Post subject: Stack software, comparative speed, optimizations and others
Posted: Fri Jul 19, 2013 11:44 am
Valued Contributor
Joined: Mon Feb 11, 2013 6:24 am Posts: 168
Hi all,
In recent weeks I have been busy but now I've had a little time I will present a report I created from stacked processing times of the most common programs and optimizations, quirks, etc. First I is to describe the testing platform.
Hardware Computer:
- CPU: Intel I5-2500 Quad core 3.30Ghz - MB: Asus P8 H67-M Pro - RAM: 8 Gb DDR3 1.333Mhz Dual Chanel - HD: SSD OCX-Vertex2 in SATA 6Gb/s - GPU: AMD Radeon HD 6850 4Gb
Hardware take pictures
- Focus rail: MacroRail Professional - Camera: Canon EOS 600D 18Mpx - Microscope objective: Nikon BD Plan 5x - Others: bellow & adaptors
All tests were performed on Windows without virtual memory and photographed images were executed and performed from HDD SSD to minimize the burden of the applications and / or photographs.
The times presented are solely the load time and process stack programs.
Instead of using a mineral I used the common fly killed a month ago. The images that are presented neither edited or retouched from the software mentioned above or from an editing program such as Photoshop, are presented as they were generated by stacking programs.
DESCRIPTION, RESTRICTIONS AND PROBLEMS FOUND
Before presenting data, images and displays certain characteristics that I used in the programs:
HELICON FOCUS. Good stacking program controlled processing of the lights when there are certain reflexes without editing. Apparently all loaded into memory to process images so if you do not have Windows virtual memory dynamically defined as we run out of memory. When loading images in memory the process is muy quick.
Stacks do not allow more than 255 images so if you want to stack more pictures of this amount have to use sub-stacks. For this reason, the 400 photos has had to use a sub-stack of 202/2. That is, 202 pictures per stack with 2 overlapping photographs.
COMBINEZP. Good free program but where there have been problems when stacking hairy insect photographs. The modes "Do Stack", "Do Soft Stack" and "Pyramid Do Stack Weighted" not valid. At the end of the stack is chosen by "Pyramid Weighted Average" method to get a proper result.
ZERENE STACKER. Program stack of great quality, especially in the treatment of hair, which can handle large volume of photographies without sub-stacks, but slower performance than the rest. It is advisable to make a Preview Stack of large stacks.
Macrophotography. A program to control the camera, the MacroRail, the stacking process / sub-stack, indicate the size of work or crop the image before stacking, retouch images, etc. It supports the focus rails; MacroRail, Macrocarril, Stack Shot or any approach lane to use a motor Unipolar or Bipolar with simple or more complex electronic CNC type.
PROCESS:
I use a EOS 600D camera pictures at highest resolution JPG (18Mpx) = 5,184 W x 3,456 H = 17,915,904 pixels.
On each test 100 and 400 photographs have been stacked as follows:
- Stacking original size picture. - Stacking cropping the size of the head of the common housefly. - Stacking ½ the original size of the picture. - Stacking ¼ the original size of the picture. - Stacking 1/8 of the original size of the picture. - Sub-stack with the original size of the picture. - Sub-stack cropping the size of the head of the common housefly.
For the stacking test 100 photographs with a spacing of 6.6 microns to total displacement of 0.66mm. For the 400 photographs, there has been a separation of 1.6 microns with a total displacement of 0.66 mm.
Here I present a table of the results I obtained. The times are in the format MM: SS.CC, where MM is minutes, SS is seconds and CC are hundredths of a second.
Looking at the table, we can see that the sub-stacks system is not recommended for small stacks of less than 100 pictures. The more pictures do more interesting is the use of sub-stacks. Especially in programs that loaded photos in memory like Helicon and Combine.
Moreover, we can see that the response speed of Helicon Focus is much faster than the other programs.
Also, sometimes piled whole picture when really only want a part of this, then the option to cut before stacking is important to save large delays.
Originally we are working with an image of 5,184x3,456 pixels, but if we tell you that we generate a stack of 1/4 of this size (1,296x864) is more than enough to see how we have our image. In fact, many times we do not use if you want this resolution to be presented at a forum as in this case. However, the stacking time being considerably reduced up to 8 times faster than the standard stack.
Sub-stacks for high number of pictures; if you use sub-stacks in Helicon Focus can gain until 30% of faster, depends of situations. In Combine ZP the difference is very little. In Zerene, the use of sub-stacks increments a little the final time.
Gradually I will publish the pictures of times achieved, working screens and insect pictures.
I feel my English is so bad, if any data is not understood please ask me.
Best regards, Oscar.
Last edited by ofarcis on Fri Jul 19, 2013 11:54 am, edited 2 times in total.
Post subject: Re: Stack software, comparative speed, optimizations and oth
Posted: Sat Jul 20, 2013 12:28 am
Valued Contributor
Joined: Mon Feb 11, 2013 6:24 am Posts: 168
Hi to all,
I've posted all the information and images relating to the tests I have done. If you have any comments or want to expand any of the information, will be welcome.
I repeated the tests four times and times of each other vary little.
