Mesh/Archive/2011-12-12
< Mesh
Jump to navigation
Jump to search
Agenda
- In the latest release and beta viewers (244864 and 245931) the mesh uploader requires BOTH that the number of materials in each LOD has to be the same AND that the names of the materials (the material attributes of <triangles> or <polys> tags) of lower LODs must be a subset of those of the high LOD. This is equivalent to simply requiring that the set of materials must be identical for all LODs. In that context, the error message (issued only in the beta viewer) that mentions subsets is not meaningful, and the error message in both about number of materials is inadequate. This suggests that the reversion (from an earlier beta) to reqiring the same number of materials may be a mistake. Please indicate if that is so or if the reversion was deliberate. [Drongle McMahon] (to Drongle and for Linden's consideration: please consider the material fix Runitai did back in July. Perhaps this is the supposed "reversion"? If it is so, then it should not be broken again. There are Maya users out there. Angel Alphaville).
- I do have specific questions -- a LOT of them -- and will no doubt continue to have them. I like to find whatever answers are already out there, before I come begging in the forum. I haven't found a good up to date central location for troubleshooting. Information is strewn all over. The answers to my questions are probably there somewhere, but I can't find them. Forum members may be able to answer many questions, but what is LL's role in supporting mesh? There is quite a lot of SL-specific information that one must know to upload anything with a decent PE, yet that information is not made easily available and accessible, or apparently kept updated. What plans does LL /the mesh team have for providing support for mesh problems? (Pamela Galli)
- The mesh uploader in the new beta has a bug in the "materials checkbox" and keeps showing the models in plain white, as opposed to showing them coloured when materials each had a different default colour. It also is in a bizarre way affected by the windlight settings that may be set at the moment. All this had appeared in the Shining branch of code and we gave the heads up for this at a meeting a month ago. The model is eventually imported right, but with plain white it is not easy to see the shading when we generate the normals and play with the values. Note that this is on an Nvidia GTX460, Win7-64 based system. [Angel Alphaville]
- I'm not able to browse for custom physic models anymore, no matter which viewer version I use. i tried different computers (iMac, MBP), different OS (Lion, windows 7 on boot camp), different networks (at my home in switzerland, and one in germany), wifi and ethernet. When I select "select from file" (or however it's called, I'm on a german version), it won't even let me browse for a file, the button stays greyed out. I already uploaded many models successfully with custom physics, now it won't even work for those. I already filed a JIRA (SH-2684), but it disappeared and I never heard again about it, last message was from Dan at 11/15, he asked me for a model, which i sent to him. He sent me the link to viewer version 245867, which was compiled for windows and linux, only, but that didn't work for me (on bootcamp, of course). I'd wish I could go on working with mesh... [Mikki Miles]
- Does the render weight code for sculpties use a hard-coded 2048 triangles? I ask because a 32 triangle sculpty (8x8 map) had about the same render weight as a 2048 triangle one the same size, which cannot be right (VWR-27827). Should it also take account of culling of degenerate triangles, which are the majority in sharp-edged LOD-proof sculpties? [Drongle McMahon] PS Mikki - try selecting "from file" twice.
