Richard House

@Michael Tetzlaff Hi Michael, I appreciate your reply. I have re-processed both options... Specular Value at 4 and Specular Value at 2. They both resolved the bright yellow color issue in the Sketchfab render. Specular setting at 4 seems to deal with the open hole area on the front surface a bit better than Specular setting 2. Here are links to these two most recent versions: Specular 4 version: https://skfb.ly/p7FFu Specular 2 version: https://skfb.ly/p7FGM Thanks! Rich

I'll keep this thread going since I'm still working on the same subject here. I tried using a ring light from slightly further back hoping it would record as a smaller, more specular light source. This time around, I am seeing some strange artifacts from Kintsugi 3D's texture mapping... specifically in the specular render. Some screenshots of the Sketchfab model renderings... Not sure where the bright lime green/yellow artifacts are coming from. This version is a second attempt at rendering the same model in Kintsugi. I got the same result from both the first and second renderings so perhaps not a glitch in that process.. Metalness render: Specular render: The Metashape exported model: https://skfb.ly/ppSXy The Kintsugi modified textures: reprocessed a second time: https://skfb.ly/ppSUA Thanks, -Rich

Brass Box Photogrammetry- Ring Flash Update: The Canon MX-14 Macro Ring light solved my issues for Metashape being able to build the model... (381 photos/ 381 aligned). I have posted two models to Sketchfab. One model is directly exported from Metashape: https://skfb.ly/p6uJ6 The second model is textured using Kintsugi 3D: https://skfb.ly/p6tvQ The main inaccuracy I encountered is that there are a few holes in this box along the front. The larger open holes along the proper left side wound up reflecting the interior bronze surface from the ring light direction and makes it appear to be a solid surface. I would need to line the interior with something black so that it doesn't reflect the bronze color back to the lens for that particular detail to read accurately. Otherwise, Kintsugi 3D does a very nice job rendering this surface. It has a dull quality to it but will appear more glossy with more direct light angles. The Kintsugi rendering reads rather true to the actual box. Thanks again for the feedback on how best to approach this one. -Rich Example where hole in front reads as a hole Ring Light flash angle reflecting the bronze interior. This occurs in many of the captures and makes it appear to be solid in the rendering.

Thank you all for the replies. This is all very helpful. Michael, I did use your data set that was shared in the workshop to do the correction of the lighting offset. That step wasn't too difficult to dial in for me. I was thinking that for the purpose of Beta testing, I should try to do a capture set with a hot-shoe flash. The shiny bronze subjects I am testing obviously are not a good fit for that so yes, I will redo these objects from scratch using a ring light flash on Manual output. I just ordered one of the Canon Macro Ring lite flashes since my other option has been a Profoto Ring light which is larger in diameter and quite difficult to dial down the power enough when close in on small subjects. To clarify the processing of images for Metashape and then for Kintsugi 3D... You suggest I pull the highlights down and open up the shadow info for Metashape? I'm guessing it would be better to reset the adjustments to a linear response curve with no highlight/shadow adjustments and re-export the same image set for Kintsugi 3D to work with. I understand that the tone calibration would reset the values on the greyscale chart to correct the tone mapped images but that's probably better done by re-exporting the from the original raw files. Am I understanding this correctly that the best workflow would be to process out two separate image sets without any cropping applied between them. Fingers crossed the ring light I ordered arrives soon so I can get these shot again soon. I will keep updating this forum and share the results when I have them. Thanks again, Rich (Kurt, I might reach out to you about your Blender render steps once I have good models to import. Using Blender to output videos is something I haven't done yet.)

Hello, My most recent attempts for Kintsugi were done with some shiny brass objects using a Canon 580EXII flash on a Canon DSLR. I wanted to test out the workflow when it requires recalibrating the shadow distance in Kintsugi. I came upon issues in Metashape's attempt to align the photos as these photos weren't ultimately lit evenly enough to match pixel data. One object is a brass box and the other is a brass lamp. These objects are less than one foot along their maximum dimension which means the camera and flash were somewhat close to the turntable during capture. Shooting horizontal, rotating left and rotating right with the hotshoe flash result in very different looks on the objects. The thing that I believe is happening is that the flash distance from the lens is significant enough that the lighting changes the look of the object too much to match pixels. With the lamp, there is an internal area that goes into shadow from these different rotations. I am thinking a ring flash is going to be necessary to get the good underlying model for these two objects and obviously will work well for Kintsugi. Is anyone else working with hotshoe flash distances that have shadows interfering with getting a good model? I think going back to a ring flash will be the best solution here. I could try using 4ft long tube led tube lighting to get soft lighting close in to the lens. I suspect my 4 ft led tubes would be too soft of a light source for Kintsugi even if they are very close to the lens. (??) I may just need to be further away and with somewhat larger objects to use an on camera flash without the shadow distance getting exaggerated by being too close in. The other concern with these objects is that when a surface is angled away from the lens, it goes quite a bit darker. When the flatter surfaces aim back to the camera, I get direct glare/flare which need to be underexposed to make out much of the surface. Will Kintsugi work with mapping textures if the light source varies in power... for example if it was set to TTL for controlling the light output? Thanks! Rich

@Michael TetzlaffMichael, are you suggesting the workflow is to decimate in Metashape down to around a 64K face model but build the normals map from the highest quality model? (regardless of the actual face count size). Or do you think that building a new normals map in Kintsugi 3d from scratch is preferred? Thanks, Rich

Thank you for these replies Michael and Dave. Michael, I suspected as much and will be mindful of the intended final viewing environment when I decimate and built the textures of my model in Metashape before working on it in Kintsugi 3D. Dave, That is a very good tip. This particular model was a fair amount too big but I have had some issues previously when the png texture map format pushes the total file size just over the limit and I had to re-export in the jpeg format instead. I'll try out a Zip archive format next time that happens. Rich

Hi. I am looking for a little clarification on how texture files work with the .obj or .glb files. I wound up processing an UltraHighQuality model in Metashape. I then used that model in Kintsugi 3D to rebuild the new texture files. Upon attempting to upload to Sketchfab is when I realized I have a model that is too big to upload with my account limitations. I am not sure if I can just go back to Metashape and decimate the UHQ model down to 64K faces and then still use the texture files built from the UltraHighQuality model or if I need to redo the Kintsugi3D process using the 64K decimated model. Thanks for any clarification on how texture files relate to the differing mesh quality models. Rich House