Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - lecaramel

Pages: [1]
Face and Body Scanning / Re: Siggraph
« on: May 21, 2014, 12:54:03 PM »
No Siggraph for me this year.
The tradeshow is going smaller and smaller each years and the Vancouver one is always smaller than the LA's ones. Even if Vancouver is great city! :)

Face and Body Scanning / Re: 3D scan for video concert
« on: May 15, 2014, 10:13:19 AM »
This performance?

I guess you mixed both pre computed scans sequence associated with live capture (with a Kinect style system)?

Anyway, great performance if it's what you was working on!

Face and Body Scanning / Re: 3D scan for video concert
« on: March 24, 2014, 01:29:22 PM »
Non, je suis ? Bordeaux, je travaille chez Pixologic, ?diteur de ZBrush.

I'll be glad to see the result if it provide the result that your client is happy with it. Motion blur isn't an issue?

Face and Body Scanning / Re: 3D scan for video concert
« on: March 24, 2014, 12:37:13 AM »
Nicow : merci, je garde ?a dans un coin.
Are you retouching your scans in 3D? I mean if you are recreating a kind of "3D animation", you are looking more for the texture aspect of the scan than the accuracy of the geometry itself, right?

Just curious :)

Face and Body Scanning / Re: 3D scan for video concert
« on: March 18, 2014, 10:51:05 PM »
Nico: which company?

General / Re: New GPU to improve speed or useless
« on: March 06, 2014, 11:36:14 AM »
Thank you for the link, I'll look that closely!

General / Re: New GPU to improve speed or useless
« on: March 04, 2014, 10:48:59 AM »
Thank you for this feedback and more important, this tip about running two instances of PhotoScan by dividing the resources.
I'm more in the process of increasing my current workflow since I'm doing mainly tests and not using PhotoScan in a production environment, then I'm processing just photos after shooting, but since one of my co worker starting to take photos for scans too, perhaps it could be a smart move to try this two instances of PhotoScan.

General / Re: New GPU to improve speed or useless
« on: February 24, 2014, 09:55:40 PM »
Thank you David for this reply and thee information.

Since dense cloud is one of the longuest step, perhaps an extra GPU could help a little more than a little bit...

Thanks again!

General / New GPU to improve speed or useless
« on: February 24, 2014, 04:09:25 PM »

I'm currently looking about improving the photoscan process by giving a boost to my computer. I don't want to change my workstation since it's a 6 months old one...

Right now, I have a i7 4770L + GTX780 and 32 Gb of RAM.

I found in an old thread (dec 2010...) that the GPU was used only on the geometry building stage, but not in camera alignment, dense cloud and texture creation.
Then if the GPU is still used for only the geometry stage, I may prefer spending some money in a new lens or an extra DSLR, but if the GPU is used and in almost or all the steps, perhaps it's worth buying a second GPU.

And about the CPU, if I disable a total of 4 cores in the preferences (2 per card since my hyper threading is enable), how things are working? only 4 physical cores will be use and not the HT ones? or it's an "unknown"?


Does someone have the info?

Face and Body Scanning / Re: improving quality on shooting time + questions
« on: February 19, 2014, 02:57:06 AM »
Thank you David for these informations.

I mentionned the 85mm because I have a prime lens which is very sharp at f/10. I'm also using Lightroom to edit my raw and then eventually denoising them, but of course, avoiding noise is I guess, a top priority!

General / Re: Zbrush - Workflow
« on: February 18, 2014, 11:41:34 PM »
The webinar is for french people, in french language. Then feel free to join, but I won't do it in english language :)

But if their is enough requests, I can organize one later if you want.

Face and Body Scanning / improving quality on shooting time + questions
« on: February 18, 2014, 05:04:48 PM »

I have several technical questions in relation to the shooting. I'll have to shoot people with just one DSLR (5D MkIII) and doing head scan and body scan. Of course I know that doing a scan with just one DSLR won't produce the best scan ever, but i already have good results, then for that, it's ok.

1 - In the documentation, it's mentioned that fixed focal lenses are advised. Is using a zoom lens (24-70mm) fixed at 50mm will be problematic? Even if it have some micro variations of zooming while shooting, between two photos?

2 - What is the best focal lens to provide good quality? 50mm 85mm? (on a full frame camera).

3 - What is the best compromise in terms of scan quality: increasing the ISO to increase the depth of field or lowering as much as possible the ISO even if it reduce the depth of field? Because I'll shoot outside, I won't be able to use studio lights. Of course, I would like to have a decent shutter speed.

4 - To improve the alignment of the photos, then the dense cloud generation (and reduce as much as possible the noise) in the final model, is it better to do a set of photos which will include the whole model in each photo for a first pass, then in a second pass, doing close up, or only close up are enough? Having these photos taken far from the model may introduce some extra noise?
I'm considering doing my scans in several pass in fact, like the full body in low quality, then shooting separately the face/head, the hands, parts of the body, then when doing, processing the scans separately, then at the end, recombining everything in one mesh in ZBrush. It may be a longer process, but perhaps can increase the quality of scans taken in not very good conditions?

