POLL: Tesseract vs. Cube 2

[qreeves]

As some of you may be aware, there's a new Cube-derived engine on the block called Tesseract. This new engine has many improvements over its predecessor, which I think is better explained by eihrul's description:

Tesseract provides a unique engine derived from Cube 2: Sauerbraten technology but with upgraded modern rendering techniques. The new rendering features include fully dynamic shadows, global illumination, HDR lighting, deferred shading, morphological/temporal/multisample anti-aliasing, and much more.

Failing to explain it with words, comparatively:


While you can see the benefits for yourself (go check it out if you haven't already!), it doesn't come without a price. The fancy new rendering does mean that older hardware may not be up to the challenge of running such a high end engine. Red Eclipse, however forward thinking, has reached the limits of what is possible within the current Cube 2 engine, and I believe it may be time for us to let go and move forward with Tesseract. Unfortunately, this is an all-or-nothing sort of deal; maps would need significant changes to use the new formats, and the new form of rendering is not compatible with the old.

This is basically your chance to have your say on the issue: Should we move to Tesseract, or would we lose too much in the process?

[ZeroKnight]

I think the biggest deciding factor is if we'll have enough people willing to help out with the conversion/upgrade process. I'd personally love to move on to a bigger, better and sexier engine so that Red Eclipse can further expand as it has been, but we might lose quite a lot of momentum in the process if we decide to switch, and we don't have enough people helping out.

To get things rolling, I'm up for the challenge (that is whenever I'm not being enslaved at my job), and I'm willing to learn as I go to help out, as always. I'd really love if we moved to Tesseract, but we'll need some dedicated help.

[Dratz-_C]

I know that Tesseract won't run at all on my computer. I would have to spend the remainder of my scholarship money for the fall semester to help pay for a newer computer to run it and save up the rest of the money myself. However, I also believe that you should continue to push forward with Tesseract. You will gain players who benefit from the change and lose players who can't run it so the net should be you will still have players. I think you should implement a compatibility mode that can run lightmaps, being just multitexturing. This way you can run old maps rather than immediately throwing them away. On this assertion, just make a deadline that is at least 6 months after you add Tesseract as the preferred rendering method after which new content must be submitted in the new format. Perhaps you have already done this since I remember your passing on Tesseract before because of the necessary map changes and system requirements. My vote: push forward to Tesseract. I will be happy to help in the conversion if I can get said hardware and wish you all the best with your portion of the work. [I've changed my vote to "no" because "Tesseract" seems, for the time being, to be unable to gracefully handle shading large maps at the quality that "Cube 2" already does.]
Cheers

[MandaPanda]

I like how the shadows and lightning are much more "crisper", I guess is the word I will use. I think the look of the environment is much better in Tesseract. Hopefully it's not too much work to get it all converted but I am really loving the look of it.

[greaserpirate]

I'm all for going forward with Tesserract as long as it is an optional upgrade.

Having both rendering pipelines on one program might be impossible, but you can certainly have two different clients playing the same map over the network, just Tesseract's maps will be modified versions.

Even though I won't be able to use it, I'm really looking forward to the seeing the Tesseract Engine develop and hopefully rise to the prestige of Unity or UDK. (I know from experience that making maps in the Cube Engine is way easier than in UDK, not to mention how much UDK crashes)

But please don't forget about us small folk. Low system requirements are one of RE's main features; people running low-end Mac and Linux computers are a huge pert of this game's demographic. For these people, RE has the market pretty much conquered. That's how I found Sauerbraten, and eventually RE, and from what I've heard from fellow players the story is very common.

And let's not forget, the Cube 2 engine still has the potential to make amazing-looking maps, like cphills, skycastle, and collusion.

[Fallen]

Having both rendering pipelines on one program might be impossible, but you can certainly have two different clients playing the same map over the network, just Tesseract's maps will be modified versions.

