PDA

View Full Version : FPS in 3ds max and MAYA


myarchitect
10-14-2008, 09:40 AM
hello guys!
anyone knows why the fps in 3ds max is often 5-10 times slower than it is in maya,with same hardware configs,same scenes,and latest drivers on my computer???

I am really confused about the big difference,thanks for answers!!

dan1el
10-14-2008, 02:25 PM
preferences perhaps ?

Maya default I think is to play at 25 fps, dropping frames if it can't handle it, you can change it to display every frame.

Don't know about Max, since I don't use it, but probably have similar option.

myarchitect
10-14-2008, 08:44 PM
how to set the fps defulat in maya??

dan1el
10-15-2008, 12:26 PM
Window/SettingsPreferences/Preferences

go down to Timeslider (Under Settings) there you have Playback speed, default is "Play every frame" I see...so it might get slower to be able to play every frame, so it was opposite of what I said in the other reply (I've just reinstalled with new settings) but check it out what it says in your installation.

CornFielD
10-15-2008, 01:03 PM
I think that he is referring to the viewport "refreshing" fps. not the animation settings.
I do agree that mayas viewports are way quicker than maxes

dan1el
10-15-2008, 03:14 PM
I think that he is referring to the viewport "refreshing" fps. not the animation settings.
I do agree that mayas viewports are way quicker than maxes

My replies always referred to the viewport speed, not the animation settings.

maantas
10-15-2008, 04:43 PM
might be interesting for u to have a look at this:
http://4.bp.blogspot.com/_biEEh3rn_3c/SOOWo5ozF5I/AAAAAAAAAJo/SACw7F3Fq94/s1600-h/benchmark.jpg
http://4.bp.blogspot.com/_biEEh3rn_3c/SOOWo5ozF5I/AAAAAAAAAJo/SACw7F3Fq94/s1600-h/benchmark.jpghttp://4.bp.blogspot.com/_biEEh3rn_3c/SOOWo5ozF5I/AAAAAAAAAJo/SACw7F3Fq94/s1600-h/benchmark.jpghttp://4.bp.blogspot.com/_biEEh3rn_3c/SOOWo5ozF5I/AAAAAAAAAJo/SACw7F3Fq94/s1600-h/benchmark.jpghttp://4.bp.blogspot.com/_biEEh3rn_3c/SOOWo5ozF5I/AAAAAAAAAJo/SACw7F3Fq94/s1600-h/benchmark.jpghttp://4.bp.blogspot.com/_biEEh3rn_3c/SOOWo5ozF5I/AAAAAAAAAJo/SACw7F3Fq94/s1600-h/benchmark.jpg
http://4.bp.blogspot.com/_biEEh3rn_3c/SOOWo5ozF5I/AAAAAAAAAJo/SACw7F3Fq94/s1600-h/benchmark.jpg
http://softimagenews.blogspot.com/2008/10/results-are-in-xsi-beats-3ds-max-and.html
http://www.jonpeddie.com/special/WhitePapers/qualifying-performance-in-3d-modelling.php

Emil3d
10-15-2008, 05:58 PM
...
I see...so it might get slower to be able to play every frame, so it was opposite of what I said in the other reply .... Most of the time it gets slower depending on how fast your computer is and the complexity of the animation, but sometimes with simple scenes and a powerful computer playing every frame may get faster than Real Time.
I think that he is referring to the viewport "refreshing" fps. not the animation settings. .... There is no rendering animation settings for FPS in Maya.
The Real Time that appears in the Time Slider preferences is set completely arbitrarily in the Settings section of the Preferences for the Time units option. Changing the Time units only changes what it will say for Real Time in the Playback Speed in the Time Slider section, it doesn't have any effect on your rendering. You only render a sequence of images and those individual images don't carry any information that they are part of an animation or the speed they suppose to be played back when used as animation sequence. You have to assign that in your composting program with a value that matches the speed you used to preview your animation in Maya. Even if you render as AVI your animation will be rendered at 30 FPS regardless of you Time settings in the preferences. So if you render to AVI (which is not very good idea most of the time) you have to set your Time settings to 30 FPS, to preview your animation correctly.
Think of it this way: Time settings which is what appear for Real Time in the time slider settings is a previewing animation tool that lets you set the viewport playback speed in Maya to match the intended speed of your final animation that will be created with another compositing program.