- Any word on getting a floor for previewing rigged and offset mesh uploads? At the moment, I'm having to use a prim after an initial Aditi upload to figure out the offset value. This double-uploading takes extra time, and really doesn't need to. [Tiberious]
- The preview of physices shape on mesh upload dialog was changed from yellow wire frame? It is hard to check the relation between Object and Physics Shape.[Yuzuru Jewell]
Meeting Notes
Attendees List
- Bubbalicious (bubbalicious.chauveau)
- Mumbles (charlar.linden)
- Dan Linden (dan.linden)
- Drongle McMahon (drongle.mcmahon)
- Eleanora Newell (eleanora.newell)
- kikofelix98 (kikofelix98)
- Medhue Simoni (medhue.simoni)
- Mikki Miles (mikki.miles)
- Nal (nalates.urriah)
- Nyx Linden (nyx.linden) (User's Wiki Page)
- davep (runitai.linden)
- Koli Contepomi (sahkolihaa.contepomi)
- Second Life (second.life)
- Tiberious Neruda (tiberious.neruda)
- Vir Linden (vir.linden)
- Yuzuru (yuzuru.jewell)
Transcript for Monday December 12, 2011
[11:54] | Bubbalicious (bubbalicious.chauveau) | hello Charlar |
[11:54] | Mumbles (charlar.linden) | hey all! |
[11:54] | Koli Contepomi (sahkolihaa.contepomi) | Hey Mumbles. |
[11:55] | Medhue Simoni | hey Charlar |
[11:55] | Koli Contepomi (sahkolihaa.contepomi) | Hey Runitai. |
[11:56] | davep (runitai.linden) | brb, gotta fix speakers |
[11:56] | Koli Contepomi (sahkolihaa.contepomi) | Heh... |
[11:56] | Tiberious Neruda | hey Medhue |
[11:56] | Koli Contepomi (sahkolihaa.contepomi) | Welcome back. :o |
[11:56] | Medhue Simoni | Hey Tib |
[11:58] | Koli Contepomi (sahkolihaa.contepomi) | Hey Nyx. |
[11:59] | Eleanora Newell | hiya, nyx |
[11:59] | Tiberious Neruda | Nyx! |
[11:59] | Nyx Linden | heya |
[12:00] | Koli Contepomi (sahkolihaa.contepomi) | And, hey Dan. |
[12:00] | Mumbles (charlar.linden) | We'll start in just a minute - give a few more people a chance to get here |
[12:00] | Dan Linden | hello |
[12:00] | Bubbalicious (bubbalicious.chauveau) | hi |
[12:00] | Mumbles (charlar.linden) | We have a fairly large agenda, lots of wordy questions :-) |
[12:00] | Mumbles (charlar.linden) | https://wiki.secondlife.com/wiki/Content_Creation/Mesh_Import_User_Group |
[12:00] | davep (runitai.linden) | no kidding |
[12:01] | davep (runitai.linden) | we should make a 250 character limit or something |
[12:01] | davep (runitai.linden) | "from now on, agenda items must be tweets" |
[12:01] | Nal (nalates.urriah) | I guess whatever you did to get more items in agenda worked. |
[12:04] | Mumbles (charlar.linden) | OK, starting... NOW |
[12:05] | Mumbles (charlar.linden) | #1 In the latest release and beta viewers (244864 and 245931) the mesh uploader requires BOTH that the number of materials in each LOD has to be the same AND that the names of the materials (the material attributes of <triangles> or <polys> tags) of lower LODs must be a subset of those of the high LOD. This is equivalent to simply requiring that the set of materials must be identical for all LODs. In that context, the error message (issued only in the beta viewer) that mentions subsets is not meaningful, and the error message in both about number of materials is inadequate. This suggests that the reversion (from an earlier beta) to reqiring the same number of materials may be a mistake. Please indicate if that is so or if the reversion was deliberate. [Drongle McMahon] (to Drongle and for Linden's consideration: please consider the material fix Runitai did back in July. Perhaps this is the supposed "reversion"? If it is so, then it should not be broken again. There are Maya users out there. Angel Alphaville) |
[12:05] | davep (runitai.linden) | the "subset" requirement should be accurate -- it's OK if a lower LoD is missing a material that the higher LoD has |
[12:05] | davep (runitai.linden) | but the lower LoD MUST NOT contain a material that is not present in the higher LoD |
[12:06] | Medhue Simoni | Do i need text to speech today? |
[12:06] | davep (runitai.linden) | Drongle/Angel: is it requiring that the material count matches, or is it just enforcing the subset requirement? |
