hope the german authority doen not take THAT long.
i still continue to update the current version.
right now im nearly finished with implementing an update mechanism. with this you can check and update the tool inside the tool.
the last days i added some new features and fixed some bugs
added context menues for selected clusters (works inside cluster browser and main window)
do selections (invert, all etc.)
change colors (group, unique, rest)
change properties( canrotate, isrectangle etc.)
groups ( create, move , clear)
apply current settings
add save comparsion image ( see examples below )
improved import (some performace tweaks)
fixed some import errors when polys are found without uvs)
fixed selection crashes ( WPF binding magic)
fixed threading crashes
started with Professional version
added optional commandline argument (xml file)
parameters can change nogui, maxruncount, desiredusedarea, write comparsion image and alot more
add xml setting file per mesh ( override global values)
can alter resolution, aspect height, maxruncount etc.
added batch operation to process and pack mutliple meshes
im not sitting around and waiting for a better day ^^
here are some examples of the new comparsion image the tool can write (click to enlarge)
Example 4 is a multi mesh file. it was loaded and alligned to grid and then repacked to one uv
Based on descriptions and comments, this looks like a badass tool. Will check it out when I get the chance. Thanks for the hard work and frequent updates!
I apologise if this has already been answered somewhere in here, but will there be some control over the rotation of the pieces, ie if there is a way to keep as many pieces vertical as possible, ie not rotated and still packed as tightly as possible?
nice work! to those involved. the industry's solutions for uv tools have been scattered for years which is why it's been such a hurdle for many artists.
newest features for the upcoming release (finished and tested with internal testing crew)
update inside the tool
tool will check for a newer version, will ask for close and update if newer is found. on completion it will restart the tool
no more prompt when creating a new file
Window will be resized to 95% the Primary screen size if window is not fitting inside
Autoload Obj on Startup
tool will create a directory named StartUpDefault in the IPackThat directory. If obj files are located there, it will load the first one found on startup
User can now save global settings (will be loaded and use each time a file is loaded)
User can save current mesh settings ( will be used if this file is loaded)
Update mechanism finished. Checks for update on startup and ask to update. Tool will restart when finished
Added button to see the changelog of the updates.
Fixed some crashes
Nearly finished steam implementation. Using web API to validate ownership. That way you don't have the steam client running or installed. You need to enter the username and pass if course ^^
cluster browser can now be toggled on or off via button
Hello. First - Brilliant tool!
I have a questions about margins too.
In 'UVLayout' (for example) 8 pixel setting == 16 pixels between the shell edges - 8 per shell ( no intersecting shell bleeds)
Here 8 pixel margin == 4 pixels for bleeding per shell
Which one is technically correct ?
Thank you.
Radical! Does the margin take texture compression into account?
if you are referencing to the special pixel treatment for unreal lightmaps (edge pixels need to be on pixels with power of 2) it doesn't. to be sure that you dont get bleedings you need to make the margins a bit larger hten usual.
Hello. First - Brilliant tool!
I have a questions about margins too.
In 'UVLayout' (for example) 8 pixel setting == 16 pixels between the shell edges - 8 per shell ( no intersecting shell bleeds)
Here 8 pixel margin == 4 pixels for bleeding per shell
Which one is technically correct ?
Thank you.
they both are right ^^. its just a matter of the perspective. when i implemented the margins i set them to the wanted size between each shell. but the other way is valid also. i could easily change the margin math so you insert the shell margin around each shell ( 4 pixels each shell means 8 pixels total).
one of my testers found some nasty memory leaks the last days. currently try to solve it. very ugly bug thats involing bindings and wpf ( thanks microsoft *G* )
the leak only occurs inside the running app when you load another file. hope i can solve this fast.
I did the first test with mergin prepacked clusters. selected clusters can be merged and use at is and will be packed as one cluster in the final solution. still some problems with hole detection but im on it ^^
about the german authority, still nothing from them. the stuff need to go up from the lawyer up to the court *sigh*
Fixed a precision bug with low margins and low resolution. This was producing overlapping bugs
Minimum margin is now 1 ( to avoid low extreme precision bug )
Minimum resolution is now 256 ( to avoid low extreme precision bug )
fixed undo on numerical up down controls
max undo levels on text input fields are set to one
clusters can be merged now ( will be used as is and used as one cluster)
entire cluster settings are stored vie mesh config as well.
