I have done over the weekend some fun test about mantra and i wanted to know how far i can go with mantra. so i set a scene up and the resolution went to 100k x 100k and i pressed render.
after 2 days rendering i had a .pic file with a size of 44.3GB :-)

so now my problem comes when i open the image viewer i see only black. whats the best way to look a xxxxl image ?
i know this resolution isn't common but mantra surprised me with a file and i want to open it :-)

Try converting it to a tiff (use iconvert from the command line tools) and view using Photoshop?

What about using COPs?

If you use COPs, I'd make sure that Proxy Cook (Edit > Composite Project Settings) is set to something small like 1/8th first just in case.

edward
If you COPs, I'd make sure that Proxy Cook (Edit > Composite Project Settings) is set to something small like 1/8th first just in case.

Also in Preferences -> Compositing -> Cooking, change the “Max Res” from 10k x 10k to something a bit bigger.

Thanks for the tips i will try it when i'm back home…

Wolfwood

edward
If you COPs, I'd make sure that Proxy Cook (Edit > Composite Project Settings) is set to something small like 1/8th first just in case.

Also in Preferences -> Compositing -> Cooking, change the “Max Res” from 10k x 10k to something a bit bigger.

If you are comfortable in the command line, you can also try:
% mplay -w 10000 -h 10000 foo.pic
% mplay -z 10 foo.pic

Even if you aren't comfortable with the command line, if you start mplay, there's a dialog where you can specify the resolution to load the image (look near the bottom left of the dialog).

render or convert your file in .rat format , get a grid , planar map your picture and render … it worked for me with a 55000x55000 stitched panorama.

Thanks for the tips:

here is my current result:

1. mplay -w 10000 -h 10000 foo.pic
mplay couldn't load the file, i think i have go a memory limit that blocks the loading.

2. COP doesn't load the Image
i think it's the same problem as above

3. iconvert
i was getting a segmentation fault (core dumped)


i think i have got on this machine to less memory (8GB) i will try it at work there i have got (32GB) maybe it helps. :-)

mandrake0
Thanks for the tips:

here is my current result:

1. mplay -w 10000 -h 10000 foo.pic
mplay couldn't load the file, i think i have go a memory limit that blocks the loading.

2. COP doesn't load the Image
i think it's the same problem as above

3. iconvert
i was getting a segmentation fault (core dumped)


i think i have got on this machine to less memory (8GB) i will try it at work there i have got (32GB) maybe it helps. :-)

You're right, I had problems shrinking the image. Though
% icomposite -w 100000 huge.pic = Mandril.pic
worked, but the opposite
% icomposite -w 100 small.pic = huge.pic

didn't work (ran out of RAM). Which means that mplay and COPs would also likely fail.

However, .rat files seem to work:
% iconvert huge.rat = huge.pic
% iconvert -w 100 small.rat = huge.rat
% mplay -w 1000 -h 1000 huge.rat

This is likely due to the fact that .rat can read partial data from the file. Even so, some operations (like “mplay -z 10 huge.rat”) do run out of memory.

Note also, that .pic files read their first scanline from the bottom of the image, while .tif files read from the top of the image. This means that to flip the image, you need to load the whole image into RAM.. – which would need over 40 GB of RAM.

thank you for the information…. it looks like i need to buy much more memory :-)

Well, lots of discussion about technicalities, what was the picture?

You may check DeepZoom tech from Microsoft. The only problem is, it doesn't work with *.pic .

@harry_the_cat i haven't seen the picture…. whats inside the picture is nothing special just a grid with a mountain sop… thats it :-) for me it was more like a test how far mantra can go and it can go far… when i have got time i will try to do some other test but for now my work takes my spare time….
one of them was that i wanted to try if i can simulate at 1 billion voxels (1000x1000x1000) :-) it crashed at about 400-600m voxels because of memory limits…. maybe i will render a still image with that data :-).


@mantragora i know deepzoom looks intresting sadly i safed it in the wrong fileformat .rat or .exr would be better.