Messages

16

General / Re: Photo Realignment

« on: March 04, 2014, 08:51:57 PM »

Are you using masks ? In some cases when cameras won't align (i am doing human scans also), i delete masks from non-aligned photos and do realigment and it works in 90%. Another thing to try is to use Generic pair selection (in aligment stage) or Disable if you are using pair preselction.

I tried deleting the masks for the NA cameras, and the realignments still would not work.  I normally use Generic pair selection during alignment.  What do you mean specifically by pair preselection, are you specifying this via Ground Control?  I'm unaware of anywhere else to specify this type of relationship for the cameras, but it sounds very useful since our cameras are in a fixed position and pairings.  Thanks!

17

General / Re: Photo Realignment

« on: March 04, 2014, 10:47:15 AM »

I am doing human scans, with a similar issue.  On particularly difficult subjects that have very bland clothing, I get say 10% of the cameras that won't align.  After the initial solve, I try to right-click "Align Selected Camera" on these NA cameras, and they still won't align.

Is there any other way to deal with this?  Is this one of the situations that coded targets are meant to help with specifically?  Can one load a camera XML from a previous solve, with this initial positioning being a 'suggestion' of where to start?  Could one hack the "ground control" system to allow this?

18

General / Re: New GPU to improve speed or useless

« on: March 04, 2014, 04:54:08 AM »

Our computing machine has an i7-4930K with two R9 280X cards in it, and we are using 60 pictures at 18MP apiece (Canon 600D).  Timing a 'high' dense cloud reconstruction using the various optional numbers of GPUs we got:

0 GPU: 67 minutes
1 GPU: 23 minutes
2 GPU: 16 minutes

So while the first GPU got us a massive speedup, the second GPU got us just 30% faster, and sped up the entire process (photos->textured model) by 15%.  Given that each card is 12% of the system cost, it barely makes sense to add the second GPU.  What's more interesting though, is dividing the cores and GPUs amongst two PS instances, and running them at the same time.  Then you get a real nice speed up!  Assuming you have the need to run multiple PS instances simultaneously.

19

Face and Body Scanning / Re: Calibrating Lenses for scan perfection

« on: March 04, 2014, 03:11:28 AM »

This can work but I can say from experience, over time from micro vibrations and other factors, imported camera alignment will fail over the course of a shoot. Say if you shoot over a couple of hours and keep re-importing from an original calibration set from the start. If you take continual update shots as you go of the reference object, it will help.

Otherwise re-using the start reference capture set, you will start to notice noise and misalignment creep in.

Thanks Lee. I have certainly noticed this too, when poor subject matter (black suit pants for example) resulted in large blind spots in the solve, as about 10% of cameras failed to align.  I loaded in cameras from another solve from photos taken about 30 minutes prior.  Even though all cameras are now "aligned" and the poorly-solved areas are vastly improved, others areas of the model now have more noise than before.

Is there any other way to deal with this?  Can the cameras be re-solved with this initial location being a 'suggestion' of where to start?  If I re-solve all cameras, they seem to ignore their previous location/orientation, but if I try to right-click "Align Selected Camera" on the NA cameras after the 90% of cameras have aligned, the remaining cameras still won't align.  It looks like the use of coded targets helps specifically to prevent this issue, I'm certainly keen to try that.

20

Face and Body Scanning / Re: Automatic Masking

« on: March 03, 2014, 10:03:16 AM »

We name our photos with the following convention: [D]/[B4]/[C], then you just load the subject and mask images from the respective batch directories.  I will often rename the batch directory to something more descriptive, then the same name + "_bg" for the mask images.

This does mean that we often have to "reset batch number" via the options screen when cameras fail to fire, it's a little annoying but keeps things pretty well-organized.  I will probably script up the reset one day and make it more easily accessible. Francis from the SS team, has also said he will move this function to a higher menu level soon.

