Game Research Forum
https://forum.xentax.com/
Good luck! You're doing the lords work here! Take all of the time you need!RandomTBush wrote: ↑Thu May 02, 2019 4:02 am I should have some time next week to try again with figuring out this rigging nonsense.
RTB said he may have had time to work on it this week. All we can do is hope for good progress! But as the famous Chef God once, said "You can have it now, or you can have it right"
Gordon?Tecosaurus wrote: ↑Fri May 10, 2019 2:34 pmRTB said he may have had time to work on it this week. All we can do is hope for good progress! But as the famous Chef God once, said "You can have it now, or you can have it right"
With a little patience, we'll have a top notch script ready for us to use in no time!
Nicely said, lol.Tecosaurus wrote: ↑Fri May 10, 2019 2:34 pmRTB said he may have had time to work on it this week. All we can do is hope for good progress! But as the famous Chef God once, said "You can have it now, or you can have it right"
With a little patience, we'll have a top notch script ready for us to use in no time!
That hare has really been through the ringer in terms of reuse. Started off in Game of Thrones, got reused for TWD3, then again for TWD4. I don't think telltale ever reused a model that much (maybe the raven from TWD got reused in TWAU and GOT??)RandomTBush wrote: ↑Sat May 11, 2019 12:13 pm *shrug*
Either way, I'm back to work on this finally. And thankfully I found a model that's pretty much identical between TWD3 and TWD4 (the hare), so that seems like the best starting point for determining how to fix this issue.
(EDIT: I'm definitely getting somewhere with this now! I've so far got two-bone weights reading correctly.)
Or, in scripting terms:"Why fix what isn't broken?" didn't apply to Telltale, it seems.
This was way too frustrating to figure out, so I'll grumble here to explain how this crap works.
First, you have to read all four weight bytes as a "long" value, and then break that apart into 2/10/10/10-bit binary segments.
Those are used for weights 2, 4, 3 and 2 respectively. Why is 2 listed twice? The upper 2 bits add an extra 0.125 each to the second weight's value (0.375 max).
And then the three sets of 10 bits each are the weights in descending order (#4 -> #3 -> #2), and need to be divided by 1023 (0x3FF) and then again for the following:
2nd = divide by 8 (0.125 max) + 0.125/0.25/0.375 from the upper bits, 3rd = divide by 3 (0.333 max), 4th = divide by 4 (0.25 max).
And finally weight #1 is the remainder, 1.0 minus #2, #3 and #4 combined.
In retrospect, I can see how this works... but what, exactly, was the problem with using float values for this sorta thing again???
Either way, thanks to that recycled hare model for being there, making me not want to rip out my hair.
Code: Select all
WeightVars = readlong f
Weight2 = (((bit.and (WeightVars) 0x3FF) as float / 1023) / 8) + ((bit.shift WeightVars -30) as float / 8)
Weight3 = ((bit.and (bit.shift WeightVars -10) 0x3FF) as float / 1023) / 3
Weight4 = ((bit.and (bit.shift WeightVars -20) 0x3FF) as float / 1023) / 4
Weight1 = (1 as float - Weight2 - Weight3 - Weight4)Nice!RandomTBush wrote: ↑Sat May 11, 2019 4:08 pm The rigging issue has finally been solved. I'll just copy-paste my comments from the script regarding it:Or, in scripting terms:"Why fix what isn't broken?" didn't apply to Telltale, it seems.
This was way too frustrating to figure out, so I'll grumble here to explain how this crap works.
First, you have to read all four weight bytes as a "long" value, and then break that apart into 2/10/10/10-bit binary segments.
Those are used for weights 2, 4, 3 and 2 respectively. Why is 2 listed twice? The upper 2 bits add an extra 0.125 each to the second weight's value (0.375 max).
And then the three sets of 10 bits each are the weights in descending order (#4 -> #3 -> #2), and need to be divided by 1023 (0x3FF) and then again for the following:
2nd = divide by 8 (0.125 max) + 0.125/0.25/0.375 from the upper bits, 3rd = divide by 3 (0.333 max), 4th = divide by 4 (0.25 max).
And finally weight #1 is the remainder, 1.0 minus #2, #3 and #4 combined.
In retrospect, I can see how this works... but what, exactly, was the problem with using float values for this sorta thing again???
Either way, thanks to that recycled hare model for being there, making me not want to rip out my hair.Now once I get static meshes working correctly, I can finally release the update! @_@Code: Select all
WeightVars = readlong f Weight2 = (((bit.and (WeightVars) 0x3FF) as float / 1023) / 8) + ((bit.shift WeightVars -30) as float / 8) Weight3 = ((bit.and (bit.shift WeightVars -10) 0x3FF) as float / 1023) / 3 Weight4 = ((bit.and (bit.shift WeightVars -20) 0x3FF) as float / 1023) / 4 Weight1 = (1 as float - Weight2 - Weight3 - Weight4)
Great! You are making very good processRandomTBush wrote: ↑Sat May 11, 2019 4:08 pm The rigging issue has finally been solved. I'll just copy-paste my comments from the script regarding it:Or, in scripting terms:"Why fix what isn't broken?" didn't apply to Telltale, it seems.
This was way too frustrating to figure out, so I'll grumble here to explain how this crap works.
First, you have to read all four weight bytes as a "long" value, and then break that apart into 2/10/10/10-bit binary segments.
Those are used for weights 2, 4, 3 and 2 respectively. Why is 2 listed twice? The upper 2 bits add an extra 0.125 each to the second weight's value (0.375 max).
And then the three sets of 10 bits each are the weights in descending order (#4 -> #3 -> #2), and need to be divided by 1023 (0x3FF) and then again for the following:
2nd = divide by 8 (0.125 max) + 0.125/0.25/0.375 from the upper bits, 3rd = divide by 3 (0.333 max), 4th = divide by 4 (0.25 max).
And finally weight #1 is the remainder, 1.0 minus #2, #3 and #4 combined.
In retrospect, I can see how this works... but what, exactly, was the problem with using float values for this sorta thing again???
Either way, thanks to that recycled hare model for being there, making me not want to rip out my hair.Now once I get static meshes working correctly, I can finally release the update! @_@Code: Select all
WeightVars = readlong f Weight2 = (((bit.and (WeightVars) 0x3FF) as float / 1023) / 8) + ((bit.shift WeightVars -30) as float / 8) Weight3 = ((bit.and (bit.shift WeightVars -10) 0x3FF) as float / 1023) / 3 Weight4 = ((bit.and (bit.shift WeightVars -20) 0x3FF) as float / 1023) / 4 Weight1 = (1 as float - Weight2 - Weight3 - Weight4)