Can be assigned via context menu.
Recovers all cluster properties and also merged clusters.
Only works if the obj file has not changed since the mesh config was written.
change log of updates can be viewed inside the tool
tool checks for updates on start and ask to update. Will restart tool when finished.
optimized code and did some minor fixes.
internal packing methods now make us of parallel operations ( better usage of available cores )
fixed refresh of margin preview when transforming clusters manually or enable the margin preview
Started finally today with the fbx sdk. Was able to load a obj file via fbx importer and could access the mesh. Didn't traversed through the mesh layers yet ^^ but it's good start.
I was using this whenever i needed it, up until recently when i noticed i could no longer save out the results... is there any way i can pre-order and regain the ability to save? It's extremely frustrating not being able to use this anymore.
Hey just tried downloading this software, it just says loading forever when I try to load my .obj file. This could be because my .obj is quite high poly, but just wanted to let you know this tool does not work for me.
It might have been asked before but is it intentional that IPackThat resets the imported UVs regardless of the % of wasted area at the start?
I mean, for example if the imported UVs have a wasted area of say 18%, then the program ignores that and starts packing things again at less optimized % instead of respecting the imported one it until it founds a more optimized %.
Just mentioning this since currently I was checking the demo and the program is still packing at a less optimized % than originally imported after 20 minutes.
@ tach
i believe you want to see ipackthat do something, even if the packed% is less optimized than your imported mesh. in this case you'd stay with your original mesh anyway.
otherwise it might look like a bug, when the programm packs for ages and actually won't produce any changed % values than your imported one.
Indeed. It just seemed a little impractical since unless you specifically write down the initial % then there's no prompt or anything saying that the end result is less optimized or comparing it to the initial %.
I was using this whenever i needed it, up until recently when i noticed i could no longer save out the results... is there any way i can pre-order and regain the ability to save? It's extremely frustrating not being able to use this anymore.
the demo never was able to save. we do as fast as possible. right now we are waiting for german authority
It might have been asked before but is it intentional that IPackThat resets the imported UVs regardless of the % of wasted area at the start?
I mean, for example if the imported UVs have a wasted area of say 18%, then the program ignores that and starts packing things again at less optimized % instead of respecting the imported one it until it founds a more optimized %.
Just mentioning this since currently I was checking the demo and the program is still packing at a less optimized % than originally imported after 20 minutes.
the packer works different than normal packers. i have to restart and do a shape shape test for every single shell. thats why i cant reuse an existing on. where do is start? topleft to downright or down to up?
if you import an allready very optimized uv pack and the tool cant find a better one, then its better to stay with the original version. the tool is not the holy grail thats always will do the best packing possible. its something to prevent manual packing in the first way. can you post an image of the original version? sometimes the imported version did not respect the margin correctly and the values for used area without margin are totaly off. if the margins are treated correctly the values will be true.
Hey just tried downloading this software, it just says loading forever when I try to load my .obj file. This could be because my .obj is quite high poly, but just wanted to let you know this tool does not work for me.
there was a bug until recently with not initialzed uvs within the mesh. if the mesh contains polygons without uv, the demo produces errors on import. in the currently internal beta version this bug has been fixed.
additional progress update
clusterbrowser visibility is now used and writen with the global default configs
additional packing improvement ( somewhat around 25% to 40% faster, depending of you amaount of cores)
currently integrating fbx to the tool. right now im only able to see one uv channel (can switch to different ones per code so far *g*) but its a good start.
need to change the scene handling but the basics are in now. the fbx sdk i used is the latest 2015.1 sdk version.
Our CEO is trying to push things with the german authority to get all legal stuff finished. but he has to be carefull not to push to much ^^
[*]if you import an allready very optimized uv pack and the tool cant find a better one, then its better to stay with the original version. the tool is not the holy grail thats always will do the best packing possible. its something to prevent manual packing in the first way. can you post an image of the original version? sometimes the imported version did not respect the margin correctly and the values for used area without margin are totaly off. if the margins are treated correctly the values will be true.
