Jump to content

Recommended Posts

Posted

Yeah, I've heard the SSD's don't have quite the lifespan of their mechanical cousins.  It's nice for the OS drive, but still kinda pricey if you're looking at the 500gb -> 1TB range. 

 

What got me thinking about it was the time it took to transfer 200GB of data across my network to my storage drive.  Connecting to a non-encrypted mechanical backup drive across a GB switch interface, my current throughput is ~310Mb/s.*  Doing the math and assuming a 40% increase due to layer 2/3 overhead, best case scenario is two hours to transfer.  So while my GB switch was a nice upgrade, I'm still nowhere near its potential.

 

*May try a direct connection via Cat-6 between Desktop <--Switch Port--> Network Drive to make sure the problem isn't my existing wired in cable.

 

Using Windows 7 Robocopy command via a script, it took FIVE hours. :|  However, since a bottleneck would still exist on the network drive ( being mechanical ) I'll probably just get another mechanical drive to replace the desktop drive if it comes down to it. 

Since i can't afford having glas fibre everywhere, i went with bridged gb network. i'm not good at setting this up, as a friend told me that if i went into it, i could use one conn for up and one for down, but due to my limited  knowledge i use the software bridge with default settings. For now this does the trick for me and it works like a charm.

But as the write speed to the HD is one of your priorities, you could use one of these caching cards or use an SSD as cache for the mechanical HD in your server (needs a special controler as well, unless your server board already supports it in hardware (b/c software sucks when controlling streams of data]).

Or use a stripe raid. Gain is ofc not 100%, but high enough to notice it. Ya, not good for data integrity, but the cheapest solution.

 

But then again: Still can't wait for you to find some motivation again and (maybe) finally see the finished flame atronach animation. ^^

 

P.s.: if you go for an SSD solution, make sure you get one that supports TRiM in hardware. Those have a longer life time and a waaaayyyy lower error rate. Not to mention that they can be used on systems that don't have direct support for SSDs, like WinXP and some *nix distributions.

Posted

 

Yeah, I've heard the SSD's don't have quite the lifespan of their mechanical cousins.  It's nice for the OS drive, but still kinda pricey if you're looking at the 500gb -> 1TB range. 

 

What got me thinking about it was the time it took to transfer 200GB of data across my network to my storage drive.  Connecting to a non-encrypted mechanical backup drive across a GB switch interface, my current throughput is ~310Mb/s.*  Doing the math and assuming a 40% increase due to layer 2/3 overhead, best case scenario is two hours to transfer.  So while my GB switch was a nice upgrade, I'm still nowhere near its potential.

 

*May try a direct connection via Cat-6 between Desktop <--Switch Port--> Network Drive to make sure the problem isn't my existing wired in cable.

 

Using Windows 7 Robocopy command via a script, it took FIVE hours. :|  However, since a bottleneck would still exist on the network drive ( being mechanical ) I'll probably just get another mechanical drive to replace the desktop drive if it comes down to it. 

Since i can't afford having glas fibre everywhere, i went with bridged gb network. i'm not good at setting this up, as a friend told me that if i went into it, i could use one conn for up and one for down, but due to my limited  knowledge i use the software bridge with default settings. For now this does the trick for me and it works like a charm.

But as the write speed to the HD is one of your priorities, you could use one of these caching cards or use an SSD as cache for the mechanical HD in your server (needs a special controler as well, unless your server board already supports it in hardware (b/c software sucks when controlling streams of data]).

Or use a stripe raid. Gain is ofc not 100%, but high enough to notice it. Ya, not good for data integrity, but the cheapest solution.

 

But then again: Still can't wait for you to find some motivation again and (maybe) finally see the finished flame atronach animation. ^^

 

P.s.: if you go for an SSD solution, make sure you get one that supports TRiM in hardware. Those have a longer life time and a waaaayyyy lower error rate. Not to mention that they can be used on systems that don't have direct support for SSDs, like WinXP and some *nix distributions.

 

 

Going to test it again tonight.  I have two systems that are SSD equipped, so I'm going to load up one big ass file ~10gb or so and transfer it between systems across the gigabit switch and watch the throughput.  That should give me pretty much the fastest possible transfer rate I can have ( in theory ) without going all Raid crazy which will rule out the switch and eth cabling. 

 

 

Today I have been forcing myself to practice rigging in Max.  While I loathe it, building the rig from scratch, skinning and weighting it are kind of necessary skills if I ever want to consider customizing / tweaking the meshes and skeletons in Skyrim. 

