Jump to content

WTF is "all alpha hair" ???


You are about to reply to a thread that has been inactive for 2259 days.

Please take a moment to consider if this thread is worth bumping.

Recommended Posts

(I mentioned this in a post across the street in a thread about the evils of no-mod content. Apologies to anybody encountering it for the second time.)

Some creators are issuing no-mod hair and making a great deal about how it's "100% alpha blended textures" or "all alpha hair", etc., apparently believing that this means the hair cannot be left with Modify permission. This isn't just the new, possibly fly-by-night creators, but long-established brands that have enabled Mod perm for years and years, now reverting to no-mod ostensibly for reasons associated with this "all alpha" thing.

Ignoring for the nonce the disastrous effect blended alpha has on rendering frame rates, does anybody know what these creators are talking about?

There are well-known problems with alpha-sorting that I understand can be (partially?) addressed by sorting the vertices (of a single material?) within a single mesh. That's certainly an interesting discovery, but I don't know why that should affect what Mod perms to set -- it's not as if one can modify that sort order -- so this doesn't seem a reason.

These are all unrigged (to be scaled by script and moved around in the editor), so I don't think this has anything to do with a known alpha rendering problem with rigged attachments.

We all know about other alpha-rendering problems, too, that arise under different lighting conditions and viewer settings, but I don't see how any of those have anything to do with Mod permissions, either.

But before I call BS on all this: Does anybody know if these creators could be talking about something real?

Link to comment
Share on other sites

I've got a guess based on the fact that Stealthic was one of companies that was mentioned. A number of hair makers are coming out with hair that can be re-configured with "styling HUD's" that allow "styling" by showing or hiding various parts of the hair. I believe that Stealthic was one of the early adopters of this. Depending on how this is done the hair makers may be trying to avoid the system being banjaxed when the hair is modified, possibly disrupting link orders, etc. 

(I'm not saying this is a GOOD reason, mind you, just a POSSIBLE reason.)

  • Like 1
Link to comment
Share on other sites

This hair looks ok in Windows, but it's an abomination on Linux.

Serious abomination.

 

Edit: As for no-mod, all the men's cuts seem to be moving this way now. Something happened recently and they all flipped from allowing us to change them, to locking us out. I'd guess stealthic are trying to hide their particular textures and alpha settings so others can't copy the Linux faulty technique, although it's trivial to see for most people who really want to know.

Edited by Callum Meriman
Link to comment
Share on other sites

These styles do nothing but add unnecessary avatar complexity.

They include multiple styles all in the same linkset and allow one HUD to change parting, bangs, over the shoulder, over the chest options, windy or non-windy style.

I've seen several makers doing this now making linksets that rack up tri counts in the several hundred thousands, the worst I've seen being over 740k.

As far as the no mod mentality, it could also be avoided if creators would simply bother referencing OBJECT_CREATION_TIME for each link with the applier/HUD RX scripts which would also allow people to "trim the fat" if needed as well as consolidate their attachments into one linkset/outfit to save attach slots. I have yet to find hair, clothing or accessories capitalizing on it.

  • Like 2
Link to comment
Share on other sites

1 hour ago, Lucia Nightfire said:

These styles do nothing but add unnecessary avatar complexity.

They include multiple styles all in the same linkset and allow one HUD to change parting, bangs, over the shoulder, over the chest options, windy or non-windy style.

Oh goodie! Since there is a thread about the importance of low poly counts and optimization, that's their response to the issue LOL!

  • Thanks 1
Link to comment
Share on other sites

Yeah, Stealthic is one of the merchants, and the "styled" bits are indeed separate links, not distinct faces of a single mesh, so with Mod perm it would be possible to relink out of order or rename a link so the script couldn't so easily find it, etc., but I'm still getting the vibe that there's something else they're worried about, somehow associated with the alpha textures.

I notice the Dura hair is a single mesh with a single material face, but in contrast the Stealthic hair includes multiple links so in that case the vertex-ordering trick to defeat the blended alpha-ordering glitch shouldn't work (as I understand it anyway). I don't seem to see the alpha-ordering problem with the Stealthic hair, but that doesn't necessarily mean anything.

I'm not sure it's a clue, but the Stealthic instructions are adamant that Basic Shaders and Avatar Hardware Skinning must be set (to avoid the tragedy of a bad hair day or something).

Link to comment
Share on other sites

7 minutes ago, OptimoMaximo said:

Oh goodie! Since there is a thread about the importance of low poly counts and optimization, that's their response to the issue LOL!