[12:06] | Drongle McMahon | That's what I thought, but it's giving the number of textures error as well. If it's not supposed to, I will do a jira. |
[12:07] | Tiberious Neruda | it's requiring both |
[12:07] | davep (runitai.linden) | sounds good, enter a jira and we'll pick it up during triage |
[12:07] | Drongle McMahon | It's insisting on both subset and same number. |
[12:07] | Mumbles (charlar.linden) | next... |
[12:07] | Mumbles (charlar.linden) | #2 I do have specific questions -- a LOT of them -- and will no doubt continue to have them. I like to find whatever answers are already out there, before I come begging in the forum. I haven't found a good up to date central location for troubleshooting. Information is strewn all over. The answers to my questions are probably there somewhere, but I can't find them. Forum members may be able to answer many questions, but what is LL's role in supporting mesh? There is quite a lot of SL-specific information that one must know to upload anything with a decent PE, yet that information is not made easily available and accessible, or apparently kept updated. What plans does LL /the mesh team have for providing support for mesh problems? (Pamela Galli) |
[12:07] | Tiberious Neruda | essentially, forcing usage of all materials on each LOD, as well as the physics mesh |
[12:07] | Tiberious Neruda | else, it throws up the MAV_BLOCK_MISSING error |
[12:08] | davep (runitai.linden) | yeah, that's busted -- physics mesh shouldn't need materials at all and lower lods shouldn't need to have every material from the higher lod |
[12:09] | Drongle McMahon | Physics is ok on the viewers I tested. |
[12:09] | Mumbles (charlar.linden) | We're aware that we don't have best practices documents up - we'd like to very much, but as I said in the forum response, we just don't have the time. I am going to raise the priority of doing something like this, and I hope that we can start sharing some of that in the next month. |
[12:10] | Drongle McMahon | It looks like the physics fix reverted to requiring lod meshes to have all the high lod materials. |
[12:10] | Mumbles (charlar.linden) | We'll only handle it from the collada on inwards. Dealing with the source programs is not something I'm comfortable with. |
[12:10] | davep (runitai.linden) | drongle: please include that info in the jira |
[12:10] | Drongle McMahon | ok |
[12:10] | Nal (nalates.urriah) | May be if you create the pages users can add content in the Wiki... |
[12:10] | Tiberious Neruda | that's acceptable, IMO. Might want to link to good modeler-specific info, though |
[12:11] | Mumbles (charlar.linden) | Nal - yes, I'm hoping for that. |
[12:11] | Drongle McMahon | There is tons of info in the forums. Just need somebody to extract it. |
[12:11] | Tiberious Neruda | (even if you guys don't WRITE it, just having it onhand will be a help) |
[12:11] | Nal (nalates.urriah) | Plus there are some great theads in the forum. We just need links collected where they can be found. |
[12:11] | Mumbles (charlar.linden) | OK teamwork FTW! |
[12:11] | Mumbles (charlar.linden) | next item... |
[12:12] | Mumbles (charlar.linden) | #3 The mesh uploader in the new beta has a bug in the "materials checkbox" and keeps showing the models in plain white, as opposed to showing them coloured when materials each had a different default colour. It also is in a bizarre way affected by the windlight settings that may be set at the moment. All this had appeared in the Shining branch of code and we gave the heads up for this at a meeting a month ago. The model is eventually imported right, but with plain white it is not easy to see the shading when we generate the normals and play with the values. Note that this is on an Nvidia GTX460, Win7-64 based system. [Angel Alphaville] |
[12:12] | davep (runitai.linden) | sounds like a regression we missed with the gl compatibility work |
[12:12] | davep (runitai.linden) | is there a jira on that? |
[12:12] | davep (runitai.linden) | (test content always appreciated) |
