Dear Chimera team, Recently, I have some trouble saving images with POV-RAY option ³true². The process of rendering is too slow (both on the mac and windows version). To give the feeling of how slow: rendering an image with 1700 lines takes 7 hrs (with quality 9 in pov-ray options) and 2 hrs (quality 5). I used to get the images rendered much faster than that (with older versions; however, I¹m not exactly sure of when I started to see such a slow rendering behavior). There is another related problem but only with the windows version: If I want to save an image, with depth cue, using the fast mode (POV-RAY false), the image output lacks any depth used in the model, with poor quality and darker than it should be. Somehow, the camera orientation is messed. In addition, the depth cue will be no longer available for the rest of the session (I have to restart over again). To get the depth cue I want, I need to turn the POV-RAY option ³true²....but then too slow to create a good figure. I¹d appreciate any comment or explanation from Chimera people; and what is the best way to overcome this problem? P.S. I have the PPC version for the MAC running OSX 10.4; the PC running Windows XP. Many thanks in advance for your help. Ibrahim -- Ibrahim M. Moustafa, Ph.D. Biochemistry and Molecular Biology Dept. 201 Althouse Lab., University Park, Pennsylvania State University PA 16802 Tel. (814) 863-8703 Fax (814) 865-7927
On Mon, 2009-06-22 at 18:53 -0400, Ibrahim Moustafa wrote:
Dear Chimera team,
Recently, I have some trouble saving images with POV-RAY option “true”. The process of rendering is too slow (both on the mac and windows version). To give the feeling of how slow: rendering an image with 1700 lines takes 7 hrs (with quality 9 in pov-ray options) and 2 hrs (quality 5). I used to get the images rendered much faster than that (with older versions; however, I’m not exactly sure of when I started to see such a slow rendering behavior). ...
I have seen the same behavior. I want to create some images for posters in my lab, but the time has gone up roughly 7 to 10 fold for images at the same settings as before. An additional problem is that 2D label letters are not properly rendered, with the result looking like the letters have lost their tops, being sliced off at a slanting angle. There is scattered "debris" as well in the images. Kenward -- In a completely rational society, the best of us would aspire to be _teachers_ and the rest of us would have to settle for something less, because passing civilization along from one generation to the next ought to be the highest honor and the highest responsibility anyone could have. - Lee Iacocca
On Jun 22, 2009, at 4:26 PM, Kenward Vaughan wrote:
On Mon, 2009-06-22 at 18:53 -0400, Ibrahim Moustafa wrote:
Dear Chimera team,
Recently, I have some trouble saving images with POV-RAY option “true”. The process of rendering is too slow (both on the mac and windows version). To give the feeling of how slow: rendering an image with 1700 lines takes 7 hrs (with quality 9 in pov-ray options) and 2 hrs (quality 5). I used to get the images rendered much faster than that (with older versions; however, I’m not exactly sure of when I started to see such a slow rendering behavior). ...
I have seen the same behavior. I want to create some images for posters in my lab, but the time has gone up roughly 7 to 10 fold for images at the same settings as before.
As per my recent mail to Ibrahim, try increasing the "Antialias threshold" to 1.0 or even higher.
An additional problem is that 2D label letters are not properly rendered, with the result looking like the letters have lost their tops, being sliced off at a slanting angle. There is scattered "debris" as well in the images.
I have not seen this behavior. 2D labels look fine in the images I rendered. Does this happen on just one machine? Please use the "Report a Bug" menu item to send in a bug report so that we can find out the details of the system you're using. --Eric Eric Pettersen UCSF Computer Graphics Lab http://www.cgl.ucsf.edu
On Jun 22, 2009, at 3:53 PM, Ibrahim Moustafa wrote:
Dear Chimera team,
Recently, I have some trouble saving images with POV-RAY option “true”. The process of rendering is too slow (both on the mac and windows version). To give the feeling of how slow: rendering an image with 1700 lines takes 7 hrs (with quality 9 in pov-ray options) and 2 hrs (quality 5). I used to get the images rendered much faster than that (with older versions; however, I’m not exactly sure of when I started to see such a slow rendering behavior).
