Technical questions regarding mod-use/exporting

[jiyorude]

Hi all! I am currently in early production of two more Quake III Short films. Therefore my questions might be a bit technological in nature, my apologies for that. At the moment I am running into some issues while creating testfootage and I really hope anyone can help me with this. It would be much appreciated!

1. Is it possible at all to render demo's in 4K DCI (4096x2160) or at any resolution higher than 4K on that matter? I would rather have a True 4K projection than a upscaled 2K-from-FullHD projection if it were to be featured.

2. I am currently experiencing issues in which Q3MME does not want to render any demo I put in my demo capture list on 4K resolution and gives me the 'megs' error. Something like 'you need at least [insert number here] megs'. I have had this error before and managed to find the solution. However, I am not able to find it anymore. Can anyone help me with that?

3. Is it possible to render the centerprint 'You Have Fragged' messages separately in its own image sequence aside from the PNG sequences that contain the depthmaps and the footage? I find it a bit troublesome that the depth of field maps affect the fragmessage by blurring it from time to time. Masking is possible, but sometimes does not give me the desired effect and sometimes it is really visible that I am masking it out of the depth of field map. It would be much easier for post production if i were able to render these on two different video tracks.

4. If this is not possible, can I only include frag messages related to my frags in the chat screen above? With this, I am able to fake centerprint messages myself in post-production as I know the exact moment when I fragged someone and who I fragged. It also gives me more freedom design-wise in post-production as sometimes some of my custom fonts are having issues in Q3MME. If this is not possible, is there any way of how I can give myself a heads up when and who i fragged without the use of centerprint as a sort of a 'marker'?

5. Can I change the quality of the image sequences to 16, 24 or even 32 bits for example, so they contain more color inormation during color grading?

6. If this is possible, will the bits of the depth of field maps change accordingly by themselves?

7. I am a bit confused regarding the depth of field feature in the mod. Is it correct when I say that the position of the grid in space represents where the focus point of the camera is, and the thick/thinness of the grid determines the shallowness/thickness of the blur? How can I manipulate the falloff of the blur? Do mme_depthfocus and mme_depthrange represent anything different compared to the features in the grid, or are these the console equivalent of the grid?

8. If I had two characters in a shot, one standing in front of the camera and one in the back of the camera, and I want to primarily focus on the character in the front and do a focus pull to the character in the back, with not too much depth of field (so you are able to see the character somewhat in the background), how would I pull that off? Sorry for asking this, but then I get a bit of an idea how to animate/how to setup the depth of field in the mod.

Thanks in advance for your help guys!

[ent]

1. Possible, seek for off-screen capture in Tutorials section.
2. What is the exact message?
3. Impossible. You should try the new way of DOF without depth mask.
4. As chat - no, but the console prints who killed who.
5. The game captures everything in 8 bits per channel. There are not assets in the game that contain more then 8 bits per channel so it's not needed.
6. I don't understand the question.
7. You are correct about the grid representation. Use keypoints like in other editing mods to save different radius in different moments of time and animate between them.
mme_depthFocus and mme_depthRange are used in the old DOF capture method with depth masks. The new one only uses mme_depthFocus and a new cvar mme_dofRadius if there are no keypoints. But if there are keypoints then focus and radius data is used from them.
8. Same as above, use keypoints to animate the focus point.

[jiyorude]

Thanks for the answers so far! I tried rendering today and the error showed up again in the Quake 3 console after booting up the demo while using off-screen capturing

The error is:
"Failed to allocate 530841600 bytes from the mme_workmegs buffer"

I reset the mme_render width and height and booted up again. Once in Q3MME i changed workmegs to a far higher number and restarted. After adding the workmegs command in the console it showed that it was changed succesfully. Then I did offscreen capturing again, but the error still persists.

I have 16 gigabytes of ram and a 4GB 1050TI Gaming X Card. I am using savedepth and animated depth keyframes, and a low amount of blur. Demo was supposed to capture on 4096x2160 on 24fps.

My capture bat:
@echo off
Start quake3mme.exe +set fs_game "mme" +set mme_renderWidth "4096" +set mme_renderHeight "2160" +set fs_extragames "defragmaps" +set r_multisample "16" +set r_multisampleNvidia "1" +set r_anisotropy "16" +set r_picmip "0" +set cl_noprint "1" +set r_overBrightBits "0" +demolist "DEPTHMania.txt"

[ent]

I see the problem now.
The limit of mme_workMegs is 512, which is not enough by modern standards.
I will need to increase the limit. I don't know when I can get to the code to provide new builds, but stay tuned.
Even if you have a lot of RAM, the engine is an x86 program, so it can allocate total up to 2GB (4GB with special hints on my side).
You completely missed the point, if you are saving depth, then you are not using the new DOF system. Disable saving depth, and use it directly in regular shots.

[jiyorude]

I managed to get it to work by changing the resolution to 1920x1080. The new way of capturing DoF also gave me the same error when trying to render on 4K. Since it should be possible to output on 4K and I have seen other people do it, do you maybe think there is a different issue going on here? Do you think that increasing the ram usage to 4GB will solve the problem?

[ent]

Are you using motion blur?
Lowering amount of any blurring will help (motion, overlap, DOF).
And yeah, increasing the working megs even up to 1 GB will fix the issue for sure.

[jiyorude]

Thanks for your help on the ram usage. I managed to figure it out.
I have two more issues currently and I hope you are able to help me with those as well (I have a lot of issues i guess, lol).

1. I am trying to run Q3 CPMA + q3mme on Windows 10 through Apple Bootcamp on a 27" 5K iMac. I installed hi-res textures and High Quality quake. CPMA runs fine and I can play as much as I want. Whenever I want to open Q3MME, i get: "Windows encountered a error and had to close..." Sometimes it only opens the q3 console and then shuts off. Custom width and height are 1920x1080 and mme_renderwidth and height are 0 so that Q3MME should be forced to open. I even changed the display resolution from Windows to 1080p to see if that would fix it. I have also added the configuration files from my other pc. Nothing seemed to work. How can I succesfully run Q3MME on this machine? I am using this computer for work on a daily basis and I want to use this machine to capture demo's all day in the background while I am doing work.

2. I have captured a Capture The Flag demo with 12 contestants (me plus 11 bots) of 15 minutes. I want to do a visual experiment by rendering every single person's first person point of view experiment with it in editing. Although that would mean I have to render 12x 15 mins of gameplay including a few static camera povs to cut to, it is feasible enough and I can render it in 2-3 hours per run. I figured it out that it works by going to chase/chase and hold spacebar and scroll through pOVs with 'Q' and 'E'. However, although I am not using an extremely big map, I find that the point of view of other players often revert back to me because I am too far away from that particular player, and thus the data was not captured. How can i make sure that everyone's movements inside the map is captured consistently throughout the game?

[jiyorude]

It took longer than I wanted, but I figured out a way to run and batch capture demo's "natively" (not completely native of course) on Mac OS through Cross Over. Topic may be closed. If anyone is interested, I can make a tutorial and post it up here.