Welcome to the new 8mm Forum!
The forum you are looking at is entirely new software. Because there was no good way to import all of the old archived data from the last 20 years on the old software, everyone will need to register for a new account to participate.
To access the original forums from 2003-2019 which are now a "read only" status, click on the "FORUM ARCHIVE" link above.
Please remember registering with your first and last REAL name is mandatory. This forum is for professionals and fake names are not permitted. To get to the registration page click here.
Once the registration has been approved, you will be able to login via the link in the upper right corner of this page.
Also, please remember while it is highly encouraged to upload an avatar image to your profile, is not a requirement. If you choose to upload an avatar image, please remember that it IS a requirement that the image must be a clear photo of your face.
Thank you!
Yes Final resolution (F) is the result of Lens (L) and Sensor resolutions from the inverse formula 1/F = 1/L + 1/S.
After all if a line is resolved and falls on two cells of the sensor it registers on both and that halves the resolution seen in the scan.
I don't fully understand the formula, but it was the way that physics textbooks and photo magazines explained resolution on film and why it was lower than both the lens resolving power and that of the film. I think it works on the eye's detection of the contrast difference between the black line and the separating white one. As they get nearer to a mid grey you can't detect the line against the background and that is the mathematical formula for it. (Too much information, I know - I must be suffering from Covid lock-down boredom )
Hi Bruce and Brian,
Thanks for the comments. Here are a few my comments and possible answers to the dilemma. Have not spent too much time so take it with a grain of salt.
Bruce's input - The factor of 2 in Nyquist formula comes from the sampling theory. The nyquist frequency equals twice the max frequency of the system that is being sampled. For example in audio, if your max frequency of the speech is 4000Hz you have to sample it at 8000Hz sampling frequency. You can say it is one sample for the sine wave positive and one sample for the negative.
In our case it is not signal that changes with time but it changes spacialy but the reasoning is the same. The lens produces dark and white lines and we need 1 pixel per dark and one per white line. So the density of the pixels has to be twice the density of the dark white lines.
But there is an issue and we know this well, you cannot sample at the nyquist frequency because you will get aliasing (digital artifacts). That is why the density of the white dark lines has to be lower which lowers the lens requirements. Different factor can be used but Bruce found 2/3. And that sounds reasonable but could be different for different sensors.
And Brian's input is valid... 1/L is lens degradation and 1/S is sensor degradation and when combined the total degradation is the sum of individual degradations
DF = DL + DS
and conversly
1/F = 1/L + 1/S
But back to lens selection. The point is that the 12mm and 16mm lenses are similar quality and matched to the sensor. So going with a better lens for the test camera will not help. 16mm will give some improvement over 12mm because the number of pixels increases.
Hi Stan, thanks for the clarification, I understand some of the above. Your previous calculations seem to prove that these lenses (12mm & 16mm) are similar in quality (pixels per mm) even though one is rated at 5M pixels and the other at 10M pixels. I assume this has something to do with the image circle size of the lens (1/2" and 2/3"). I will continue to read and inform myself.
Regards - Bruce
Yes, it is to do with the size of the sensor Bruce.
Not really sure how we can leverage from the new camera. Possibly somewhat better final rez because the images have more pixels. But that does not justify the extra cost.
I ordered the new USB cable from TIS and will see if that improves the frame rates.
Not really ready to upgrade the system to 10 bit.
Regards,
Stan
Thank's for your help Stan finding the problem with the perpetual rewind.
That was a lot of detective work. You mentioned pin 8 a couple of times. I didn't know which was pin 1! There is no marking on the chip (msp430g2231) other than a recessed circle on one end and recessed half slot on the other. The documentation does not indicate pin 1.
Eventually with the schematic and the board drawings, I found pin 8. The problem was a wire on the rewind switch that had fatigued and broken loose. I would have found that earlier, since I knew it was caused by something I did, but the wire was held in place by the heat shrink tubing!
Too much handling and the switch panel was either hanging by the wires or clamped to the front face.
For now Hawkeye is working and closed up, should be safe!
5.1.1.2 12-Bit Packed Bayer Raw
This format transmits data using 3 bytes for each pair of 2 consecutive pixels.
I'm still puzzled by the file sizes and color depths listed in the IC, of which the Technical Reference Manual for the 37ux226 makes no mention. Only RGB_64 or Y8.
Sounds like the choice, for IC, is only 8bit or 12bit. That is pretty a significant jump in color from 256 to 4096 shades per channel. How much of that can the sensor see? I can't say without a test case.
If I wasn't still shell shocked about removing the camera, I'd shoot a light gradient on the wall and look for a continuous tonal range in the tiff.
Or is this about frame rate? That I can test, if I haven't already. I'm using a usbC, that should be usb3 too. I have run at every speed. Should I slow the shutter to a certain value?
I did not try do sustained captures at RGB 64. I was most likely using RGB 24. Both the single frame test tiffs I captured RGB24 or RGB32 show as 8bit in file properties even thought the file sizes are different.
I received my built-up Hawkeye 2 weeks ago from Stan and I must say what a great machine. My main goal as a historian was to be able to have clean, clear snapshots of the film that I was scanning as well as having the best film transfer possible. I have been using a Wolverine from its inception and must say that what Stan delivered to me was so far superior I give it 12 out of 10. Thanks, Stan for taking your time and effort in developing such an excellent machine. I don't think you could improve much more on the final quality of the snapshots or the video. Your product is more than I expected. Thanks again and keep up your good work....and Stay Safe.