CornFielD
10-15-2008, 06:06 PM
Emil3d I am totally aware of that...

but I think that there is some confusion in this thread. correct me if I am wrong, but people seems to be discussing about playback fps here, and I think that the original poster is referring to the viewport refreshing fps which can be shown from the headsup display. :beer:

Emil3d
10-15-2008, 06:51 PM
Yep, you are right, I didn't read very carefully. Sorry for the confusion.

edit: I have both Maya and Max on the same computer and before updating to the more recent video drivers (the latest was released today) max viewport performance was a total disaster comparing to Maya but now now it is doing very well may be even better, I'll make a test an let you know.

Emil3d
10-15-2008, 07:24 PM
I just made a test with a simple cube with the default shading of both programs - versions 2009. In both programs units set to centimeters. Cube is 1x1x1. Distance to travel 26 units for 3000 frames. I matched the camera views by eyeballing in both programs. Frames to render 3000 with looping palyback.
On my computer,
Maya vewport playback FPS: stayed in a very narrow range between 230 - 260
Max vewport playback FPS: stayed in a much wider range but very high 400 - 470, however it dropped a few times in a several loops test to around 300 for a very breif moment.
With this test Max performance with the latest driver on my computer is twice as fast than Maya's

CornFielD
10-15-2008, 08:59 PM
ya they might have speeded up something there. I am still using max 2008 at work, so don't have any xperience with 2009. Good to hear that they have done something about it.

I don't completely still trust your test. I think that u have to do a test which is a bit "heavier". From my xperience max started to be sluggish when there was high polycount.

martinw
10-16-2008, 04:15 AM
Maya vewport playback FPS: stayed in a very narrow range between 230 - 260
Max vewport playback FPS: stayed in a much wider range but very high 400 - 470, however it dropped a few times in a several loops test to around 300 for a very breif moment.
With this test Max performance with the latest driver on my computer is twice as fast than Maya's

Emil, your test is not really very realistic. I am not that familiar with what goes on in the Max viewport, but from the Maya side once you up to 200fps or so you are really just timing the overhead of the general Maya update loop, ie the time to update the HUD displays like FPS, the grid, view axis, UI callback queries, redrawing camera name in viewport etc. The amount of time spent actually updating the 1x1x1 cube in the scene is completely overwhelmed by these other factors. You would be better off creating a much heavier model and bring the frame rate down to a few frames per second to do viewport performance comparisons. Those will be much more significant to a user than the difference between 200fps and 400fps. Even then, there are so many different code paths in both apps that you would need quite a comprehensive set of tests to get a good overall comparison.

I'm not trying to push Maya here. We talk to Max developers regularly and share ideas for performance improvements. Over time we certainly hope both will see increasing performance.

Emil3d
10-16-2008, 06:23 PM
Martin, I'm not convinced myself either that my tests are realistic and have any meaning but since I happened to have the latest versions of both programs on my computer, it was not a big deal for me to give some information (realistic or not) to the original user who can draw his own conclusions. Thank your for your expert explanation how to interpret these kind of tests.

Meanwhile, I made another quick test. This time with a wall of cubes animated to move in 3000 frames as a group from left to right on the screen facing the camera. The wall is built with 100 cubes spaced one cube apart. Each cube is 10x10x10 cm and with 100 subdivisions in all dimensions. This makes the total triangle count in the scene 12 000 000 or 6 000 000 faces. Again, I matched the cameras between the 2 applications by eyeballing. I removed all display elements, like grid, view cubes, and left only FPS displays and the name of the camera and world axis since I don't know how to hide those in Max. Both viewports were with the same size on the screen.

Maya test: view - smooth shaded with the default lambert material
FPS very steady in all playbacks 7.4 most of the time and from time to time 7.5 during the entire animation. Even when in some playbacks I tumbled freely the camera the FPS stayed steady.

Max test: smooth shaded + highlights (equivalent to Maya's) with the default material which looks more like Maya's Blinn. Max vewport has two fixed default lights and I disabled one of them to mach the single light viewport in Maya.
FPS very unsteady between each playback test ranging between 8.x to 12.x but in most playbacks it was not redrawing all cubes. The wall looked like those electronic billboards with flashing lights. However in a couple of playbacks it went through without those redraw problems with 12.6 - 12.8 FPS., and actually tumbling the camera during the playback seemed to fix the redraw problem during the rest of the playback.

CGTalk Moderation
10-16-2008, 06:23 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.