Old Hair Complexity

[Ashke Navarathna]

I have some old  hair (like 10 years old) that I love, but when I put it on, my avatar complexity jumps by about 100k. It can be worn down or up, and I only ever wear it up. Is it the ability to wear it up what's causing it to be so complex? Is there a way to remove that ability and make it just up all the time? I've looked for newer updo hair like it, but can't find anything I like as much.

Thanks,

Ashke

[Ariel Vuissent]

You may be able to fiddle with it, if it's modifiable. I wouldn't, unless it's also copy-able, since if it's not and you make a mistake, it could easily become unwearable. 

If it's both, rez a copy on the ground, preferably in the style you prefer - the updo - if possible. ctrl+alt+t, I believe, will show transparent prims and whatnot. Edit, check the "Edit Linked" box, and choose one of those invisible prims. Unlink and delete it. Repeat until all unnecessary prims are gone. You can also rip out the scripts by going to the Contents tab and deleting them, although you'll have to isolate the script or scripts that affect the style change vs any other scripts that may be in the hair, including possible resizing scripts, if you still need them. The latter may not affect rendering cost, but can still affect the amount of lag you cause to yourself and others. 

You can also post an image of the hair here, and see if anyone can suggest similar styles that you may not have found. Older hair is not necessarily bad, but a number of things have improved in the last ten years...

Edited August 22, 2018 by Ariel Vuissent

[Ashke Navarathna]

I just checked. It's not modifiable. I should've checked that before posting. My bad. The complexity is 60k lower when the hair is down. Weird. I've attached 2 pics of front and side view. I love the color (main, and highlights on back), the way it frames my face, and the messy  bun look. Suggestions appreciated!

[Resi Pfeffer]

4 hours ago, Ashke Navarathna said:

and the messy  bun look. Suggestions appreciated!

Have a look at the inworld store of Monso (marketplace does not offer the whole variety). They have a bunch of messy and windswept hairs.

Edited August 22, 2018 by Resi Pfeffer
typo

[Marianne Little]

No.Match has very low complexity. I haven't seen render cost over 10 k, most is 3-6 k. They have some messy buns, take a look. https://www.flickr.com/groups/2755584@N20/pool/129183655@N02

And their hairs is mod.

In my opinion (and I know you didn't ask) your hair really need an update with more realistic looking messy hair strands from the bun.

Edited August 22, 2018 by Marianne Little

[Callum Meriman]

4 hours ago, Ashke Navarathna said:

The complexity is 60k lower when the hair is down.

 

I think it's likely the huge complexity is going to be those tortured prims that make up the twisted and tapered, curly/messy bits. (Your necklace looks like it might also suffer the same sort of torture, so if I may gently suggest to inspect that complexity too!)

All the best of luck!! I hope you can find a suitable new cut at those shops mentioned above (No Match is pretty great)!  

[ChinRey]

10 hours ago, Ashke Navarathna said:

I have some old  hair (like 10 years old) that I love, but when I put it on, my avatar complexity jumps by about 100k.

If there's any comfort, the actual complexity is likely to be considerably less than any fitted mesh hair.

Due to the fitted mesh LoD bug, the render complexity is way off for fitted mesh. If the render complexity calculation had worked the way it was intended, a fitmesh hair would typically have showed up with a render complexity somewhere between 5 and 300 times as high as it does. (More details here.)

My best advice is to just forget about that render complexity thing. It's broken beyond repair.

Edited August 22, 2018 by ChinRey
Did the math and corrected the numbers

[Fionalein]

10 minutes ago, ChinRey said:

My best advice is to jsut forget about that render complexity thing. It's broken beyond repair.

I wish you would tell that to the events that enforce it...

[ChinRey]

7 minutes ago, Fionalein said:

I wish you would tell that to the events that enforce it...

Wear a fitmesh hair at those events and a more sensible (and usually better looking) old style one when you're in the civilized parts of SL.

Or even better, wear a full body alpha and nothing else at those events. A completely invisible avatar with no attachment will have a calculated render complexity of either 0 or 1,000 (depending on how the alpha layer was made) and it has no scripts of course. Low enough to satisfy even the most agressive fake avatar lag monitor.

Edited August 22, 2018 by ChinRey

[Bree Giffen]

Could the hair be no mod because it contains no mod scripts? I forget how that works. I think either every prim has a script or there is one fairly complex script in the root prim. I'd suggest, if you can make a copy, to try rezzing it to the ground and try to edit it. If you can, then do as suggested above. I'd definitely remove those curly strands.

[Nalates Urriah]

17 hours ago, Ashke Navarathna said:

I have some old  hair (like 10 years old) that I love, but when I put it on, my avatar complexity jumps by about 100k. ... Is it the ability to wear it up what's causing it to be so complex? Is there a way to ...