Thank you very much Merv for the encouraging words. Very happy that the machine met and exceeded your expectations. It is not my work alone but several members here contributed to it and were actively involved in the way the final product turned out.
Glad you figured it out the takeup issue David. The front panel is not very robust. Guess in a hind sight I should have used a pcb for it.
Test camera issues:
Looks like the drivers support 8 bit or 12 bit modes only. Would be nice to have something in between.
The usb driver that we use uses the packed format only which is OK. That gives smaller file size since there are no padding bits.
The disadvantage is that packing and unpacking chews up resources making the file operations slower but on the other hand the transfer is faster.
I do see the difference in image properties for the old and new camera. The old one is 32 bit and the new one is 64bit and file size is larger.
The file size in the file properties cannot be used for size calculations because the tiff format uses LZW compression.
Don't think the light gradient would give us much new. The issue is that the display would have to have deeper resolution.
When I compare the 8 bit and 12 bit images in photoshop they look very similar. When I check the pixel values the 12 bit version is just a scaled value of 8. This is because in IC Capture my monitor is 8 bit and I adjust the 12 bit image to that. So the 12 bit whites all go into saturation because I adkust for vivid mid tines.
My hunch is that with a 10 bit monitor I would be adjusting the exposure reference different. The whites would not be as much saturated and the mid tones would still be vivid. That way you get better dynamic range. Similar reasoning goes for the shadows.
Have you checked if on your system if you can get the FPS listed in the manual?
8-Bit Bayer Raw
Width Height Maximum Frame Rate
4000 3000 30
3840 2160 30
1920 1080 30
640 480 30
I don't get even close to these figures. Maybe the new cable will help.
You can only capture in RGB32 (8-bit per 4 channels) or RGB64 (packed 12bit x 4 channels no padding, no 64 bit boundary alignment) with IC Capture in color mode.
Update:
Got the new cable from TIS and that made a whole world difference. The frame rates are 30FPS for 8bit and 20FPS for 12 bit. The IC Capture respons fast and there no more issue with sticky sliders.
David sent me one of the 16mm test lenses. Not top quality but good for testing the frames rates and image sizes.
It was free so I am not complaining
Anyways, with the longer standoffs and lens holder extension got a capture image at 2200x1800 pixel.
Will provide the details later.
A few more notes.
With the new cam the lens holder is made from metal and it helps dissipate the camera heat (just FYI for Kamel).
An extension is required with the 16mm lens used.
Had to rotate the camera by 90 degrees in order to get the cable to fit. No rotation required for 12mm.
"What is required to achieve the higher frame rates listed in the documents?". Instead of responding, they want to know the make and model of my computer and do I have a usb 3.0?
I don't. I thought that all usb C was 3.0, it's not. I tried this with a newer computer and it reports that my connection is really only 2.0.
Beyond all that at 4kx3k at 64 bits, there are capture problems with the image. I was able to capture 64bit at 2048x1536 without problems at 1 FPS.
The list grows at the lower size to include 1-7.41 FPS.
That's enough for me but, since my long cable is attached inside of the housing, I think I should should get a usb3.
Glad that worked for you. If you need a lens extension I have those coming out of my ears. I ordered some from China in late December, that were reported lost. Having trouble with my printer, I ordered more in May. They both showed up 2 weeks ago. They are aluminum, 1 cm long with abt 6mm of spacer each.
Sure David. Really appreciate your help.
Now I have to complete one Hawkeye board for myself because the one I have is a butchered V11. Your bag of spare parts will be put to good use
I will also order the 16mm lens that you have and then run a test clip. Will see what we get.
USB type A SS connector has different pinout than regular A type connector.
It is still backwards compatible with USB2 but obviously you get only usb2 data rates. The cable that you use has to have usb3 symbol stamped on it and SS written next to it. The same goes for your usb port on the computer.
In order to get the usb3 connection on Hawkeye at the back I would have to redesign the USB area of the board.
Stan and David, interesting progress, great news about the TIS USB3 cable achieving rated speeds, you had to trim a considerable amount, thanks for keeping us informed. What are the FPS speeds reduced to when connecting to USB2 port with the new cable.
Is the lens available in the US or does it have to be purchased from China.
Bruce, you might find this useful, it's certainly interesting.
Edmund recently sent me their physical catalog after I bought the pocket resolution card for $6!
I would have been happy with the online catalog. https://quad2.mydigitalpublication.c...i=645491&p=174
Great stuff, for a lot of money!
I'd get Stan to weigh in on the 25mm!!
If you could get the 25mm lens to work you might find you will need to build xy positioning for the camera, as It won't be centered. It might have a greater min. focus distance than the 16mm. They are both spece'd at .1m minimum. Longer lenses tend to need more. I still over drilled the main circuit board mounts for a small rotational adjustment.
Switching from Super8 to Reg 8 films will need to have an adjustable option in the y axis.
The Lens from cnAICO comes with an IR cut filter. ScorpionVision will add one for extra cost and time. I never found anyone else selling the 16mm.
Kamel said his lens is from IS at f2.8. There is plenty of light for f5.6 if using HDR. For daylight scenes my capture shutter speed was just over 1/20sec. That should be enough. With HDR, my shutter speed goes up to 1/60sec. and more. That's without adding brightness or gain.
Found these on amazon. I'll just let the usb C hang out the rear vent. I have a laptop with usb C input or a PC with usb A
Tweet
)
Comment