Fomenting Mutiny

[Legacy_Pstemarie]

               

OldTimeRadio wrote...

Now, I ask you, did I just single-handedly revolutionalize CTP Babylon by doing this?

If I didn't, what did I do?

Ummm...you just redid (or undid) some (if not all) of the work that OMB - the creator of Clean Models - did when he ran ALL of CTP Babylon through CM3 years ago prior to its final release. 

This behavior is exactly what I - and others - was describing in earlier posts and is why ANYONE using CM3 should read the CTP forums on Harvest Moon or contact OMB - or Bannor - directly.

OTR, we can't stop people from throwing caution to the wind, ignoring the advice of MANY others, but I certainly hope, for the sake of people that use CEP, TAD won't just throw in "fixes" willy-nilly without verifying their veracity. 
               
               

               

                     Modifié par Pstemarie, 01 février 2014 - 02:34 .
                     
                  

            

[Legacy_Lord Sullivan]

               There is no such thing as "215 000" errors or fixes to be made to a model. At the most a tipical single tile model
can have anywhere between "16" to "70" errors at the most if modeled carelessly including the modeling errors
by the modeling program because it can't read your mind.

I never used CM3 as I take the time to make sure my work is proper when done.
               
               

               

                     Modifié par Lord Sullivan, 01 février 2014 - 03:12 .
                     
                  

            

[Legacy_OldTimeRadio]

               Yeah, that's more or less what I think is going on, Pstmarie.  It's an easy trap to fall into, though, and I certainly fell into holes like that with a lot of tools and processes at one point or another.  I think the main thing is, when there is disagreement on a technical point, the best we can do is to provide some kind of empirical data to at least show why we believe what we do.  I think in cases where empirical data and reproducable tests are part of any dialog or, perhaps, disagreement, things are bound to head in a positive direction and work out just fine.

'>

In regards to using CM3, I don't use it very much either- but that's not because of any deficiency on its part.  Like Lord Sullivan, I try to avoid CleanModels by what I think of as embracing the "best practices" that the tool embodies.  I'm not going to lie, though: I once had a problem so vexing I did a Hail Mary pass and CM3 caught it and got the ball all the way into the endzone for me.  Since I usually tend to take the inability to solve a problem personally, CM3 kinda hurt my feelings that day.  But in a good way, lol.  Anyway, speaking of "best practices", the 20-page PDF that accompanies the program is worth downloading whether or not they actually use it.

There's a lot of wisdom in there.
               
               

               
            

[Legacy_Carcerian]

               So Clean Models is NOT a virus scan?

Admittedly I was also planning to run every possible mdl thru it on a project too, for the sake of optimization, I'm glad i read this post '>

If it ain't broke, don't fix it?
               
               

               

                     Modifié par Carcerian, 01 février 2014 - 05:52 .
                     
                  

            

[Legacy_OldTimeRadio]

               This whole discussion is pretty specific to the existing models in CEP.  The only reason I brought CTP Babylon into it was because I was certain the models in it had received a level of technical attention that was as good or better than any model CEP had in it.

This is simply not the case with most models out there, though.  So the same conclusions aren't going to apply. 
               
               

               

                     Modifié par OldTimeRadio, 01 février 2014 - 06:24 .
                     
                  

            

[Legacy_Carcerian]

               Well, d20 modern (The material in question) is a very similar situation to CEP, a lot of older content made by different artists of varying skill and ways of doing things. It was proposed to run content thru CM to optimize for a future PW project.
               
               

               

                     Modifié par Carcerian, 01 février 2014 - 06:43 .
                     
                  

            

[Legacy_Pstemarie]

               

OldTimeRadio wrote...

This whole discussion is pretty specific to the existing models in CEP.  The only reason I brought CTP Babylon into it was because I was certain the models in it had received a level of technical attention that was as good or better than any model CEP had in it.

This is simply not the case with most models out there, though.  So the same conclusions aren't going to apply. 

Yeah, CTP Babylon was painstakingly gone over with a fine-tooth comb by the CTP Team. Although I wasn't a member of the CTP Team, I hung around with them enough to be in the loop on its design and was one of the playtesters for Babylon.

Sully, is absolutely correct when he says that no model has that many errors to fix. The output from CM3 is solely dependent upon the quality of the model passed through it, the number of components making up the model, and the settings used for the program. Personally, I mainly use CM3 to fix lighting and shadow issues and change tileset heights. This being said, like OTR, CM3 has hit a homerun for me on more than one occaision - most recently when I passed Sigil through it and let CM3 fix the shadow issues. It saved me weeks of work.

Much of the discussion recently has turned to improving CEP content. CM3 is not going to magically make a low-poly, low quality model into something the same quality as the content being released in Project Q or by authors such as Six, LoW, Lord Sullivan, etc. All you'll have after a pass through CM3 is a low-poly, low-quality model that is HOPEFULLY free of glaring errors. If you truly want to fix CEP's older content and upgrade it to "modern" standards, the content needs to be rebuilt from the ground up. 
               
