Camera Resolution Test

Discuss about cameras, projectors, calibration panels, turntables etc.
Oteck
Posts: 82
Joined: 11 Nov 2016, 21:21

Camera Resolution Test

Post by Oteck »

I've been testing out a telescopic camera i've got recently and see how it compares. This is done on a 240mm calibration and so far the gain is somewhat better. The texture grab is also much more accurate on monochrome and the details along the edges are more sharper but it could be much sharper with a larger C-mount lens as I out resolve the current one which is only for 1/2.7" vs 1/1.4". It's also vignetting so that's not helping lol

Exposure = 1/16s
K132 = max brightness
FPS limit = 5 (usb2.0 limitation)

240mm 1920x1080 Colored
C1 1920x1080.png
120mm 1920x1080 Colored
C1 1920x1080 120mm.png
240mm 3096x2080 Monochrome
ASI 3096x2080.png
User avatar
Micr0
Posts: 586
Joined: 15 Nov 2016, 15:20
Location: New York City

Re: Camera Resolution Test

Post by Micr0 »

What's interesting (and Jeff seemed to see this too) is that when you get to the much higher resolutions, the edge definition is better but the surface texture seems to get rougher. BTW can you try it with an exposure that is an even factor of the frame rate of the projected frame rate?
µ
Oteck
Posts: 82
Joined: 11 Nov 2016, 21:21

Re: Camera Resolution Test

Post by Oteck »

I couldn't do it with the mix camera setup as the C-1 usb2 is the bottleneck. once i remove it i'll be limited on the fresh of the projector on my k132 as the usb3.0 allows me to go 60fps (10bit) or 30fps(12bit) in the highest resolution in which i'll start to see banding. I tend to shoot slower to allow more light to gather so i'm aiming at 20-25fps depending if i'm shooting black stuff.

The lens is pretty much preventing me from getting any sharper image so i have some 1/2" c-mount on the way to take advantage of the higher MP as the CS i have are only good for 3mp while the sensor is 6.4mp.

Also note those shots were done at 1 sharpening and very little on the qc check

https://astronomy-imaging-camera.com/pr ... 78mm-mono/

10bit ADC/14bit ADC
3096×2080 60fps / 30fps
2560×2048 62fps / 31fps
2048×1080 116fps / 58fps
1280×960 130fps / 65fps
800×600 204.7fps / 102.3fps
640×480 253.1fps / 126.5fps
320×240 479.7fps / 239.8fps
Oteck
Posts: 82
Joined: 11 Nov 2016, 21:21

Re: Camera Resolution Test

Post by Oteck »

I think im hitting more issues than i though, my pc can't process the images fast enough unless this is how it suppose to be.
If set to 30fps at quality (60 patterns) it should take 2 seconds to complete no? unless it's not a pattern per frame as it's taking 24sec to complete a scan

Anyways i tried to do 30fps, 3096×2080, 1/32s on this slow pc and it's becoming very sensitive to vibration.
2017-04-17_19-50-17.png
Curiousjeff
Posts: 121
Joined: 16 Nov 2016, 22:31

Re: Camera Resolution Test

Post by Curiousjeff »

That's damm good !

Do you have fine adjustement with this camera or are you just using a standard 1/60 shutter speed ?

I find your surface quality excellent. No waves.

Regarding speed, I would not be concerned about it.

My camera has 121 fps, my computer is very fast with 64GB of ram, etc.. and still I don't get that frame rate during aquisition.

I am talking about scan time (patterns being projected) and not the extra time to process.

Usually, a scan at full camera resolution 2048 x 1536 with about 22 patterns takes 6-8 seconds.

If I set the resolution to fullhd, it goes down to 4 seconds.

With my HDMI capture card and the GH4, David says I have 60 fps, but it is even slower.

If you have good speed and you are not scanning living items, I would not worry about it.

You mention 60 patterns. Are you scanning horizontal and vertical at the same time ?

Jeff

P.S.