2 diffrent versions means twice more work to do. Maybe not right now, but updating will be painfull and will take much longer. This idea will work only while there will be 2 teams - one working on Cube2 Engine, and the second one working with Tesseract, but still - connecting both players means doubled number of bugs (if not more than doubled), also - I'm not sure if you can make excatly same map in both engines by converting or whathever - if not, then remaking the map will be painfull and sometimes there will be diffrence between thoose maps...
And ofc community maps... Community most likely will support only Cube2 or Tesseract - only some ppl will support both (as some ppl will not even want to download Cube2 version just to remake map which they've done with Tesseract...).

I think that it'll be much better to give up RedEclipse project with Cube2 Engine, and countinue it as RedEclipse: <enter_subtitle_here_or_whathever>

I'm fully about pushing RE with Tesseract, but as something diffrent - with keeping `classic` Cube2 RE master-server online to let ppl playing it.

[Ulukai]

My vote goes to Tesseract all the way as well and I will do everything what's in my power to help this project succeed. You can count on my to help making the necessary changes to the maps. When we'd start doing this, we should have some guidelines set out so that we know what w eshould do different than how we do it now. Maybe some maps will need to be abandoned, I don't know, but it might be a good opportunity to clean the list if necessary.

As quin mentioned that this will be an all or nothing operation, I think the chances are limited that 2 versions will / can be maintained at the same time. It would need lots of time to maintain the basic functions of the engine, and maps should be optimized to run on either one of these versions. I believe you can play Tess maps in RE and vice versa, but performance won't be optimized so we should choose only 1 way to go. I believe Tesseract will be a bright future ahead for RE. This amazing work done by eihrul could place RE on the radar for more players because it will get a more modern look. It would provide us quite some coverage and I'm sure our playerbase will benefit from this as well.

For the players with older hardware... I understand this might be less pleasant, but maybe Tess can be tweaked a bit as well to gain some performance. Anyway, there are plenty of Linux users with decent computers and solid GPUs who can play this as well, the argument that all Linux players play such titles because there are not much other alternatives is no longer a valid argument since the booming of Linux gaming lately with the Humble Bundles and Steam.

Finally, imo keeping the present version alive besides a Tess version is not a good idea. There is the risk of fragmenting the community and playerbase. All our noses should point in the same direction.

[Fallen]

For the players with older hardware... I understand this might be less pleasant, but maybe Tess can be tweaked a bit as well to gain some performance. Anyway, there are plenty of Linux users with decent computers and solid GPUs who can play this as well, the argument that all Linux players play such titles because there are not much other alternatives is no longer a valid argument since the booming of Linux gaming lately with the Humble Bundles and Steam.

There always will be a people with older hardware - but when they get a new hardware then they leave games which they've played only because they had older hardware.. Also... Most ppl with `older hardware` should be able to run Tess without any problems on minimal settings I think. People with antic hardware usually doesn't play games at all or stick to that same game..
And... do you suggest that Linux users always have worst hardware? It's a lie. Nowadays I think that avergee hardware of Linux users is much better than avergee hardware of Windows users (I do not compare Mac, 'coz it's totally a diffrent story...), also - don't forget Desura feeding us with games and not only - I'm a Linux gamer and I didn't just started playing games while HIB appeard - a long time earlier there were games such as ET: Quake Wars and ofc thousands of free titles, and as well half (if not more) of the games were running perfectly under Linux then (right now probably 80-90% of the games runs fine under wine)...

Keeping RE with Cube2 is to stay with ppl with older hardware - and that group is getting smaller and smaller day by day... There's no way to gain more players without move forward...

[Ulukai]

For the players with older hardware... I understand this might be less pleasant, but maybe Tess can be tweaked a bit as well to gain some performance. Anyway, there are plenty of Linux users with decent computers and solid GPUs who can play this as well, the argument that all Linux players play such titles because there are not much other alternatives is no longer a valid argument since the booming of Linux gaming lately with the Humble Bundles and Steam.