Post subject: Re: Stack software, comparative speed, optimizations and oth
Posted: Sat Jul 20, 2013 8:36 pm
Valued Contributor
Joined: Mon Feb 11, 2013 6:24 am Posts: 168
Hi to all,
In the first post i talk about the CombineZP problems with high number of phtographs to process:
Quote:
COMBINEZP. Good free program but where there have been problems when stacking hairy insect photographs. The modes "Do Stack", "Do Soft Stack" and "Pyramid Do Stack Weighted" not valid. At the end of the stack is chosen by "Pyramid Weighted Average" method to get a proper result.
I solve this problem with two ways; one of the is to use Sub-stack another is to change to Pyramid Weighted Average" method
You can see an example of the problems that present when i use these modes that present problems:
Post subject: Re: Stack software, comparative speed, optimizations and oth
Posted: Mon Jul 22, 2013 10:26 am
Valued Contributor
Joined: Fri Jun 21, 2013 8:14 am Posts: 108
If you can put a zip file with the individual images up somewhere for me to grab, I'll run a test with Photoshop's auto-align/auto-blend functions and see how it performs. I use it a lot.
Post subject: Re: Stack software, comparative speed, optimizations and oth
Posted: Mon Jul 22, 2013 10:54 am
Valued Contributor
Joined: Mon Feb 11, 2013 6:24 am Posts: 168
Hi Jester,
Have a problem... The total size of photographs in JPG format for the Stack 100 images need 335 MB of space. The Stack of 400 photographs have 1,30Gb. I can't put a zip file of these.
The unique way is to upload in FTP server to you!!! But need a lot of time. Are you sure?
Another possibility is to say me wath process are in Photoshop and I replicate and measure in time.
Post subject: Re: Stack software, comparative speed, optimizations and oth
Posted: Mon Jul 22, 2013 1:19 pm
Valued Contributor
Joined: Fri Jun 21, 2013 8:14 am Posts: 108
ofarcis wrote:
Hi Jester,
Have a problem... The total size of photographs in JPG format for the Stack 100 images need 335 MB of space. The Stack of 400 photographs have 1,30Gb. I can't put a zip file of these.
The unique way is to upload in FTP server to you!!! But need a lot of time. Are you sure?
Another possibility is to say me wath process are in Photoshop and I replicate and measure in time.
Regards, Oscar.
I've got a 300 MB FIOS connection, so 1.3 gig is no problem for me. Do you have an FTP server to put them on?
Post subject: Re: Stack software, comparative speed, optimizations and oth
Posted: Mon Jul 22, 2013 1:33 pm
Valued Contributor
Joined: Mon Feb 11, 2013 6:24 am Posts: 168
Hi Jester, I upload in this moment the 400 images but probably need some hours to upload. When is done I advise you for PM.
Hi Tim, the Zerene quality is best in some circumstances, I use in the last year and the quality results is very good especially mineral needles (i don't know if is the correct translation of agujas, acículas). But tends to burn bright areas as reflecting on facets in minerals and this is a big problem, in this case, is better the Helicon Focus method B.
For example, for this mineral or gems or whatever you want is more better the Zerene Results:
Post subject: Re: Stack software, comparative speed, optimizations and oth
Posted: Mon Jul 22, 2013 5:05 pm
Moderator
Joined: Thu Dec 13, 2007 3:42 pm Posts: 4091 Location: the Netherlands
Capice! Interesting... Jester's suggestion is certainly worth a try too.
What do you think about the number of photos one uses in a stack? I've been living under the assumption that less is better but here i see you using copious amounts of pics... What's your experience teaching you? The more the better?
Post subject: Re: Stack software, comparative speed, optimizations and oth
Posted: Tue Jul 23, 2013 1:30 am
Valued Contributor
Joined: Mon Feb 11, 2013 6:24 am Posts: 168
Hi Tim,
Some people that stack insects need to take 1.500 or more photos. But my experience say that less is better ever of the depth of field is completely covered.
In gems, don't need a bigger number of photos, really only need to covered the depth of field. This is necessary to have an good stack.
Exist some formulas to calculate the correct depth of field to sure it's correct:
DOV = 2 * CoC * f * ((m+1)/(m+m))
Another easy way is to make an Excel table or similar that you appoint some important data like:
1.- Magnification used 2.- Field of view 3.- objective 4.- Ideal number of Photos per millimeter etc.
Take some sequence images and check for the displacement if the depth of field is covered.
For example, if you use an 10x magnification, and displacement the focus rail 20 microns and take 10 photos. Can stack and check if all is covered or exist blurred slots. In this case, the displacement need must be less, for example 15 microns and test again. When the field is overlapped between photos about 30%, this is a good displacement.
Next only nee to see what is the total displacement and calculate with the correct displacement. In the example, if you need to displacement to cover all the depth of field 1mm can divide 1mm / 15micros = +- 67 photos per millimeter. This is the 4º point that i mentioned (Ideal number of Photos per millimeter).
Now, when you use ever this objective at 10x, ever need to use 15 microns between photos, or 67 photos per millimeter.
You can see in the next image a bad coverage in the left, and good coverage in the right:
Post subject: Re: Stack software, comparative speed, optimizations and oth
Posted: Thu Jul 25, 2013 9:15 am
Valued Contributor
Joined: Fri Jun 21, 2013 8:14 am Posts: 108
I have not tried the 100 stack yet, but Photoshop CS6 has yet to successfully complete a 400 stack (it crashed after 45 minutes last time I tried, on a system with 32 gigs of RAM.)
Users browsing this forum: No registered users and 2 guests
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot post attachments in this forum