I'm guessing you're using transparent surfaces. We've made improvements in the fidelity of translation of transparent colors between Chimera and POV-Ray since the 1.3 release. In 1.3, the transparent colors were much darker than they appeared in Chimera. This change increases the rendering time because the amount of supersampling that POV-ray performs depends on how much color variation there is in the image and the new scheme provides a lot more variation than its uniformly dark predecessor. One example I tried produced slightly more than a factor of two in rendering times between the 1.3 version and the current daily build. Now, if you're rendering such a high-resolution image you don't really need as much effort expended on each individual pixel as you do for a smaller image. I find that if I increase the "Antialias threshold" POV-Ray option from 0.3 to 1.0 then I get an almost unchanged image in approximately the same time as as the 1.3 version required. You might even be able to go higher than that and still get a good image.
There is another related problem but only with the windows version: If I want to save an image, with depth cue, using the fast mode (POV-RAY false), the image output lacks any depth used in the model, with poor quality and darker than it should be. Somehow, the camera orientation is messed. In addition, the depth cue will be no longer available for the rest of the session (I have to restart over again). To get the depth cue I want, I need to turn the POV-RAY option “true”....but then too slow to create a good figure.
I understand that Greg (who works on this code) recently made some depth-cueing related bug fixes. Please try the very latest daily build and see if you still get this problem. If you still do then please use Help->Report a Bug to send us a bug report which will automatically include detailed information about the system you're using.
I’d appreciate any comment or explanation from Chimera people; and what is the best way to overcome this problem?
An alternative to ray tracing is to turn on "glossy lighting" (from the Lighting tab of the Side View). That may look good enough for what you want and works with Chimera's non-raytraced image saving. --Eric
Thanks Eric for the reply. Actually, I¹m not using transparent surfaces. The image I tried to make is a normal cartoon model. For the ³glossy lighting² option: the button with the default ³normal² is inactive. How to turn it on? I could not figure it out. Thanks, Ibrahim On 6/23/09 1:24 PM, "Eric Pettersen" <pett@cgl.ucsf.edu> wrote:
On Jun 22, 2009, at 3:53 PM, Ibrahim Moustafa wrote:
Dear Chimera team,
Recently, I have some trouble saving images with POV-RAY option ³true². The process of rendering is too slow (both on the mac and windows version). To give the feeling of how slow: rendering an image with 1700 lines takes 7 hrs (with quality 9 in pov-ray options) and 2 hrs (quality 5). I used to get the images rendered much faster than that (with older versions; however, I¹m not exactly sure of when I started to see such a slow rendering behavior).
I'm guessing you're using transparent surfaces. We've made improvements in the fidelity of translation of transparent colors between Chimera and POV-Ray since the 1.3 release. In 1.3, the transparent colors were much darker than they appeared in Chimera. This change increases the rendering time because the amount of supersampling that POV-ray performs depends on how much color variation there is in the image and the new scheme provides a lot more variation than its uniformly dark predecessor. One example I tried produced slightly more than a factor of two in rendering times between the 1.3 version and the current daily build.
Now, if you're rendering such a high-resolution image you don't really need as much effort expended on each individual pixel as you do for a smaller image. I find that if I increase the "Antialias threshold" POV-Ray option from 0.3 to 1.0 then I get an almost unchanged image in approximately the same time as as the 1.3 version required. You might even be able to go higher than that and still get a good image.
There is another related problem but only with the windows version: If I want to save an image, with depth cue, using the fast mode (POV-RAY false), the image output lacks any depth used in the model, with poor quality and darker than it should be. Somehow, the camera orientation is messed. In addition, the depth cue will be no longer available for the rest of the session (I have to restart over again). To get the depth cue I want, I need to turn the POV-RAY option ³true²....but then too slow to create a good figure.