 

 

Regrius:  The binary file is pretty much the same scene just choosing the binary format vs Ascii prior to exporting it.  If you wish, I can build another of the exact same scene and send them along for comparison. 

 

Posted

Going to test it again tonight.  I have two systems that are SSD equipped, so I'm going to load up one big ass file ~10gb or so and transfer it between systems across the gigabit switch and watch the throughput.  That should give me pretty much the fastest possible transfer rate I can have ( in theory ) without going all Raid crazy which will rule out the switch and eth cabling.

i only meant a raid for the drive that's needs the biggest capacity and inside the server. Everything that runs in stripe will make each connected drive last longer and make the overall read/write process faster. So it can be M2, SSD, mech. HD or the new SATA that has drives that are just boxes, running 2 M2 in stripe, whereby 2 10k RPM SAS in Stripe mode are ofc the best and most expensive solution, aside from 4TB and bigger SSDs.

 

Today I have been forcing myself to practice rigging in Max.  While I loathe it, building the rig from scratch, skinning and weighting it are kind of necessary skills if I ever want to consider customizing / tweaking the meshes and skeletons in Skyrim.

Rigging was about applying the skin to a skeleton, right? Even if i'm wrong, please tell me i'm right so i sound at least a little bit like i know what i'm talking about here. :(

Posted

 

Going to test it again tonight.  I have two systems that are SSD equipped, so I'm going to load up one big ass file ~10gb or so and transfer it between systems across the gigabit switch and watch the throughput.  That should give me pretty much the fastest possible transfer rate I can have ( in theory ) without going all Raid crazy which will rule out the switch and eth cabling.

i only meant a raid for the drive that's needs the biggest capacity and inside the server. Everything that runs in stripe will make each connected drive last longer and make the overall read/write process faster. So it can be M2, SSD, mech. HD or the new SATA that has drives that are just boxes, running 2 M2 in stripe, whereby 2 10k RPM SAS in Stripe mode are ofc the best and most expensive solution, aside from 4TB and bigger SSDs.

 

Today I have been forcing myself to practice rigging in Max.  While I loathe it, building the rig from scratch, skinning and weighting it are kind of necessary skills if I ever want to consider customizing / tweaking the meshes and skeletons in Skyrim.

Rigging was about applying the skin to a skeleton, right? Even if i'm wrong, please tell me i'm right so i sound at least a little bit like i know what i'm talking about here. :(

 

 

Okay, test complete. 

 

10GB file transfer from SSD <-- GB Switch --> SSD was about what I expected it to be.  842mb/s transfer if my math is right.  Took 95 seconds start to finish.  ( 80gbits / 95sec = 842mb/s )

 

So, either Windows was slowing it down with the file queries ( likely ) or the NAS box was slowing it down doing the same ( or both ). 

 

I can always do a Raid setup with a pair of Raptor drives, but I do have this thing about quiet and a pair of 10k rpm drives don't typically contribute to that.  Here is how bad it is . . . . lol

 

I have an enterprise class switch ( Cisco 2970 ) that has a pair of fans that sound like it's going to hover in place and fly off when they spin up.  Great for data-center deployment, not so great in my closet.  I nipped that in the bud by dropping by Fry's and grabbing a huge case fan. Pulled the tiny bazillion rpm fans out, templated and cut a hole in the top of the switch case and mounted the case fan to it.  Runs dead silent and still moves more air through the switch than the fans it came with could ever dream of. :D

 

***************

 

Rigging:  You're mostly right.  From my experience, there are three factors that make or break the way a model moves when you animate the underlying bone structure.

 

1)  Mesh topology.  This is more modeling than rigging, but if your mesh is destined for animation, then you really need to put some thought into where the edge loops are and how the topology is laid out. Screw this up and even a miracle in the next two phases won't fix it. 

 

2)  Bone placement. Got your Blue Ribbon winning topology all setup and ready to go.  Great.  Obstacle number two is making sure the bones are placed in the right spots so the joints conform to that amazing topology you spent so much damn time tweaking.  You don't line the bones up correctly and the model will distort in ways you never thought possible.  ( My biggest issues are normally the hips and shoulders )

 

3)  Weighting.  Your super awesome topology meets equally amazing bone setup.  They just don't MAKE a cape big enough for you  . . . . . lol.   The last critical step is linking the mesh to the bones. . . also known as skinning.   Once done the application will apply default weighting to the mesh determining the strength of Bone X to Mesh Section Y.  At this point you get to spend a few hours tweaking this to make sure the weights are accurate to ensure your mesh deforms properly when you move it.  Left to defaults, you can get crazy deformations if the weights are too weak or too strong for any given bone.  Eg, why is the ass deforming when I move the Spine X bone ?   Or why does the inner thighs collapse when I bend the character at the hip ?