I unserstand. Simply wanted to know if the reset was intentional. And I'll download the newest version later and send you some images via pm then (nda stuff).
It's also worth mentioning that I was testing things on an organic model. Hence why I think it struggles a little with optimization. This tool is probably most awesome for hard surface stuff where uvs tend to be more squared.
Replies
i still continue to update the current version.
right now im nearly finished with implementing an update mechanism. with this you can check and update the tool inside the tool.
the last days i added some new features and fixed some bugs
- added context menues for selected clusters (works inside cluster browser and main window)
- do selections (invert, all etc.)
- change colors (group, unique, rest)
- change properties( canrotate, isrectangle etc.)
- groups ( create, move , clear)
- apply current settings
- add save comparsion image ( see examples below )
- improved import (some performace tweaks)
- fixed some import errors when polys are found without uvs)
- fixed selection crashes ( WPF binding magic)
- fixed threading crashes
- started with Professional version
- added optional commandline argument (xml file)
- parameters can change nogui, maxruncount, desiredusedarea, write comparsion image and alot more
- add xml setting file per mesh ( override global values)
- can alter resolution, aspect height, maxruncount etc.
- added batch operation to process and pack mutliple meshes
im not sitting around and waiting for a better day ^^here are some examples of the new comparsion image the tool can write (click to enlarge)
Example 4 is a multi mesh file. it was loaded and alligned to grid and then repacked to one uv
I should let the last one run longer, had to shut down my computer.
*also smaller images next time
*In exchange for appropriate sums of currency, of course.
Cheers
Looking forward to this then !
I would love to integrate the tool in my workflow and my maya scripts.
It has just above 5200 uv islands.
I have a questions about margins too.
In 'UVLayout' (for example) 8 pixel setting == 16 pixels between the shell edges - 8 per shell ( no intersecting shell bleeds)
Here 8 pixel margin == 4 pixels for bleeding per shell
Which one is technically correct ?
Thank you.
one of my testers found some nasty memory leaks the last days. currently try to solve it. very ugly bug thats involing bindings and wpf ( thanks microsoft *G* )
the leak only occurs inside the running app when you load another file. hope i can solve this fast.
I did the first test with mergin prepacked clusters. selected clusters can be merged and use at is and will be packed as one cluster in the final solution. still some problems with hole detection but im on it ^^
about the german authority, still nothing from them. the stuff need to go up from the lawyer up to the court *sigh*
Well, I´d second that
That function confused me also.
Started finally today with the fbx sdk. Was able to load a obj file via fbx importer and could access the mesh. Didn't traversed through the mesh layers yet ^^ but it's good start.
Cheers
Are there any news on a release date yet? :-)
It's about time that manual unwrapping time can be cut in half or completely eliminated !
I mean, for example if the imported UVs have a wasted area of say 18%, then the program ignores that and starts packing things again at less optimized % instead of respecting the imported one it until it founds a more optimized %.
Just mentioning this since currently I was checking the demo and the program is still packing at a less optimized % than originally imported after 20 minutes.
i believe you want to see ipackthat do something, even if the packed% is less optimized than your imported mesh. in this case you'd stay with your original mesh anyway.
otherwise it might look like a bug, when the programm packs for ages and actually won't produce any changed % values than your imported one.
additional progress update
currently integrating fbx to the tool. right now im only able to see one uv channel (can switch to different ones per code so far *g*) but its a good start.
need to change the scene handling but the basics are in now. the fbx sdk i used is the latest 2015.1 sdk version.
Our CEO is trying to push things with the german authority to get all legal stuff finished. but he has to be carefull not to push to much ^^
cheers
I believe it runs 10 tiles across in U, and then extremely far up in V.
http://bneall.blogspot.co.uk/p/udim-guide.html
I unserstand. Simply wanted to know if the reset was intentional. And I'll download the newest version later and send you some images via pm then (nda stuff).
It's also worth mentioning that I was testing things on an organic model. Hence why I think it struggles a little with optimization. This tool is probably most awesome for hard surface stuff where uvs tend to be more squared.