21

General / Re: Canon EOS 600D Corded Remote trigger

« on: February 28, 2014, 10:43:36 AM »

We experimented with full corded setups, as well as a combination of wireless and corded setups (one wireless transceiver per column, split out to all cameras).  We got weird behavior with the latter, and have since just settled on all wireless transceivers as the hassle of the wires was too much, but we may switch back because swapping out dead batteries on 60 cameras is indeed a pain...especially because who knows when each one will go out.

http://www.amazon.com/dp/B0099SGFZI

If you do an all corded setup, you can get away with straight audio cables/splitters/plug converters, mostly bought through monoprice.com, be careful though because you do indeed need good quality all the way through the system.  No 'easy' solution here, but I can say we had an all-wired setup working with 60 cameras, and no issues with impedance, or needing to boost the signal at any point.

22

Face and Body Scanning / Re: Calibrating Lenses for scan perfection

« on: February 28, 2014, 01:44:12 AM »

Not sure whether upfront calibration will improve things. I once read you might be better off placing some well-structured object in your recording volume and calibrate the cameras from that.

Do you remember where you read this?  I am very interested in setting up something like this to:

1) initially focus the cameras
2) using coded targets, establish world units as well as orientation
3) get a rock-solid solve on camera positions, which can then be re-loaded to other sets of images (particularly uncooperative smooth/featureless surfaces that don't align nicely otherwise)

I'm thinking a mannequin with noisy spraypaint applied, plus coded targets on the ground and a vertical bar, could produce this result.  Curious if anyone has attempted this already?

23

Face and Body Scanning / Re: Most reasonable DSLR for multi-cam face capture setup

« on: February 13, 2014, 10:48:40 PM »

What is the appeal of this model, given that a T3i/600D sells on Amazon for $50 less, and seems close to same in spec with the addition of an adjustable LCD display?

http://www.amazon.com/dp/B004J3V90Y
http://www.amazon.com/dp/B00IB1BTWI

Will it ultimately be priced lower than a T3i/600D?

24

Face and Body Scanning / Re: Most reasonable DSLR for multi-cam face capture setup

« on: January 22, 2014, 09:23:46 PM »

Thanks Lee!  So far my tests had been on Medium, I ran some more tests on High and the results got much closer.  Our lighting is pretty diffuse, but our tests have been on a turntable so far, so there is probably some directional bias that would favor the Nikons as you've suggested.  Not to mention the Canons are significantly less $$$, which means more cameras for coverage.

Given the extra ease of use, MLU support, standard connectors, superior textures, and overwhelming community endorsement, Canon makes sense!

25

General / Re: Benchmarking a GPUs

« on: January 22, 2014, 12:11:53 PM »

Wishgranter, any luck with a nice list?  I keep flipping through the posts here yet don't feel like I quite have it all straight.

Please oh please has anyone benched an R9 290X yet?  It seems from latest Anandtech review, the 290X and 280X dance around each other, and the 290 is maybe 5% slower than the 290X:

http://www.anandtech.com/show/7481/the-amd-radeon-r9-290-review/14

Given that the 290 ran about 25% faster than the 7970 earlier in this thread, the 290X should then only manage 30-35% faster than 7970/280X, sound right?

Just about to plunk down on a new system here, and trying to decide between two 280X, and a single 290X.  The single 290X would provide option for a second 290X later, however it sounds like people are having power and cooling problems with getting a 2 x 290X system running nicely...is anyone running a 2 x 280X system run alright with just air cooling?

26

General / Re: Strobe-based Image Projection

« on: January 20, 2014, 09:55:45 AM »

Just got our LB, using it with a Yongnuo YN 560 III, and Canon 18-55mm kit lens, and the results are pretty dim.  You can get full body coverage if you put it on its side, you probably want 3 or 4 for full 360 body scan.  As others have noted, output is dim such that the room must essentially be dark otherwise, any strobe combination results in it getting washed out.

Anyone have ideas for driving more light through the LB?

