Messages

31

General / Anandtech- Which Sample Set they use?

« on: May 11, 2015, 09:14:27 PM »

I could not find the info:

What set of images does the site anandtech uses for their cpu testing in agisoft photoscan?

I gues its the one on the site here, but which one?:

http://www.agisoft.com/downloads/sample-data/

The doll, monument, spherical panorama, coded targets or Building?

32

Face and Body Scanning / Idea for cheap full body capture with 1 camera?

« on: April 03, 2015, 11:20:08 PM »

Hello, i thought about possibilities for full body 1 camera captureof humans.
1) Face done separately from body, photos taken without stative, the person must not move for about 2-3 minutes- doable.

2) but not moving for a full body doesnt seem realistical, since more photos from more angles are needed, thus i get the idea of using real person clothes on some sort of real life sized figurine or manequine..., something like this:

http://www.ebay.com/itm/Male-Mannequin-Manequin-Manikin-Dress-Form-Display-MD-HMB2BK/161498910830?_trksid=p2047675.c100005.m1851&
_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D1%26asc%3D28790%26meid%3Ddaa7963c2e324f5e98c21137ba065aec%26pid%3D100005%26rk%3D1%26rkt%3D6%26sd%3D350304748442&rt=nc

or

http://www.ebay.com/itm/Male-Plastic-Unbreakable-Mannequin-Display-Dress-Form-Display-Ken-Free-Wig/131383270329?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D1%26asc%3D28790%26meid%3D0a192f655a784f41b103bd9b8ba2d359%26pid%3D100005%26rk%3D5%26rkt%3D6%26sd%3D131471327332&rt=nc

the problem is that these would need to be poseable, at least be able to make a T or A pose for a character (cloth withotu face) scan.

My question is simple, do you know somebody about such figurine/manequine, that would allow such posing? (T or A pose)?

33

General / Re: Benchmarking a GPUs

« on: April 03, 2015, 11:15:23 PM »

Hm... maybe i saw this aswel, but photoscna uses open cl 1 (or 1,1) i think, the test uses probably later versions of open cl...?

Can anyone confirm that this test is useable for gpu agisoft comparison?

34

General / Re: How do you know you got a good sparse point cloud??

« on: April 03, 2015, 10:14:48 PM »

Yes thats my questions aswell... if anyone has an advice please dont hesitate to share it :-).

35

General / Re: What does Optimize Cameras actually do?

« on: April 03, 2015, 09:58:14 PM »

isnt it somehow dependent also on your physical camera calibration done with agisoft lens? (or is it a wrong thought?)

36

General / Re: RAW, JPG and different types of noise

« on: March 02, 2015, 04:29:31 AM »

BTW its an older topic but still... did the smaller size of jpgs also results in faster processing in photoscan? (like quicker photo allign or dense generation)?

And what about lossless compresion of tiffs (lzw) does it contrary slow down processing in photoscan (due to the need to "uncompress" the files when (before) reading them?

37

General / Re: LuxMark 2.0 similiar performance with agisoft in GPU reviews?

« on: February 24, 2015, 11:40:33 PM »

Basic back-of-the-napkin estimations seem to indicate that choosing a graphics card based on LuxMark tests for OpenCL isn't a bad idea. Of course, the guy running 8 Titans is on top of the rankings. If you want the best and don't mind paying for it, wait for the GM200 Maxwell Titan or the R9 390X which should be out later this spring, the R9 390X looks like it might have the edge, but hey, I'm an nvidia guy. Used, I came down on the side of the 780ti for price/performance, but if you want more bang something in the titan range isn't a bad idea. If you have great cooling (especially if you want to run more than 1 gpu) AMD offers some good alternatives. Realistically, at the top end of the gaming level you're probably not going to choose a bad card.

thanks for reply, no im thinkinh, because of budget between gtx 960 or radeon r9 280 or 280X...

Is here someone other that can confirm that performance in Luxmark is similiar to performance in photoscan? im not a programmer so i cannot say... but its all openCL so it should be very similiar,. the performance....?

38

General / Re: LuxMark 2.0 similiar performance with agisoft in GPU reviews?

« on: February 23, 2015, 01:37:03 AM »

Could someone elaborate on this please? I really need to know this

39

General / Re: Graphic card: GTX 960 or R9 280 ?