This is a common problem that comes from the changing nature of SL. When the hair was made optimum design criteria was not what it is now. Today's standards try to discourage many of the things that were done years ago. The result is you see a high ACI on old stuff. Not all old stuff... Those things that didn't use generally good game-modeling optimization have high ACI.

To get an idea of what can be done, check Penny Patton's article: http://pennycow.blogspot.com/2017/03/draw-weight-revisited.html

[ChinRey]

2 hours ago, Nalates Urriah said:

This is a common problem that comes from the changing nature of SL. When the hair was made optimum design criteria was not what it is now. Today's standards try to discourage many of the things that were done years ago.

Sorry Nalates, but that is simply not correct. An old prim/sculpt hair will nearly always be less laggy than a modern fitted mesh one, no matter how many sculpts and toruses it's made from. The reason why a fitted mesh hair usually shows up with a lower render complexity is that the calculation is wrong. With fitted mesh, the render complexity value is essentially just an arbitary number and it has nothing whatsoever to do with how laggy that item actually is.

Edited August 22, 2018 by ChinRey

[Nalates Urriah]

1 hour ago, ChinRey said:

Sorry Nalates, but that is simply not correct. An old prim/sculpt hair will nearly always be less laggy than a modern fitted mesh one, no matter how many sculpts and toruses it's made from. The reason why a fitted mesh hair usually shows up with a lower render complexity is that the calculation is wrong. With fitted mesh, the render complexity value is essentially just an arbitary number and it has nothing whatsoever to do with how laggy that item actually is.

We'll have to disagree. ACI/ARC is a indicator of render/lag cost. So...

Mesh versus tortured prims is no contest. Mesh hair these days in general has an ACI or 1,000 to 20,000. Only a few of the  prim and/or sculpty hair items I've seen get that low. 

Some beautiful prim hair I haven't been able to trash yet has an ACI of 170,000... I have yet to see a mesh hair get that high an ACI. That is 170 times worse than my best mesh hair.

 Ashke has run into the problem many of us do with old prim hair having way high ACI. A primary cause is textures. But, prims when twisted/tortured can exhibit explosive polygon growth, which can be a major reason for increased ACI. You can see that by adding materials to a twisted prim so that it switches over to the 'mesh' land impact accounting. Toruses are the worst and for that reason are seldom used in new builds as mesh is way LI cheaper and ACI lower.

If the problem with Ashke's hair is textures, Penny's article may help.

[Ashke Navarathna]

Thanks to everyone for help and suggestions! I found a messy bun at No Match that I think will work when I want to put my hair up. It's not the same, but I'm changing the rest of my body (switched to mesh) so I guess it's time to change the hair!

[ChinRey]

21 minutes ago, Nalates Urriah said:

We'll have to disagree. ACI/ARC is a indicator of render/lag cost.

It is supposed to be an indicator of render lag/cost but the system is completely broken. For fittes mesh we're talking about a mragin of error of more than 99% here and that's ridiculous of course.

This is not something to disagree or agree about, this is a well proven and documented factL Fitted mesh is far laggier than what the render compelixty calculation claims. In some cases i can be several hudnred times as laggy as the number says. If you don't believe me, read these two threads:

https://community.secondlife.com/forums/topic/419469-rigged-mesh-lod-bug/

and

https://community.secondlife.com/forums/topic/426621-what-the-fitmesh-lod-bug-actually-means/

If you can find any errors in the tests and calculations in those threads, well, I have to apologize. But if you can't, accept it.

Edited August 22, 2018 by ChinRey

[Dementia Obviate]

1 hour ago, Ashke Navarathna said:

Thanks to everyone for help and suggestions! I found a messy bun at No Match that I think will work when I want to put my hair up. It's not the same, but I'm changing the rest of my body (switched to mesh) so I guess it's time to change the hair!

Don't forget to show us your new look.

[Theresa Tennyson]

10 hours ago, ChinRey said:

It is supposed to be an indicator of render lag/cost but the system is completely broken. For fittes mesh we're talking about a mragin of error of more than 99% here and that's ridiculous of course.

This is not something to disagree or agree about, this is a well proven and documented factL Fitted mesh is far laggier than what the render compelixty calculation claims. In some cases i can be several hudnred times as laggy as the number says. If you don't believe me, read these two threads:

https://community.secondlife.com/forums/topic/419469-rigged-mesh-lod-bug/

and

https://community.secondlife.com/forums/topic/426621-what-the-fitmesh-lod-bug-actually-means/

If you can find any errors in the tests and calculations in those threads, well, I have to apologize. But if you can't, accept it.

I seem to remember when the jellydoll feature first came out someone did some tests with three avatars.