               

               

                     Modifié par Pstemarie, 01 février 2014 - 07:32 .
                     
                  

            

[Legacy_Zarathustra217]

               Yeah, naturally running things through CM3 won't magically improve the quality of the content, the idea is just to fix all those basic errors, such as bad pivoting causing shadow issues, unwelded vertexes and improper lighting settings. If you guys read back, I already mentioned having found 7000 of those (presumably some of them are insignificant, but I have noticed a lot that are). I've also found a large amount of other errors like duplicate faces, even duplicate meshes, as well as zero face meshes that prevents models from properly compiling.

I already mentioned though that it of course would need some testing and verification.

As for content overhaul, I also gave an example of that too. If you take your time to visit the wiki, you can also read the fairly narrow definition and terms of overhauls. And the idea is even that all overhauls should be approved by popular vote.

All in all, you should have nothing to worry about.
               
               

               

                     Modifié par Zarathustra217, 01 février 2014 - 09:31 .
                     
                  

            

[Legacy_Proleric]

               

OldTimeRadio wrote...

Zarathustra217 wrote... (13 hours ago)
Unless there's objections, I'm still hoping to do general fixes to the existing CEP by using CM3, but it'll take a bit more time than I had hoped due to various issues that arise. I will probably try to approach it per hak for now.

...I just wanted to go on record as generally objecting very strongly to any decompilation and mass-manipulation of existing CEP content without some really clear reason what, exactly, is going to be achieved by it and what assurances would be provided that errors or needless degradation isn't going to occur as a result of it...

From the context, am I right to assume that your objection only refers to the use of Clean Models?

Clearly, there are many other reasons for "decompilation and mass-manipulation", such as my proposed fix for Mounted Skeletons, which I trust is not a cause for alarm.

Regarding CleanModels, I have no technical opinion, but, knowing how sensitive the issue remains, I'd suggest that the wiki might be a more constructive format for building a consensus of the cognoscenti, as it allows issues to be broken out and resolved one by one.

Having said that, on the bigger picture of change in general, I'd support the principle that every change proposed on the wiki should have a clear objective, with some form of peer review / quality control on the outcome. We all recognise that CEP changes cannot be allowed to break things, so it needs a more formal approach than, say, CCC, in my humble opinion.
               
               

               
            

[Legacy_Estelindis]

               I've said it before, but, in the context of the present discussion, it seems to bear repeating.  One must use the right settings if one wants CM3 to fix the kinds of bugs that it can fix.  The best settings will vary from one type of model to another.  The pdf guide explains it all.  What this really means is that people who want to run CEP models through CM3 need to divide the models into groups.  I really think that such division would require individual examination of each model.  First, one would place the model down in the toolset, have a look at it and see if there are any obvious bugs.  Then, take a look at the file size to see if it corresponds to the size one expects for a model of that simplicity or complexity.  If it has no visible bugs but the file size is larger than one expects, it might benefit just from having CM3 weld its t-verts, for instance.  If bugs are identified, then import the model into 3dsMax or GMax, just to have a look, not necessarily to re-export.  Hopefully, that can show more clearly if there are bugs that should be fixed by hand or by CM3.

Anyway, I have purposefully refrained from giving a commitment to help with this kind of work, for the time being.  I have enough on my RL plate that, NWN-wise, I only really have time for the elven tileset at the moment.  When things get a bit freer, I would like to help with the fixes, but I don't know yet how soon that will be.
               
               

               
            

[Legacy_Pstemarie]

               

Zarathustra217 wrote...

Yeah, naturally running things through CM3 won't magically improve the quality of the content, the idea is just to fix all those basic errors, such as bad pivoting causing shadow issues, unwelded vertexes and improper lighting settings. If you guys read back, I already mentioned having found 7000 of those (presumably some of them are insignificant, but I have noticed a lot that are). I've also found a large amount of other errors like duplicate faces, even duplicate meshes, as well as zero face meshes that prevents models from properly compiling.

You missed the point. Why waste time running ANCIENT low-quality models through a program to fix "basic errors" when what they really need is to be remade and improved so that they do not standout when loaded into a scene with other, much higher quality, models? Its a colossal waste of time. As for improving them - frankly, a lot of what makes the earlier CEP models low-quality by today's standards are the poor quality, low resoultion textures used on them.
               
               

               

                     Modifié par Pstemarie, 01 février 2014 - 02:17 .
                     
                  

            

[Legacy_Zarathustra217]

               

Pstemarie wrote...

You missed the point. Why waste time running ANCIENT low-quality models through a program to fix "basic errors" when what they really need is to be remade and improved so that they do not standout when loaded into a scene with other, much higher quality, models? Its a colossal waste of time. As for improving them - frankly, a lot of what makes the earlier CEP models low-quality by today's standards are the poor quality, low resoultion textures used on them.