5a7d7808c090b_Skjermbilde(1065).thumb.png.0f8c1501f11fda32e3f42792b68aa703.png

The polycount is for just one of the four styles in the linkset, they are all about the same. The hair takes advantage of a well known bug in the weight calculation system to appear much lighter than it is. It's the same trick mesh head makers use. (What'll happen to that JIRA btw? Does LL have the courage to fix it? @Whirly Fizzle, are you there?)

 

 

  • Like 1
Link to comment
Share on other sites

46 minutes ago, ChinRey said:

5a7d7808c090b_Skjermbilde(1065).thumb.png.0f8c1501f11fda32e3f42792b68aa703.png

The polycount is for just one of the four styles in the linkset, they are all about the same. The hair takes advantage of a well known bug in the weight calculation system to appear much lighter than it is. It's the same trick mesh head makers use. (What'll happen to that JIRA btw? Does LL have the courage to fix it? @Whirly Fizzle, are you there?)

 

 

 

I presume you mean BUG-11627 - Rigged meshes LOD swap based off of Avatar bounding box instead of element or linkset bounding box.

That problem means that rigged mesh tends to not have detailed lower LODs.
LL did try to change it but it broke too much content.

Note the last comment on that JIRA issue from a ProductEngine (LL outsourced developers):
 

Quote

State update: issue on hold until 'better content policies are made' due to lots of content missing lower lods.

 

  • Thanks 2
Link to comment
Share on other sites

18 minutes ago, Whirly Fizzle said:

I presume you mean BUG-11627 - Rigged meshes LOD swap based off of Avatar bounding box instead of element or linkset bounding box.

That problem means that rigged mesh tends to not have detailed lower LODs.
LL did try to change it but it broke too much content.:

Yes, that's only to be expected.

For those who don't have time to read all the technical details in the JIRA here's an illustration:

5a7d8724dc207_Skjermbilde(1067).png.39b1f3dae9aed9077e76c1b907e2a122.png

Those not quite flat transparent sheets are the four different styles of the hair I'm wearing. What happens is that the rezzed copy of the hair is shown at low LoD, while the one I'm wearing is at mid LoD. The reason for the difference is that the rezzed object has switch points determined the usual way by the object's actual size while a worn fitted mesh has switch points determined by the avatar's size. The calculated render weight and download weight are not affected by this though, they are always calculated by according to the LoD the item has had when rezzed. The result is that small fitted mesh objects, like heads and hairs show up with a much lower render cost than they actually have.

Edited by ChinRey
  • Like 3
  • Thanks 1
Link to comment
Share on other sites

As most mesh hair looks like Barbie let Ken apply  it with a trowel.. sticking to my little flexi 'tail (which I can wear under a helmet and looks fabulous) and has a teeny render hit looks like the right call. See you on the asphalt, Barbies.

Link to comment
Share on other sites

 

11 hours ago, Qie Niangao said:

the Stealthic instructions are adamant that Basic Shaders and Avatar Hardware Skinning must be set (to avoid the tragedy of a bad hair day or something).

Yey, let's all drop our quality to see their fluffy hair properly.

No thanks. Not reducing my quality, just like I won't run in CALWL as some head makers suggest, nor bump my LOD to 9 in debug as some mesh fakers suggest, nor .... clear my cache ... as a panacea for everything.

 

Edited by Callum Meriman
Link to comment
Share on other sites

42 minutes ago, Callum Meriman said:

 

Yey, let's all drop our quality to see their fluffy hair properly.

No thanks. Not reducing my quality, just like I won't run in CALWL as some head makers suggest, nor bump my LOD to 9 in debug as some mesh fakers suggest, nor .... clear my cache ... as a panacea for everything.

 

How would enabling basic shaders and hardware skinning drop your quality? Shaders are cumulative - higher levels of quality expect the lower levels to be turned on.

  • Like 1
Link to comment
Share on other sites

45 minutes ago, Theresa Tennyson said:

How would enabling basic shaders and hardware skinning drop your quality? Shaders are cumulative - higher levels of quality expect the lower levels to be turned on.

I guess it's more unticking advanced and running in basic... so the hair doesn't look like this:

d72134ac83e520fcb74b40665b83a7fe.thumb.png.42062dfa1606478ce9ffe1f0217e23d7.png

Link to comment
Share on other sites

20 hours ago, Whirly Fizzle said:

 

I presume you mean BUG-11627 - Rigged meshes LOD swap based off of Avatar bounding box instead of element or linkset bounding box.

That problem means that rigged mesh tends to not have detailed lower LODs.
LL did try to change it but it broke too much content.