 

 

Ugh. . . . never ending issues all the way around.  :|

 

I swear to be a 3D artist, you need to have a PH.D in computer science in addition to strong artisan skills. 

 

Posted

Restored my fun and social meters by watching the entire RE series, social bunny went away.

 

I'm by no means an animator, but I always got hung up on the knees, never been able to get one to deform to my liking. Shoulder/Arm-Sockets are the same thing, only 10x worse. An actual human shoulder has lots of different bones, soft tissues, straps, and muscles that give it multiple orient points, yet every single armature I've ever seen just has an anchor for the arm-root, and parents it to the spine, which completely fails to describe how everything in between is supposed to orient.

The whole idea of bone-based armatures only really works for hardbodied stuff like robots and insects, IMHO they just don't belong with softbodied things like humans, we need another way to describe softbodied movement, and finally there's some work being done again with procedural generation that may eventually give it to us. I think the real problem is we're using straight spatial math on stuff that doesn't have a straight spatial relationship to the control object.

 

I'm not sure I'll need exactly matching text/bin files, I'm just going to assume the last bit of the files are illegible junk and drop them. Real thing to decide is whether I want to actually interpret the fbx data into higher level fbx objects, or just grab the pieces I need for animation and run. Quicker to just smash and grab, but without an interpreter, it couldn't be put back into an fbx for export, and also wouldn't be expandable to handle things like meshes, textures, etc. Only plan I had for expanding it was to allow direct translation of skeletons to fbx, in case the base ones were intentionally screwy somehow, and something was getting lost on the way through nif.

Posted

Okay, test complete. 

 

10GB file transfer from SSD <-- GB Switch --> SSD was about what I expected it to be.  842mb/s transfer if my math is right.  Took 95 seconds start to finish.  ( 80gbits / 95sec = 842mb/s )

 

So, either Windows was slowing it down with the file queries ( likely ) or the NAS box was slowing it down doing the same ( or both ).

i'm running FreeNAS and below 4GB of RAM i should not even try to copy files bigger 2GB. So i upgraded to 16GB and e'voila: Nice transfer rates.

i could have quad glas fibre, but with not enough RAM in the server that would have been a waste.

 

I have an enterprise class switch ( Cisco 2970 ) that has a pair of fans that sound like it's going to hover in place and fly off when they spin up.  Great for data-center deployment, not so great in my closet.  I nipped that in the bud by dropping by Fry's and grabbing a huge case fan. Pulled the tiny bazillion rpm fans out, templated and cut a hole in the top of the switch case and mounted the case fan to it.  Runs dead silent and still moves more air through the switch than the fans it came with could ever dream of. :D

i'm not a big Cisco fan (See what i did there? ^^) and this has nothig to do with Snowden as i avoid Cisco since 1995, so i can only say: Wow.

 

Ugh. . . . never ending issues all the way around.  :|

 

I swear to be a 3D artist, you need to have a PH.D in computer science in addition to strong artisan skills.

100% agreed. i tried and failed epic, because all that was going on in my brain, looking at what Blennder showed me, was: D'oh!

Posted

 

 

100% agreed. i tried and failed epic, because all that was going on in my brain, looking at what Blennder showed me, was: D'oh!

 

you just need a good tutorial on how to setup the UI to suit your needs. Once you got that part right, you can start on remembering the blender short cuts keys once you know that you already got 1/2 way done. The learning curve is actually knowing what what does. i have a print out of the short cut keys in front of my monitor so i can quickly see what the key combinations are if you use those keys enough time you hardly need the paper short cuts since you already memorized them then everything becomes easier to work with.

Posted

i'm not a big Cisco fan (See what i did there? ^^) and this has nothig to do with Snowden as i avoid Cisco since 1995, so i can only say: Wow

Cisco is a pain in the ass to put it nicely, but while complicated and anything but user friendly, they are extremely powerful and flexible enough to do anything you want. I run a Cisco 871w router along with the 2970 switch so it forces me to learn the Cisco IOS if I want to keep my home network running :) But then I help manage a network of roughly ~15k Cisco devices during the day, so two more isn't much to deal with lol

 

 

Blender. I have it installed, but, like many, dislike the UI. I started playing with this mess many years ago with Lightwave. Then Softimage and now Max. ( May as well toss Rhino and Zbrush in there too )

 

I prefer Softimage as that one is the one I am most accustomed to, but since Autodesk bought ( and is now killing it off ) I have to switch gears I guess. I'm still not comfortable with Max modeling tools, but otherwise it does what I need it to.