« on: February 20, 2015, 05:22:19 PM »

ATI cards have much much better Open CL drivers and optimizations compared to Nvidia which chooses to focus its time and energy on CUDA processing.

Thanks, is it a question of purely drivers, or is the also the architecture of the cards...? what im asking is if there is a chance and a possibility that Nvidia will improve greatly their drivers (software) and with this boost their performance in opencl? Is it even possible, or do the ati cards have simply better hardware architecture for open cl?

40

General / LuxMark 2.0 similiar performance with agisoft in GPU reviews?

« on: February 19, 2015, 11:07:58 PM »

Hello, i think a lot of people are having problems with choosing a GPU for a potential running of photoscan (besides other things), so they read GPU reviews on various sites... The problem is that these reviews do not include agisoft photoscan as one of their benchmarks, but i have noticed that a lof of reviews include a benchmark called "LuxMark 2.0" which uses OpenCL (as agisoft does during dense point generation). There are a lof of tests of a lot of GPUs where you can find the LuxMark 2.0 score for each card...

My question is: Is agisoft photoscan really comparable with LuxMark 2.0? Thus if a GPU X has 150% performance over GPU Y in LuxMark 2.0, is it safe to say that GPU X will have aproximatelly also 150% performance over GPU Y in Agisoft photoscan?


(Yeah there are some benchmark results on this forum but they do not include all cards etc.)

41

General / Re: Benchmarking a GPUs

« on: February 18, 2015, 12:47:36 AM »

Is it still true, that ATI cards offer in the same price range significantly better performance than NVIDIA cards?

For example i believe GTX 960 and R9 280 (and R9 280X), are almost in the same price range and their performance in games is similiar, but due to something (ATIs better drivers in openCL), these ATI cards offer huge performance increase in comparison to NVIDIA cards in the same price range?

Is it true still please?

42

General / Re: CPU and GPU benchmarks

« on: February 18, 2015, 12:39:55 AM »

This is quite amazing compare to Annadtechs tests!  You are 3 times faster than there fastest rig tested.

Do the anandtech test cover also GPU performance in agisoft? I found the cpu performance graphs, but i didnt find any graphs for gpu performance, it seems they are testing GPUs only in games, is that true or did i missed the graphs/tables?

43

General / Re: Graphic card: GTX 960 or R9 280 ?

« on: February 18, 2015, 12:30:26 AM »

Hello JohnyJoe,

According to the specification info, GTX 960 would be about 2 times slower than GTX 980, and R9 280X about 1,5 times slower than R9 290X.
For the cards that are mentioned as a reference you can find quite a low of performance output info in the benchmark thread.

Thanks for reply, and is it true, that otherwise similiar cards (performance wise in GAMES) will have huge difference in photoscan, because ATI has better opencl implementation, and even if some ATI graphics card has lets say performance of 80 fps , and some NVIDIA card has the same performace in the game 80 fps, the ATI CARD will be much faster in photoscan than the nvidia one?

44

General / Re: Graphic card: GTX 960 or R9 280 ?

« on: February 17, 2015, 11:19:08 PM »

GTX780 if your price range is around $200-250 and you dont mind used cards. I just picked up 4 for $800 just last month.

Thanks for reply, i will look into that, but if i dont want a used card? which one from those 2 would you recomend me?

also is is still true (new drivers might changed something or ne version of agisoft as well), that ati cards in the same price range are generally faster (a LOT faste) in agisoft, then nvidia cards?

45

General / Re: Dealing with DOF

« on: February 17, 2015, 10:37:23 PM »

Mesh generation is done from the point cloud, so the photos will have no influence at that stage but the texture would probably be affected  as the image is projected onto the object based from the camera position

Interesting...

1) So blurred (out of focus) areas are not generally a problem IF the parts of the model that are out of focus are IN focus (sharp) on other images of the subject with a good overlap? And this applies to dense point cloud generation and mesh generation?
Is this true...? So for dense point cloud generation and mesh generation there is NO need for focus-stacking in special programs?

2) But for texture generation out of focus areas even only in SOME photos, might be a problem, since photoscan can take this areas in those photos and make a texture out of them (not realizing, that the parts of model in this texture, might be blurred)? So focus stacking is usefull mainly only for texture generation step?

Do i get it right?