I understand that Greg (who works on this code) recently made some depth-cueing related bug fixes. Please try the very latest daily build and see if you still get this problem. If you still do then please use Help->Report a Bug to send us a bug report which will automatically include detailed information about the system you're using.
I¹d appreciate any comment or explanation from Chimera people; and what is the best way to overcome this problem?
An alternative to ray tracing is to turn on "glossy lighting" (from the Lighting tab of the Side View). That may look good enough for what you want and works with Chimera's non-raytraced image saving.
--Eric
Hi Ibrahim, The glossy lighting quality setting is not available if your OpenGL driver and graphics card does not support shaders. Most graphics cards from the past few years support pixel and vertex shaders but older ones may not. Tom -------- Original Message -------- Subject: Re: [Chimera-users] Saving image and rendering From: Ibrahim Moustafa To: Eric Pettersen Date: 6/23/09 12:26 PM
Thanks Eric for the reply.
Actually, I’m not using transparent surfaces. The image I tried to make is a normal cartoon model.
For the “glossy lighting” option: the button with the default “normal” is inactive. How to turn it on? I could not figure it out.
Thanks, Ibrahim
On 6/23/09 1:24 PM, "Eric Pettersen" <pett@cgl.ucsf.edu> wrote:
On Jun 22, 2009, at 3:53 PM, Ibrahim Moustafa wrote:
Dear Chimera team,
Recently, I have some trouble saving images with POV-RAY option “true”. The process of rendering is too slow (both on the mac and windows version). To give the feeling of how slow: rendering an image with 1700 lines takes 7 hrs (with quality 9 in pov-ray options) and 2 hrs (quality 5). I used to get the images rendered much faster than that (with older versions; however, I’m not exactly sure of when I started to see such a slow rendering behavior).
I'm guessing you're using transparent surfaces. We've made improvements in the fidelity of translation of transparent colors between Chimera and POV-Ray since the 1.3 release. In 1.3, the transparent colors were much darker than they appeared in Chimera. This change increases the rendering time because the amount of supersampling that POV-ray performs depends on how much color variation there is in the image and the new scheme provides a lot more variation than its uniformly dark predecessor. One example I tried produced slightly more than a factor of two in rendering times between the 1.3 version and the current daily build.
Now, if you're rendering such a high-resolution image you don't really need as much effort expended on each individual pixel as you do for a smaller image. I find that if I increase the "Antialias threshold" POV-Ray option from 0.3 to 1.0 then I get an almost unchanged image in approximately the same time as as the 1.3 version required. You might even be able to go higher than that and still get a good image.
There is another related problem but only with the windows version: If I want to save an image, with depth cue, using the fast mode (POV-RAY false), the image output lacks any depth used in the model, with poor quality and darker than it should be. Somehow, the camera orientation is messed. In addition, the depth cue will be no longer available for the rest of the session (I have to restart over again). To get the depth cue I want, I need to turn the POV-RAY option “true”....but then too slow to create a good figure.
I understand that Greg (who works on this code) recently made some depth-cueing related bug fixes. Please try the very latest daily build and see if you still get this problem. If you still do then please use Help->Report a Bug to send us a bug report which will automatically include detailed information about the system you're using.
I’d appreciate any comment or explanation from Chimera people; and what is the best way to overcome this problem?
An alternative to ray tracing is to turn on "glossy lighting" (from the Lighting tab of the Side View). That may look good enough for what you want and works with Chimera's non-raytraced image saving.
--Eric
------------------------------------------------------------------------
_______________________________________________ Chimera-users mailing list Chimera-users@cgl.ucsf.edu http://www.cgl.ucsf.edu/mailman/listinfo/chimera-users
participants (4)
-
Eric Pettersen -
Ibrahim Moustafa -
Kenward Vaughan -
Tom Goddard