 

One of my big issues with Blender is that it doesn't support a 3D mouse very well. It does support them, but it behaves differently than it does in the other apps. Just enough to drive me crazy :/ So I use it to open and look at various blend models to see how folks approached them, but that's about it.

 

The other issue with Blender is the lack of training available *compared to the other apps*. It's out there, just not as much of it. Need to talk Digital Tutors or Gnomon Workshop into producing some online video training for it if they truly want folks to be serious about it.

Posted

As much as i would like to learn 3d modeling and appreciate the help that can be found here and everywhere else on the web, i still decided to stick with what i do best.

"A cobbler should stick to his last." is my motto here, as i have difficulties learning new stuff with every year that passes.

There are always people i can ask to help me out with stuff i can't do and they can ofc ask me, if what they need is covered by my skills.

Yes, i really put it as simple as that and it helps getting to know new people too.

 

But enough of that off-topic chatter, because i still have a hard time waiting for your flame atronach animation and looking in this thread does not help killing my anticipation. :)

Posted

Awesome Dayelyte, I see you've started to create animations for the Netch, how's the progress? :)

 

Chuckle.  Terrible :D    Motivation has been somewhat lacking in the animating department as of late.  Both the Atronach and the Netch will require the Havok Tools Exporter vs the usual .nif one.  Though, I'm beginning to doubt any of them will make it into SexLab :P

 

Was playing around with custom skeleton and weight painting past couple of days.  Frustrating would be the word to use I think. 

 

 

As much as i would like to learn 3d modeling and appreciate the help that can be found here and everywhere else on the web, i still decided to stick with what i do best.

"A cobbler should stick to his last." is my motto here, as i have difficulties learning new stuff with every year that passes.

There are always people i can ask to help me out with stuff i can't do and they can ofc ask me, if what they need is covered by my skills.

Yes, i really put it as simple as that and it helps getting to know new people too.

 

But enough of that off-topic chatter, because i still have a hard time waiting for your flame atronach animation and looking in this thread does not help killing my anticipation. :)

 

 

Trust me when I say I fully understand the " I have difficulties learning new stuff with every year that passes " part :D

Posted

Your attention to detail and thoroughness is fantastic which excels above even some of the most famous custom animators for skyrim. You have my giant respect and I feel as though I am in some sort of debt to you, you rekindled my hope for one day, perhaps other people may make more female beasts with the same level of attention to detail as the many male ones. Perhaps, I may even see my distant curiosity and fantasy for a remodeled female riekling becomes a reality, get my desired love in a pocket one day, regardless hwo lng it may be, it most certainly seem more likely than impossible now.  it's a long shot, but your work just brought forth that same wondrous fascination when the first time I saw marky's riekling on his comic. 

I like what you're doing with the werewolf, its so appropriately risque (in LL standard lol) and you really have quite an understanding in applying character in your animation. 

Not a chicken fan, but i blew my sides at the sight and prospect of it, would still definitely add it in my game for kicks.

Posted

Love your animations Dayelyte. Im curious; will the animations for the female werewolf be updated to work for 1.57? I know a couple people have probably already asked.

Posted

I like this mod looks good but I can't get it working :(  I use the NMM and when I try to install the testing part it says something about replacement in the animation list. I have the matchmaker but when I use it on the werewolf it says that the spell didn't work. I tried to reset the animation registry and still nothing happens :( Can you help? Please :)

Posted

Dayelyte,is you werewolf animation ready for sexlab 1.50+ ?

 

The animations themselves are only that, just animations.  They are immune to version upgrades and work equally well with any version of SexLab.  Now that is somewhat of a misleading statement though, so let me explain.  For the moment, I only create the animation file.  There is some external scripting that is required in order to get the animation itself registered with SexLab.  That scripting is what usualy has to be updated as Ashal modifies / changes how his code works.  The animation files will remain exactly the same, it's just the scripting that doesn't play nice between the major SexLab releases.

 

 

Your attention to detail and thoroughness is fantastic which excels above even some of the most famous custom animators for skyrim. You have my giant respect and I feel as though I am in some sort of debt to you, you rekindled my hope for one day, perhaps other people may make more female beasts with the same level of attention to detail as the many male ones. Perhaps, I may even see my distant curiosity and fantasy for a remodeled female riekling becomes a reality, get my desired love in a pocket one day, regardless hwo lng it may be, it most certainly seem more likely than impossible now.  it's a long shot, but your work just brought forth that same wondrous fascination when the first time I saw marky's riekling on his comic. 