From a previous post:

David 5: single camera
Resolution of camera: 1920 x 1080
Profile: Speed
Number of patterns: 21
Time of scan for patterns: 2 seconds
Curiousjeff
Posts: 121
Joined: 16 Nov 2016, 22:31

Re: Camera Resolution Test

Post by Curiousjeff »

Since I had a bad experience with the PointGrey driver (you cannot use two cameras within Daivd 5), you should check the software compatibility with the manufacturer if you intend you use two cameras.

My 2ct.

Jeff
User avatar
Micr0
Posts: 586
Joined: 15 Nov 2016, 15:20
Location: New York City

Re: Camera Resolution Test

Post by Micr0 »

Oteck wrote: 18 Apr 2017, 04:01 I think im hitting more issues than i though, my pc can't process the images fast enough unless this is how it suppose to be.
If set to 30fps at quality (60 patterns) it should take 2 seconds to complete no? unless it's not a pattern per frame as it's taking 24sec to complete a scan

Anyways i tried to do 30fps, 3096×2080, 1/32s on this slow pc and it's becoming very sensitive to vibration.
2017-04-17_19-50-17.png
The time it takes to capture a scan (sequence of patterns) depends on a host of factors. The David software waits to "see" a complete pattern before is projects the next one. So even if you have a fast camera but the projector and camera shutter are out of sync then it will take longer to go through a scan sequence. This is one of the reasons a global camera shutter is better. I found that when I was playing around with the frame rate on my video card that all of a sudden the scan sequence would double in speed. By finally tuning the frame rate I got the time to capture a Default quality sequence down to 1.8 sec. BTW I almost never use the "quality" setting. The scan sequence does also depend on camera speed, projector speed, USB speed and computer speed, so obviously it isn't just one thing.

One other note: Most cameras sold for astronomy don't have an IR filter over the sensor. This could be contributing to your surface roughness.
µ
User avatar
avogra
Site Admin
Posts: 116
Joined: 11 Nov 2016, 17:27
Location: Munich, Germany
Contact:

Re: Camera Resolution Test

Post by avogra »

concerning scan times: i think david uses more than one frame to take one picture. If I understand it right, david changes the pattern. then it discards frames until image content has changed. this is to account for possible delays of beamer, camera, windows, etc. after that it discards at least one more image, because the last image might contain parts of the last pattern and parts of the new pattern, as camera and projector are not synced to each other. so in the end it needs at least 3, usually more frames to take one picture.
If scan time is crucial, you can tweak this in david preferences. i think, only if you link your camera to the beamer and measure the delays of david -> beamer and camera -> david you can get away with one pattern change each frame.
Oteck
Posts: 82
Joined: 11 Nov 2016, 21:21

Re: Camera Resolution Test

Post by Oteck »

I did play with high speed 60fps but it's too fast that david can't calibrate it and i had to drop the resolution quite abit to get there. at 800x600 it was quick so thats why i figured the cpu is the bottleneck. There is a faster pc I can use but that means taking my gaming rig to the work shop and it's big and heavy.... The 60 pattern is the quality mode which has the highest scan time, I never actually used anything lower but i ought to.

Typically id be fine for anything 30fps as my projector isn't bright enough to do it so it's best to take your time at it. The new lens is in order but I should invest in a better smaller scale calibration board as the one i got from david isn't to well made.... My tripod is also shaky when it's raised up too high and with all the weight im adding on it isn't helping the stability.

The zwo asi driver allows me moderate control so i can crop the sensor to what ever size if i needed a boost in speed. It would show in the setting what fps it can do which is very handy but so far it's fast and you can happily just let the fps counter do the work. I'm also looking at a IR/UV pass filter if it can help sharpen the scans more
Oteck
Posts: 82
Joined: 11 Nov 2016, 21:21

Re: Camera Resolution Test

Post by Oteck »

Micr0 wrote: 18 Apr 2017, 12:36 One other note: Most cameras sold for astronomy don't have an IR filter over the sensor. This could be contributing to your surface roughness.
I may look into an IR/UV block filter and hope it smooth out the roughness and get the surface like the 120mm scan. This surface should be matte and not really textured like that so im guessing the heat from the IR spectrum and the uv is giving me extra feedback
Post Reply