5 - in addition to above, is it better to do in several chunks a scan by far and a scan with closeup photos and then combining them or just having all the photos, far and close, in the same chunk?

6. The Ultra high quality of dense cloud worth the extra waiting time in terms of quality (at least for a face/head) or high is far enough? I mean if adding 10 extra hours of computing versus 10 hours (or less) of 3D retouching in ZBrush.

7. no problems on shooting photos with two different focal lens? (35mm for photos by far and 85mm for closeup)

Thank you for your help!

General / Re: Zbrush - Workflow
« on: February 18, 2014, 04:52:32 PM »
Cannot wait for 64bit and being able to sculpt over 10 million polygons reliably, as well as be able to support larger than 8k textures :)
The problem is that their is a difference between sculpting (brush strokes) and several operations, like Projections. Projection can need 4 times more memory than usual operations, then yes, if ZBrush is at its limit with the mesh itself, things can become problematic :)

ZBrush crashes when exporting more than 7 million polygons as .OBJ. Paul Gaboury mentioned this could be a legacy issue because ZB was never built to handle these types of sizes.. It's possible to use Goz to Modo but not ideal.
A trick is to try exporting through 3D Print Exporter or through Decimation Master (you have export options in these plugins), which are using a specific exporter, not the ZBrush native one).

I would be happy to share any data with Pixologic. I also have a bunch of ideas that could be useful for ZBrush 5 to help the industry that deal with scanned data. Zbrush hasn't had much love with regards to supporting the growing scanning industry and we could do with some :)

Things like intelligent hole fill, multi camera import, paint layers. FYI im on the ZBrush beta but have never heard anything since I signed the papers back in 2011!!
Well, all your ideas are more than welcome. We are now two team members at Pixologic using PhotoScan and working with scan datas and facing real issues ourselves is also the best thing to understand your problems.. and as you can imagine, it's not the same as having a "bug report" :)

@bmc130 : yes you can paint on texture directly with the help of ZappLink and Photoshop by projection, but not a "real" paint on texture like you are doing on some other software.

Sadly there are many bugs with this method, going back and forth to PS, projection errors, seams, back side projection etc Also the lack of paint layers is a huge missing link.
Yes, I can understand, in a way, I was just answer on the question on a technical point of view.

But without ZBrush it would be near impossible to work with RAW photogrammetry data from Agisoft Photoscan. It's an essential part of the pipeline and incredible at that.
And that's why all feedback is more than welcome.

At which email can I contact you / chat with you?
Also, I'm doing an online webinar in France tomorrow about ZBrush and scan datas. If you have extra samples or raw images, I'll be glad to show them, including what can be wrong in shooting, even with your fantastic setup! :)

General / Re: Zbrush - Workflow
« on: February 17, 2014, 08:07:20 PM »
Sorry for the long delay of my answer, didn't have an email notification...

I'm sorry, but we are never commenting/speaking about future features of ZBrush until they are announced... But i can speak about the 64 bits since it has laready been announced for ZBrush 5.

About GoZ for PhotoScan, in fact, it's more up to Agisoft to do the integration. I can contact them directly. But for sure, it could improve the workflow :)

@infinite: ZBrush is crashing when importing scans with more than 7 millions of polys? Or when manipulating these 7 millions after importing?
You can contact me by email if you want: thomas at pixologic dot com, eventually with your workflow. Also having some of your large scans may be helpfull for us to test (note, I bought a couple of your raw files at triplegangers to do test too..)

@bmc130 : yes you can paint on texture directly with the help of ZappLink and Photoshop by projection, but not a "real" paint on texture like you are doing on some other software.

General / Re: Zbrush - Workflow
« on: January 10, 2014, 11:49:11 PM »
Hi there!

I'm jumping in this topic a little bit late, but just to let you know that I'm working at Pixologic, makers of ZBrush, then I'll be glad to help if you have questions.

I just personally bought PhotoScan few hours ago, then of course, as soon as I'll be able to produce decent scans, I'll do my best to explain also how to simplify the workflow to clean/retouch your scans.

At least, just to simplify this workflow:

Try this:
4. Click on a ZSphere
5. Choose the uncloned high-detailed PS-mesh for "Rigging"
6. Choose the ZRemeshed low-poly mesh as a topology (and you have to move one vertex in order that ZBrush will work correctly - sometimes you have to tweak it imho)
7. Turn on "Projection" (and sometimes you have to tweak the default settings)
8. Preview an "Adaptive Skin" by choosing the desired density for the object (mostly 3-6)
9. "Make Adaptive Skin"

You don't need to use ZSpheres to do that.
Simply subdivide several time your new ZRemeshed model to have at least the same number of polygons than your original scan (to avoid losing information), then in Tool >> SubTool >> Project, press the Project All function: the visible SubTool (the scan), will be projected on the ZRemeshed one.

The good thing about this solution is that you can also mask some areas to protect from the projection, using the morph targets to protect an area, why not projecting on layers, etc.

Anyway, just ask for help!


Pages: [1]