In one case the three avatars were wearing nothing but the three of the same type of fitted mesh body.

In the second case the three avatars were wearing nothing but three different brands of fitted mesh bodies.

And in the third case the avatars were wearing nothing but flexi hairstyles and full body alphas.

Interestingly enough, with controlled testing the framerate of the scene with the three flexi hairs was worse than the framerate of the mesh body tests.

I also seem to remember that you were the one who did this test.

[Candice LittleBoots]

I don't care what happens in the engine-room. I'm upstairs on the sun deck soaking up the rays and wolf-whistles 

[ChinRey]

1 hour ago, Theresa Tennyson said:

I seem to remember when the jellydoll feature first came out someone did some tests with three avatars.

In one case the three avatars were wearing nothing but the three of the same type of fitted mesh body.

In the second case the three avatars were wearing nothing but three different brands of fitted mesh bodies.

And in the third case the avatars were wearing nothing but flexi hairstyles and full body alphas.

Interestingly enough, with controlled testing the framerate of the scene with the three flexi hairs was worse than the framerate of the mesh body tests.

I also seem to remember that you were the one who did this test.

I started a separate thread for all this to keep this thread as clean of all that boring techie stuff as possible. Is it ok by you if we move the discussion over there?

But since you ask. That was not a comparasion test between fitted mesh and flexi hair, it was a test how multiple identical mesh bodies "stacked" compared to multiple different ones. I can't remember why I added that flexihair test round too. I think I was going to do a similar "stacking" test for flexihairs too but it turned out I didn't have three identical ones to test. You can read it all here:

https://community.secondlife.com/forums/topic/375629-jelly-babies/?page=5&tab=comments#comment-1165320

I had already done the comparasion test between fitted mesh and flexi hair and posted it earlier in the same thread:

https://community.secondlife.com/forums/topic/375629-jelly-babies/?page=4&tab=comments#comment-1165072

In that test a fitted mesh body with a render weight of only 5,162 turned out to be noticeably laggier than a flexihair with a 75,801 render weight. I had forgotten that one and it's really worrying. I'll post a comment about it in the other thread.

[ChinRey]

1 hour ago, Candice LittleBoots said:

I don't care what happens in the engine-room. I'm upstairs on the sun deck soaking up the rays and wolf-whistles 

Oh. In my experience: use an avatar that looks vaguely female and you get as much wolf-whistling and attention from the guys as you can handle. SL males are easy that way.

[Ashke Navarathna]

This is the new messy bun look I picked up at No Match. It's called "No Common." 

 

[Nalates Urriah]

22 hours ago, ChinRey said:

It is supposed to be an indicator of render lag/cost but the system is completely broken. For fittes mesh we're talking about a mragin of error of more than 99% here and that's ridiculous of course.

This is not something to disagree or agree about, this is a well proven and documented factL Fitted mesh is far laggier than what the render compelixty calculation claims. In some cases i can be several hudnred times as laggy as the number says. If you don't believe me, read these two threads:

https://community.secondlife.com/forums/topic/419469-rigged-mesh-lod-bug/

and

https://community.secondlife.com/forums/topic/426621-what-the-fitmesh-lod-bug-actually-means/

If you can find any errors in the tests and calculations in those threads, well, I have to apologize. But if you can't, accept it.

I knew ACI/ARC was off/broken. I hadn't realised HOW far off it is.

I won't actually know how things compare until ARCTan is out. Then the community will re-start the discussion with a new set of numbers.

My bet is instead of changing a 30k ACI to 300k we'll see everything else go down... you know... more PC.

[ChinRey]

2 hours ago, Nalates Urriah said:

I knew ACI/ARC was off/broken. I hadn't realised HOW far off it is.

Nobody did until Optimo and Beq took a close look. What they found might actually only be the top of the iceberg, the more we look at it, the worse it gets.

I seriously can't see how LL is ever going to fix this but we can always hope.

It won't be soon though and until then: forget about this render complexity nonsense. It's just a number and it doesn't mean anything at all. If somebody complains about your render cost and that somebody is wearing any fitted mesh at all, tell them to mind their own business and stop bothering you.

The jellybean filter will kind'a work because it does derender avatars. All avatars add to the lag so getting rid of some of them will always help, even if it's the low lag ones that are derendered rather than the high lag ones. And of course, it does protect against worn graphics card crashers and for that reason alone, I wouldn't recommend anybody to siwtch it off completely.

[Alyona Su]

I believe the ARC score is a misnomer anyway. Personally speaking, my only concern about ARC is that I fall under 350k, which is the default jellydoll threshold (because I'd rather not be a jellydoll). If others complain about my ARC score then my answer is simply that their "non-imposter avatars" setting is too high. I simply recommend they set that number lower (like 6) then move away from me.