[12:13] | Vir Linden | #4 I'm not able to browse for custom physic models anymore, no matter which viewer version I use. i tried different computers (iMac, MBP), different OS (Lion, windows 7 on boot camp), different networks (at my home in switzerland, and one in germany), wifi and ethernet. When I select "select from file" (or however it's called, I'm on a german version), it won't even let me browse for a file, the button stays greyed out. I already uploaded many models successfully with custom physics, now it won't even work for those. I already filed a JIRA (SH-2684), but it disappeared and I never heard again about it, last message was from Dan at 11/15, he asked me for a model, which i sent to him. He sent me the link to viewer version 245867, which was compiled for windows and linux, only, but that didn't work for me (on bootcamp, of course). I'd wish I could go on working with mesh... [Mikki Miles] |
[12:14] | davep (runitai.linden) | I guess we'll follow up with QA on #3 |
[12:14] | Vir Linden | This should be fixed in shining fixes, will be in the next beta. |
[12:15] | Vir Linden | Do not think it's in 3.2.4 beta though. |
[12:15] | Mikki Miles | fingers crossed... |
[12:16] | Drongle McMahon | Mikki, did you try selecting "Load from File twicw"? That always worked for me. |
[12:16] | Drongle McMahon | *twice |
[12:16] | Mikki Miles | twice, trice, fource, manyce, nada |
[12:16] | Drongle McMahon | Hmm |
[12:16] | Dan Linden | I'm not aware of the bug mentioned in #3. I'll enter a bug. |
[12:17] | Mumbles (charlar.linden) | moving on... |
[12:17] | Dan Linden | "jira or it didn't happen" |
[12:17] | Mumbles (charlar.linden) | #5 Does the render weight code for sculpties use a hard-coded 2048 triangles? I ask because a 32 triangle sculpty (8x8 map) had about the same render weight as a 2048 triangle one the same size, which cannot be right (VWR-27827). Should it also take account of culling of degenerate triangles, which are the majority in sharp-edged LOD-proof sculpties? [Drongle McMahon] PS Mikki - try selecting "from file" twice. |
[12:17] | Mikki Miles | mh... |
[12:17] | Nyx Linden | we do assume a constant number of triangles in a sculpt, so that's a valid bug that should be fixed eventually, just ran out of time |
[12:18] | Nyx Linden | could you file a JIRA for this one and I'll throw it on the olde fashioned todo list? |
[12:18] | Drongle McMahon | I did already. |
[12:18] | Vir Linden | (btw on #4, try canceling out of the whole uploader dialog and restarting it - that fixed the browse missing problem when I tested it) |
[12:18] | kikofelix98 | brasileiros aqui ? |
[12:18] | Mikki Miles | did that, too |
[12:18] | Nyx Linden | ah right I see the jira ID now |
[12:18] | Drongle McMahon | vwr-27827 |
[12:18] | Nyx Linden | thanks! |
[12:19] | Nyx Linden | probably won't account for removal of degenerate triangles as that would be difficult to pre-compute |
[12:19] | Nyx Linden | but accounting for sculpt resolution should be possible |
[12:20] | Dan Linden | #3 is SH-2782 |
[12:20] | Medhue Simoni | Gotta go pick my son up, leaving my avatar here, can boot me when the meeting's over. |
[12:21] | Mumbles (charlar.linden) | ok, moving on. |
[12:21] | Mumbles (charlar.linden) | and afterwards, we can all takes pics with Medhue. |
[12:21] | Bubbalicious (bubbalicious.chauveau) | lol |
[12:21] | Mumbles (charlar.linden) | #6 Any word on getting a floor for previewing rigged and offset mesh uploads? At the moment, I'm having to use a prim after an initial Aditi upload to figure out the offset value. This double-uploading takes extra time, and really doesn't need to. [Tiberious] |
[12:23] | Tiberious Neruda | yeah, this one's me. I'd kinda thought the "What does this model represent" dropdown would do that, but... |
[12:23] | Vir Linden | That's not currently scheduled. If there's a JIRA we'll manage it on the backlog. |
[12:24] | Vir Linden | Other priorities right now. |
[12:25] | Mumbles (charlar.linden) | next one - and last one currently... |