Not all the old CEP stuff are low models, most of the things having issues are otherwise quite fine. Overhauling and fixing errors are two seperate issues.
               
               

               
            

[Legacy_Zarathustra217]

               

Estelindis wrote...

I've said it before, but, in the context of the present discussion, it seems to bear repeating.  One must use the right settings if one wants CM3 to fix the kinds of bugs that it can fix.  The best settings will vary from one type of model to another.  The pdf guide explains it all.  What this really means is that people who want to run CEP models through CM3 need to divide the models into groups.  I really think that such division would require individual examination of each model.  First, one would place the model down in the toolset, have a look at it and see if there are any obvious bugs.  Then, take a look at the file size to see if it corresponds to the size one expects for a model of that simplicity or complexity.  If it has no visible bugs but the file size is larger than one expects, it might benefit just from having CM3 weld its t-verts, for instance.  If bugs are identified, then import the model into 3dsMax or GMax, just to have a look, not necessarily to re-export.  Hopefully, that can show more clearly if there are bugs that should be fixed by hand or by CM3.

It depends on the kind of issue you are hoping to get CM3 to fix. What I'm working on is just fixing basic welding, repivoting and lighting settings. I'm sure more issues could be fixed by more in-depth examination, but we are talking about tens of thousands of models here. It's either mass operation or nothing.
               
               

               

                     Modifié par Zarathustra217, 01 février 2014 - 03:31 .
                     
                  

            

[Legacy_TheOneBlackRider]

               I'm not sure, if this makes much sense, but I'd like to offer my help.
Point is, that I do have (too ?) manny kettles boiling - besides RL the PW(s) I'm involved in.
I have converted NWN2 placeable models but not with a complete retexture.
Anyways, if there are small (simple, cause I'm still no wiz like any of you out there) jobs, please contact me. I'd see, what I could get done.
               
               

               
            

[Legacy_OldTimeRadio]

               

Carcerian wrote...
Well, d20 modern (The material in question) is a very similar situation to CEP, a lot of older content made by different artists of varying skill and ways of doing things. It was proposed to run content thru CM to optimize for a future PW project.

Yeah yeah but there's a difference: Both the models in CEP and Babylon were "worked over" (by hand, in Max, or using CM3 or some other tool) in ASCII and then compiled.  Decompiling and then mass-manipulating them is not necessarily going to be a wise or flawless process without some specific fix in mind in the first place. 

Anyway, in any case where you're dealing with a rag-tag corpus of models which have, more or less, existed in ASCII and never been worked over by something like CM3, then that's when a tool like CM3 (or any other mass manipulation process) could potentially give the greatest benefit with the least likely harm done.  And, at least in some capacity, that's what I believe Acaos did when he was cleaning up the models from CEP 2.0 in preparation for CEP 2.1, and then moving forward.

The above is why decompiling models currently in CEP or CTP Babylon and re-using CM3 on them is not necessarily a good idea but why, with something like D20 which (AFAIK) has never been cleaned up, it could be.  Even then, though, CM3 is a tool to assist a competent modeler and isn't a substitute for one.

Proleric1 wrote...
From the context, am I right to assume that your objection only refers to the use of Clean Models?

Clearly, there are many other reasons for "decompilation and mass-manipulation", such as my proposed fix for Mounted Skeletons, which I trust is not a cause for alarm.

Not really specific to CM3.  CM3 is just the easiest one for people to get themselves in trouble with. 

CM3 is kind of like the sorcerer's hat in the Sorcerer's Apprentice.  Mickey snatches the sorcerer's hat at night and uses it to solve very specificific issues which make his life easier.  For the most part it works out well for him- though there are
unintended consequences on occasion.  But the real trouble starts once he gets lazy, falls asleep wearing the wizard's hat and, dreaming he is the wizard he is the apprentice to, waves his hands in ever grander jestures which cause all manner of havoc to occur.

+ =

I want to avoid a situation where people dump models en mass from CEP, run them through CM3 not really understanding what they're doing, and then turn around and dump those on The Amethyst Dragon and say "Here.  I fixed these 500 models for CEP!" and then have the burden on The Amethyst Dragon to figure out what, exactly, happened and make an evaluation about if some benefit was actually achieved.

I've already said quite a bit on the model health of the existing CEP content and unintentionally going down the rabbit hole with CM3 so I'll just kind of link to those instead of going on about it.

When it comes to any modification of prior CEP content, I have no problems with specific plans of limited scope to fix signifigant or glaring errors.  Frankly, I don't think anyone should even mention digging back into the prior CEP content unless they can make a reasonable petition to do so which meets the criteria of the three things I bolded above.  It just seems like a good, common-sense approach given the situation.  One that shifts the balance of the burden more from the Amehtyst Dragon to those desiring to make changes.
               
               

               

                     Modifié par OldTimeRadio, 01 février 2014 - 05:30 .