There always will be a people with older hardware - but when they get a new hardware then they leave games which they've played only because they had older hardware.. Also... Most ppl with `older hardware` should be able to run Tess without any problems on minimal settings I think. People with antic hardware usually doesn't play games at all or stick to that same game..
And... do you suggest that Linux users always have worst hardware? It's a lie. Nowadays I think that avergee hardware of Linux users is much better than avergee hardware of Windows users (I do not compare Mac, 'coz it's totally a diffrent story...), also - don't forget Desura feeding us with games and not only - I'm a Linux gamer and I didn't just started playing games while HIB appeard - a long time earlier there were games such as ET: Quake Wars and ofc thousands of free titles, and as well half (if not more) of the games were running perfectly under Linux then (right now probably 80-90% of the games runs fine under wine)...

Keeping RE with Cube2 is to stay with ppl with older hardware - and that group is getting smaller and smaller day by day... There's no way to gain more players without move forward...

I totally agree. I was not suggesting that Linux users use inferior hardware, as I run a powerful PC with a GTX 760 on Linux myself. I was referring to what greaserpirate said. But just like you say, there will always be players with older hardware, we have them these days in our community and we should at least keep their fate a little in mind. For the greater good though, I strongly believe that Tess will be the best move we can make.

[Calinou]

I say Tesseract. We need to move forward if we want to attract more players, not to mention it makes mapping easier and more user-friendly.

[TheLastProject]

Personally, although Tesseract looks nice, I'm fairly hesitant.

I know graphics are a plus, but is it even that important? Red Eclipse has always stood out due to its gameplay, and the gameplay is independent of the graphics. Although I would like seeing Red Eclipse look better, I am unsure what the performance difference will even be. I'll probably come back here with a final opinion if I get the chance of trying out Tesseract, so I can see just how much worse the performance would be, but I do wonder if it's worth it. Red Eclipse is a game which has always catered to the low-budget hardware folks, of which I believe a rather big part of our community is, and I'm not sure if we won't lose more members than we'll gain new members.

[Ulukai]

Personally, although Tesseract looks nice, I'm fairly hesitant.

I know graphics are a plus, but is it even that important? Red Eclipse has always stood out due to its gameplay, and the gameplay is independent of the graphics. Although I would like seeing Red Eclipse look better, I am unsure what the performance difference will even be. I'll probably come back here with a final opinion if I get the chance of trying out Tesseract, so I can see just how much worse the performance would be, but I do wonder if it's worth it. Red Eclipse is a game which has always catered to the low-budget hardware folks, of which I believe a rather big part of our community is, and I'm not sure if we won't lose more members than we'll gain new members.

When trying out the Tesseract renderer it should actually be done with an optimized map... but I don't know if there are many good test maps available for this. I think it is fairly important given the fact that we would probably do it for the official maps as well, and in the Tesseract readme the performance difference of good lighting usage is stressed as well:

To make shadowmapped point lights, just make light entities as normal. Optionally you can control the shadowing properties of the light using attribute 5: 0 = shadows, 1 = no shadows, 2 = static shadows - world and mapmodels only (no playermodels or pickups). Shadowmaps are cached from frame to frame if no dynamic entities such as playermodels are rendered into them, so static lights (2) will generally be much faster than normal lights (0) during gameplay and should be used where possible. Unshadowed lights (1) are even faster as they further reduce shading costs, so they should also be used where possible.

Source: https://github.com/lsalzman/tesseract

I don't think we would lose such a big deal of the community if you take a look at the results of the poll And attracting some more players would be very nice, I would love to play this game at any time of day and see that at least 3 or 4 servers are occupied, not only at night.

[Evropi]

Tesseract's closest match is Unreal Engine 3. The lighting looks absolutely amazing, and it's as well-optimised as any Cube 2 derived work has ever been. I've only tried it briefly on my local system, but what can I say, colour me impressed!

At the start of this year I would have made all sorts of amends to you to prevent something like this. That's because I had a godawfully outdated computer (Intel Celeron 330, GeForce 6800GT)! Not too dissimilar from TheLastProject, one might say. Now, I would have no qualms whatsoever.

