way - most render farms today use GPU for the rendering.

RGDS
Ragnar