27

Face and Body Scanning / Re: Most reasonable DSLR for multi-cam face capture setup

« on: January 20, 2014, 09:34:22 AM »

Lee/all, one more question about the Nikons for y'all if you don't mind!  I dig the Canons.  However I've done several tests with 6 each of Canons and Nikons stacked right on each other in stereo pairs, and the Nikons generally result in 20-30% more points resolved, and the resulting mesh is definitely less noisy.  Sounds like some have found otherwise?

I suspect that while the Nikons images are not pin-sharp at the pixel level, perhaps overall they still contain more detail than a sharp 18MP image.

Of course, to make it interesting, the resulting textures from the Nikons aren't as sharp or color-accurate as the Canon, on the same geometry.  And great textures is a big reason we're all here...

Would it be madness to attempt an array of mixed Canons and Nikons, using all images for the mesh solve, and only the Canon images for textures?  That makes my head hurt. 

28

Face and Body Scanning / Re: Max cameras per computer over USB

« on: January 17, 2014, 10:53:38 PM »

Update: we bought a generic SIIG PCIe card at Fry's, which shows up in Win 8 as a "Renesas Electronics USB 3.0 Host Controller".  It maxed out at 22 devices when plugging in a USB 3.0 hub directly, but when plugging in a USB 2.0 hub directly, we are at 25 devices and counting.  So it looks like we'll be okay, thanks all!  Still getting Win7 on the next machine though.

29

Face and Body Scanning / Re: Max cameras per computer over USB

« on: January 17, 2014, 01:00:08 AM »

Thanks so much everyone for the quick and helpful replies!  The machine has Intel B85 Express chipset and Intel USB 3.0 eXtensible Host Controller, so it is probably having the problems you are referring to Wishgranter.  The PCIe card I bought has a VIA USB 3.0 eXtensible Host Controller, which sounds suspiciously similar to the Intel, so is probably why it has the same problems, even using VIA's driver.

Using USB 2.0 hubs or plugging into the USB 2.0 ports helps us add a few devices, but just a few.  The connections show up under the same Intel USB 3.0 controller even when plugged into USB 2.0 ports, so I think we are screwed without a clever driver hack, or installing Windows 7. 

My Win7 laptop happily accepts 24 devices with no signs of stopping, and says that each camera reserves just 1% of system bandwidth.  This is the case for both its Intel USB 2.0 controller, and separate AsMedia USB 3.0 controller (except it does not show bandwidth usage).  Looks like it is our Smart Shooter machine for now!

Is anyone who is having success using a Win8 machine?  If so, what motherboard/USB 3.0 controller are you running?

30

Face and Body Scanning / Max cameras per computer over USB

« on: January 16, 2014, 06:56:02 AM »

Hi all, coming up against the USB max devices problem here in building out our array.

We have been using 12 cameras connected to a single Windows 8 machine, works great.  However upon plugging in a 13th camera, Windows 8 rejects it, citing "Not Enough USB Controller Resources".  I suspect that each camera is getting allocated a certain amount of USB bandwidth, and because the cameras only connect via USB 2.0, they run out of the needed bandwidth rather quickly.

To compensate, I got a PCI-e USB 3.0 controller card, hoping that a dedicated card might fare better, or at least allow another 12 to be connected.  However the card only allowed 7 more cameras to be connected before throwing up the same error.  Which seems bizarre, since the PCIe x1 slot that it's in should have adequate bandwidth available to it (500Mbps) that USB 2.0 requires (480Mbps)...could it be the B85 chipset on the motherboard coming up short somehow?

http://www.newegg.com/Product/Product.aspx?Item=17Z-0002-00002

Unfortunately I am unable to conclusively confirm the exact problem on my Windows 8 machine, as I cannot find a utility for querying each device's bandwidth allocation (it used to be conveniently available in Win7's Device Manager).  Does anyone have a favorite utility for checking this in Windows 8?

The above aside, is everyone just going the 12 cameras per computer route, using minimal machines, and having them all write to a shared network location?