I believe Tesseract adds several extensions to the map format, or so I heard. It is in some ways not directly compatible with Cube 2 maps.* I think that the move would annoy a great many people unless a migration tool for maps is in place (and this is very unlikely considering BSP is a binary format). If there is, I doubt it would be a big difficulty to get someone to manually curate all the maps in the game (hell, I'd try my hand); removing a bunch of lights here and there should not be an issue.

I'm not a graphics whoor, but the fact is that better graphics make me enjoy all games more, no matter what. I can't quite place why, but that's what is happening. I imagine the situation is not too dissimilar for others. I believe better graphics would draw in a lot more players. Let's put it this way; I would never have even looked at Red Eclipse if I didn't have a bad GPU back then. Same for the roguelike genre for instance. RE should strive to pull in players who are not just looking for games to conform to the technological limitations imposed upon them; for those people, Tesseract is massive step forward.

With that sole precondition filled, I can't see a reason not to migrate to Tesseract.

* I am going on LED's information here so don't shoot the messenger!

[TristamK]

Tesseract is a cool thing. But there are a lot of nuances. The most important of them - it's performance. Those who have tested it get 50% drops of performance (i find this information on the different sources). It is unacceptable. Players choice Re because they have a high fps on theirs machines. But if drops of performance will reduced to the 10-20% then migrate to Tes.
Also, I would be interested to know, how the transfer of old maps work? Would it be possible to just run the old map in editor with a calculated oldest light and simply replace it with a new one? (I ask this also because then I will need to remade all map from map pack)
Also, how much time can take recreating of engine and recreating of maps ? Because if it will take a more then 1.5...i think it's unacceptable too...
P.S. What about other upgrades ? What about changing Cube Script -> Lua ? :P

[greaserpirate]

Having both rendering pipelines on one program might be impossible, but you can certainly have two different clients playing the same map over the network, just Tesseract's maps will be modified versions.

2 diffrent versions means twice more work to do. Maybe not right now, but updating will be painfull and will take much longer. This idea will work only while there will be 2 teams - one working on Cube2 Engine, and the second one working with Tesseract, but still - connecting both players means doubled number of bugs (if not more than doubled), also - I'm not sure if you can make excatly same map in both engines by converting or whathever - if not, then remaking the map will be painfull and sometimes there will be diffrence between thoose maps...

I think that it'll be much better to give up RedEclipse project with Cube2 Engine, and countinue it as RedEclipse: <enter_subtitle_here_or_whathever>

Not true at all. In fact, that's currently how Tesserat works with Sauerbraten, and it works just fine. I've seen people on Sauer coopedit servers who use Tesseract alongside the normal Cube2 players, and nobody experienced any problems. It would actually take more work to have Tesseract use a different masterserver.

Keeping it this way would be an effortless way to avoid a lot of unnecessary suffering. You wouldn't need to fork the game off, you could keep the community a whole and you would avoid forcing a large portion of the playerbase that can't afford newer computers, including me and Dratz-C, from being able to participate in the community we love.

For the players with older hardware... I understand this might be less pleasant, but maybe Tess can be tweaked a bit as well to gain some performance.

I talked to Hirato and he said Tesseract's as optimized as it's going to be .

[Calinou]

Having both rendering pipelines on one program might be impossible, but you can certainly have two different clients playing the same map over the network, just Tesseract's maps will be modified versions.

2 diffrent versions means twice more work to do. Maybe not right now, but updating will be painfull and will take much longer. This idea will work only while there will be 2 teams - one working on Cube2 Engine, and the second one working with Tesseract, but still - connecting both players means doubled number of bugs (if not more than doubled), also - I'm not sure if you can make excatly same map in both engines by converting or whathever - if not, then remaking the map will be painfull and sometimes there will be diffrence between thoose maps...