[12:25] | Mumbles (charlar.linden) | *7 The preview of physices shape on mesh upload dialog was changed from yellow wire frame? It is hard to check the relation between Object and Physics Shape.[Yuzuru Jewell] |
[12:25] | davep (runitai.linden) | Might be a regression, I'll follow up with QA if there isn't already a jira |
[12:26] | Mumbles (charlar.linden) | that is all we have, folks... |
[12:26] | Yuzuru (yuzuru.jewell) | http://kanaemesh-e.blogspot.com/2011/12/mk103-make-mesh-object-by-kumiki-133.html please check. |
[12:27] | Yuzuru (yuzuru.jewell) | http://lh4.googleusercontent.com/-mRyk_i8HOEM/TuFHzil413I/AAAAAAAAJMM/ES41PXGMAtE/s940/MK103-20.png |
[12:28] | Yuzuru (yuzuru.jewell) | It is on beta viewer. |
[12:28] | Mumbles (charlar.linden) | hey, btw, as of yesterday 15.5% of all regions have mesh objects rezzed on them. |
[12:29] | Tiberious Neruda | well, maybe a question here... maybe not relevant yet, but... now that mesh is, at least partially, here, what's the next new shiny being worked on? Materials? |
[12:29] | Drongle McMahon | Any estimate of how many viewers can see them? |
[12:29] | Nal (nalates.urriah) | Neat. Any idea how many are using mesh render capable viewers? |
[12:29] | Nal (nalates.urriah) | lol |
[12:30] | Mumbles (charlar.linden) | I don't know, off hand, Drongle/Nal. |
[12:30] | Mumbles (charlar.linden) | You should ask Oz, but I will go and see what we can release, if anything. |
[12:30] | Drongle McMahon | I gathec Phoenix-with-mesh just came oput. |
[12:30] | Nal (nalates.urriah) | With Cool, Firestorm, Phoneix and a couple of others I am guessing we are over 50%. Oz tells me he can get that info but never says.... :/ |
[12:30] | Drongle McMahon | *gather |
[12:31] | Tiberious Neruda | yeah, Phoenix with mesh JUST launched |
[12:31] | Drongle McMahon | Hopefully that will qhosh the idea that mesh is inefficient. |
[12:32] | Mumbles (charlar.linden) | ok, it looks like we're wrapped here... |
[12:33] | Nal (nalates.urriah) | looks like... |
[12:33] | Eleanora Newell | hah, got back in time to say bye :) |
[12:33] | Mumbles (charlar.linden) | thanks all, good questions |
[12:34] | Eleanora Newell | minutes will be posted as always? |
[12:34] | Mikki Miles | thanks, and bye |
[12:34] | Mumbles (charlar.linden) | We're happy to start fielding more general content creation questions. |
[12:34] | Mumbles (charlar.linden) | cheers! |
[12:35] | davep (runitai.linden) | thanks folks |
[12:35] | Vir Linden | Bye all. |
[12:35] | Nal (nalates.urriah) | Did you do a blog post on that? I missed it if you did. |
[12:35] | Tiberious Neruda | hmm |
[12:35] | Nyx Linden | see everyone next week! |
[12:35] | Mumbles (charlar.linden) | damn |
[12:35] | Mumbles (charlar.linden) | Nal caught my slip |
[12:35] | Nal (nalates.urriah) | :) |
[12:35] | Bubbalicious (bubbalicious.chauveau) | bye |
[12:35] | Mumbles (charlar.linden) | I did not, sigh |
[12:35] | Tiberious Neruda | Nyx, I heard Materials was going to be next up after mesh... what's this entail? |
[12:35] | Mumbles (charlar.linden) | I'll do it now, with the extra time I have... |
[12:35] | Nal (nalates.urriah) | I'm blogging about it but didn't put anything on the forum... |
[12:36] | Nal (nalates.urriah) | \o/ |
[12:36] | Eleanora Newell | charlie, you got time? was that a xmas bonus? |
[12:36] | Tiberious Neruda | ...unless that's not being worked on yet |
[12:37] | Mumbles (charlar.linden) | haha eleanora |
[12:37] | Mumbles (charlar.linden) | it's a gift from you all |
[12:37] | Eleanora Newell | :) |
[12:37] | Mumbles (charlar.linden) | Materials aren't being worked on yet. |
[12:37] | Tiberious Neruda | darn |
[12:37] | Tiberious Neruda | ...but sign me up for testing when it shows :3 |
[12:37] | Mumbles (charlar.linden) | it's not up next, not neccessarily. |
[12:38] | Eleanora Newell | ok, we'll see you all next week, thank you :) |
[12:38] | Mumbles (charlar.linden) | off to the forums! |
[12:38] | Tiberious Neruda | I think it would help creators out |
[12:38] | Second Life | Teleport completed from http://maps.secondlife.com/secondlife/Borrowdale/87/208/33 |