I like what you're doing with the werewolf, its so appropriately risque (in LL standard lol) and you really have quite an understanding in applying character in your animation. 

Not a chicken fan, but i blew my sides at the sight and prospect of it, would still definitely add it in my game for kicks.

 

Thanks for the thoughts :D   Like most, I'm never fully satisfied with the animations as there is always somethiing else that can be tweaked, moved or modified.  Though, at some point you have to call it done lest you never finish.  I decided on the female beasts since Panic was doing the male variants and we already had other animators doing the standard human male / female animations. 

 

 

Love your animations Dayelyte. Im curious; will the animations for the female werewolf be updated to work for 1.57? I know a couple people have probably already asked.

 

They have, and like I stated above, the animations themselves are viable regardless of the version of SexLab ( they don't even require SexLab actually, any mod can make use of them ), however the scritping required to drive them will need to be created / updated by the mod-makers in order to run them within SexLab.

 

 

I like this mod looks good but I can't get it working :(  I use the NMM and when I try to install the testing part it says something about replacement in the animation list. I have the matchmaker but when I use it on the werewolf it says that the spell didn't work. I tried to reset the animation registry and still nothing happens :( Can you help? Please :)

 

The files currently in the download section are just the animations,  they lack the scritping necessary to implement them into the SexLab engine at this time.  As I don't wish to cause any oddball issues with custom versions of Ashal's scripts, I have pulled all scripting and behavior files and only the raw animation files remain at this time. 

 

That being said, the only way the animations work in game at the moment is if Ashal includes them in his mod ( SexLab ) or if another mod-maker includes them within their mod.  My apologies for that. 

Posted

Refactoring ate my property reader, but in rebuilding it, got it to read bools properly, and inflate embedded compressed data, or I think it will when it's not in pieces at least. Needs more refactoring, but slowly getting into proper shape. C# is funny like that, if you get the OO part right, it's quick, easy, and intuitive. Get it even a little bit wrong, and the whole language fights you the entire way.

Posted

 

The files currently in the download section are just the animations,  they lack the scritping necessary to implement them into the SexLab engine at this time.  As I don't wish to cause any oddball issues with custom versions of Ashal's scripts, I have pulled all scripting and behavior files and only the raw animation files remain at this time. 

 

 

 

That being said, the only way the animations work in game at the moment is if Ashal includes them in his mod ( SexLab ) or if another mod-maker includes them within their mod.  My apologies for that. 

 

 

Well, I put them in my MoreNastyCritters. Simply because its a shame to not have them available for everyone.

 

I dont guarantee MNC wont cause any problems at all down the road. I cant. But it tries to play as nice as possible with SL, not changing any of Ashals scripts and so on.

That being said use at your own risk. Its an option.  :blush:

 

Posted

Spent many hours today working on the next scene for the Atronach.  Swear I could tweak on these things all day long and never be happy with them :|

 

Anyhoo, the wip image:

 

post-103460-0-25742100-1403219773_thumb.jpg

 

-edit-

 

and the laptop finally finished with the motion test render ( ~ 2 hours )

 

post-103460-0-06010900-1403233959_thumb.gif

Posted

Refactoring ate my property reader, but in rebuilding it, got it to read bools properly, and inflate embedded compressed data, or I think it will when it's not in pieces at least. Needs more refactoring, but slowly getting into proper shape. C# is funny like that, if you get the OO part right, it's quick, easy, and intuitive. Get it even a little bit wrong, and the whole language fights you the entire way.

 

.  .  and that is why I will never be a programmer.  I script in a few different shells / languages ( most of the Unix flavors, TCL / Expect ) and hate every moment of it because the syntax rules just get on my last nerve. 

 

One line will have two dozen different punctuation characters of all flavors and god forbid you forget one. . .

 

" There's your problem !   You have six commas and a semi-colon before the opening bracket and after the dollar sign while you only needed four commas and a less than sign !  Too easy !!   :|  ( shoot me now ) "

 

 

lol

Posted

Spent many hours today working on the next scene for the Atronach.  Swear I could tweak on these things all day long and never be happy with them :|

 

Anyhoo, the wip image:

 

attachicon.gifflaming_missionary.jpg

 

-edit-

 

and the laptop finally finished with the motion test render ( ~ 2 hours )

 

attachicon.gifAtronoch_S4_Motion_Test.gif

 

Looks very nice!

 

Care to release the .hkx? I'd be happy to put them into MNC as stage 2  :blush:

 

And I simply have to ask again: how are the Netch coming along?

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...