I think that it'll be much better to give up RedEclipse project with Cube2 Engine, and countinue it as RedEclipse: <enter_subtitle_here_or_whathever>

Not true at all. In fact, that's currently how Tesserat works with Sauerbraten, and it works just fine. I've seen people on Sauer coopedit servers who use Tesseract alongside the normal Cube2 players, and nobody experienced any problems. It would actually take more work to have Tesseract use a different masterserver.

Keeping it this way would be an effortless way to avoid a lot of unnecessary suffering. You wouldn't need to fork the game off, you could keep the community a whole and you would avoid forcing a large portion of the playerbase that can't afford newer computers, including me and Dratz-C, from being able to participate in the community we love.

For the players with older hardware... I understand this might be less pleasant, but maybe Tess can be tweaked a bit as well to gain some performance.

I talked to Hirato and he said Tesseract's as optimized as it's going to be .

Not possible anymore. Tesseract uses its own map format, and has a changed protocol compared to Sauerbraten. It just turned into a standalone game...

[Sniper-Goth]

I'd go with tesseract too.
But let's hope that the usual crouch+stand in small places to get out of map can't be used in it too.

[greaserpirate]

I'd go with tesseract too.
But let's hope that the usual crouch+stand in small places to get out of map can't be used in it too.

Physics isn't changed, just rendering, so you'll still be able to haxor out of the map

Harmless glitches give a lot of games more character.

[ballist1c]

I say Tesseract. We need to move forward if we want to attract more players, not to mention it makes mapping easier and more user-friendly.

+1.

I'd like to be separated from the decaying Sauerbraten, and hopefully this switch will bring us some attention. Also, fully dynamic shadows? Dynamic shaders? Anti-aliasing? Hell yeah! I'd love to see what this brings to RE

[TristamK]

I talked to Hirato and he said Tesseract's as optimized as it's going to be .

Bad news.
Also. Tesseract work a bit strange for me (maybe i don't know how work with it, sure). I get some light errors. Also with more then 7 light ents performance drops very much for me. I don't know how it will work on maps with "closed type" like cargo or depot...
But i like how dynamic light work...

[Unnamed]

I took a look at Tesseract.
The shadows and lighting looks good. That's a big difference to the Cube 2 engine and with fully dynamic shadows you don't have to calculate lights while editing.
I play RE, because I like the gameplay, but better graphics also make the game more enjoyable and maybe attract some more players.


I've added 20 light ents in the map and my fps dropped about 40%. It is how in the readme file written: normal lights slow down the performence.
edit: I also added 20 bots. They were shooting at each other so there were around 100 bodies and my fps dropped to 20 - 30.
Normal lights shouldn't be used too often and there should be an option to turn off shadows.

[greaserpirate]

Not possible anymore. Tesseract uses its own map format, and has a changed protocol compared to Sauerbraten. It just turned into a standalone game...

Isn't this just eihrul consciously making Tesseract seperate? It was compatible before and working fine, surely we can make it that way again.

Shiny new engines are nice, but goddamn it, I don't want to have the community I've grown to love and feel at home in be swept away from under my feet.

[Calinou]

I'd go with tesseract too.
But let's hope that the usual crouch+stand in small places to get out of map can't be used in it too.

You can fix this by putting clip, but we're getting off-topic here.

[Sniper-Goth]

I'd go with tesseract too.
But let's hope that the usual crouch+stand in small places to get out of map can't be used in it too.

You can fix this by putting clip, but we're getting off-topic here.

No, even with clip you can pass through it.
You didn't saw what I saw through the maps.

Anyway, i need to test tesseract here to see if i will have too much fps drops.

[cdxbow]

Very happy for MekArcade to be used as the guinea pig, if that provides a way to reduce/assess the risk better. Outdoors probably suits dynamic lights better than indoors. Need those destructibles to show it off properly. I don't really have a good idea about the performance hit, bugs, incompatibilities or other negatives, but I have looked closely at 'modern' looking commercial games, and dynamic lights is a large part of recipe.