Note the last comment on that JIRA issue from a ProductEngine (LL outsourced developers):
 

 

So does that mean they want us to depend on it never changing?  Then it should be documented as a feature.

  • Like 2
Link to comment
Share on other sites

5 hours ago, anna2358 said:

So does that mean they want us to depend on it never changing?  Then it should be documented as a feature.

Well, to sum it up, shortly after Quick Graphics with Jellybabes was introduced some clever merchants found a loophole that allowed them to fool the computers into believing their high lag meshes are low lag causing SL to be considerably heavier than it's supposed to be. This doesn't count as griefing though since they are making money from it so yes, it's unlikely that LL will do anything about it.

Edited by ChinRey
Link to comment
Share on other sites

18 hours ago, Callum Meriman said:

 

Yey, let's all drop our quality to see their fluffy hair properly.

No thanks. Not reducing my quality, just like I won't run in CALWL as some head makers suggest, nor bump my LOD to 9 in debug as some mesh fakers suggest, nor .... clear my cache ... as a panacea for everything.

 

Have you tried relogging?   (Sorry).

Seriously, I think all content creators, and certainly anyone who creates content using mesh components they've bought in, should test their products using the Official Viewer at default graphics and LOD settings, just to see what happens.      It's also instructive to test things at the lowest graphics settings, just to be aware of what things are going to look like to someone who logs in using a cheap laptop.   You don't have to make everything on the assumption everyone is going to be using low-level graphics, but it's good to know what happens if they do, so you can anticipate problems and work round them.

I've been able to debug several problems for friends that way, when they couldn't understand why lots of their customers were complaining that items weren't rendering properly.   

It's particularly a problem, I think, with LOD, since -- at least as I recall -- Firestorm's default LOD is higher than the LL default (and, therefore, the default in several other viewers, too), so many content creators never realise, until people start complaining,  that things that make that look good to them look dreadful to other users.

Test stuff in several viewers at several different settings levels, folks -- not all your customers use Firestorm, and neither do all of them have gaming-level graphics cards.

  • Like 4
Link to comment
Share on other sites

45 minutes ago, Innula Zenovka said:

Test stuff in several viewers at several different settings levels, folks -- not all your customers use Firestorm, and neither do all of them have gaming-level graphics cards

Or maybe just provide a demo. Then customers can take some responsibility for making sure what they are buying is what they want. 

Link to comment
Share on other sites

2 hours ago, Whirly Fizzle said:

Callum, that looks like this bug, which is specific to Mac & Linux.
BUG-8801 - Certain Mesh Hair with alpha triangles, Mac and Linux specific

Ugh.

Another reason to only buy after trying the demo. Not every creator is going to build for every computer, every OS, every setting configuration, every windlight, etc. 

Link to comment
Share on other sites

31 minutes ago, Pamela Galli said:

Or maybe just provide a demo. Then customers can take some responsibility for making sure what they are buying is what they want. 

Whenever possible, I like to be aware of potential issues well before things get to the demo stage.   

Yes,  people should certainly provide demos anyway, for all sorts of reasons, but I'm thinking particularly of a friend of mine who had no idea that the full-perms meshes she'd bought to use for a product she was making looked dreadful if you were using the Official Viewer at default settings.   It was only when I showed her a screenshot that she realised quite how bad the products on display in her shop looked for many of her customers.

It's not usually that difficult to make (or buy) components that at least look reasonable at lower graphics settings, even if you need higher graphics settings fully to appreciate them.    Or the creator can take a conscious decision to go ahead with selling something that a proportion of potential customers won't be able to see properly -- but that's then a conscious decision rather than a mistake someone blunders into.

 

Edited by Innula Zenovka
Link to comment
Share on other sites

1 minute ago, Innula Zenovka said:

Whenever possible, I like to be aware of potential issues well before things get to the demo stage.   

Yes,  people should certainly provide demos anyway, for all sorts of reasons, but I'm thinking particularly of a friend of mine who had no idea that the full-perms meshes she'd bought to use for a product she was making looked dreadful if you were using the Official Viewer at default settings.   It was only when I showed her a screenshot that she realised quite how bad the products on display in her shop looked for many of her customers.

 

You are making my point that buyers just need to examine demos in all possible use cases they want to consider. Because there is simply no way that you can be totally assured that a creator has done so, and there never will be. He might test all sorts of conditions but not Linux, and wind up with some bug associated only with Linux. The solution: demo. 

How do you propose to be aware of potential issues before things get to the demo stage?

Link to comment
Share on other sites

You are about to reply to a thread that has been inactive for 2259 days.

Please take a moment to consider if this thread is worth bumping.

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
 Share

×
×
  • Create New...