A few months ago, I ran a blog post about how important it is to do camera tests. It was part one in a four part series. It discussed the fact that you should not just surf the web for forums and camera review blogs, but go out there yourself and test your camera. I will provide the playbook so that you can understand my methodology when I test new film stocks and new digital cameras for my features.
I treat each new digital camera that comes out as a new film emulsion, one that needs to be tested before I can use it to help tell my stories. Finding out its strengths and its weaknesses is all part of the process. There are three parts to the process. The first is to find the sensitivity of the sensor. Part two is the latitude of the sensor. Part three is taking your camera on location to put it through the paces in different lighting scenarios, both day and night.
You should do this even if you have been shooting with a camera that you have owned or rented for some time. If you have not done these tests, please do the diligence to test it because you will learn something new.
Preparing a Slate
I prefer to use a large whiteboard. Put all of the info necessary to educate you, your crew and your director on it. We were shooting four cameras on this test, so we divided our whiteboard into four sections.
Whiteboard slate with four sections
Slate information for the ISO test
This is something you do not want to rush. Plan carefully because this is your roadmap. Without good notes and driving directions, you are going to get lost.
Understanding the Power of Your Sensor
ISO lens cap tests are the best way to understand how far you can push your sensor and how much noise your camera exhibits at each ISO. It will also show whether there is a native ISO, which means the camera’s lowest noise ratio. The Canon 5D had a native ISO at 160, 320, 640,1250. These specific ISOs should contain the lowest noise.
5D Mark II ISO menu
Rolling Your Camera in the Darkness of a Lens Cap
Derek Johnson and Laura Murphy will take you through how to perform these tests. Accuracy is paramount, so take your time so that you do not make mistakes.
Setting the ISO on the Canon 1DC
Adding ISO information onto slate
Recording five seconds of slate
Putting lens cap on and recording 20 seconds of black
After shooting the all the ISO’s on black. Take the footage into after effects to push the footage to bring out the Noise. We have had the best results by using the Exposure Effect in Adobe After Effects CS6. Adjust the Exposure level and offset to where the grain starts to colorize and you can recognize the patterns.
Screen grab of 1DC Canon log ISOs
Screen grab of 1DC neutral picture style
Adding Skin Tone to Your ISO Tests
It is so important to not just shoot these lens cap tests but to add a model to showcase skin tones and illustrate how the noise reacts to shadow areas of the face, etc. Using a grey card is essential as well because this shows noise very well. A color chart is good too. I try to put them on the same focal plane as our model so that focus is good across everyone.
Example of key and fill ratio
Part three next week will show the latitude tests which are not about charts, waveform monitors and vector scopes. Stay tuned!
Tags: Canon 1DC, Canon 5D, spon-canon
This entry was posted
on Thursday, December 13th, 2012 at 9:39 am and is filed under Camera, Configuration, Pre Production.
You can follow any responses to this entry through the RSS 2.0 feed.
You can leave a response, or trackback from your own site.
Add to Google
Shane,
Did a lot of this in a studio setting while I was a contract photographer for the USAF. One cannot stress how important this is. I still get bonked.
Coupled with the bald blue sky test at small apertures to see where smooch shows up is a good habit to get into.
Thanks for this
Be well
Laurence
Laurence Zankwski. Thanks so much for the kind words and support.
I think it’s misleading to say that these are “native” 160-base ISO. There’s been a lot of testing that shows that although “cleaner”, they are not “native”.
http://shootintheshot.joshsilfen.com/2010/05/13/canon-hd-dslr-native-iso/
Would love to hear your thoughts on this Shane.
brentwgraham, I checked out the website and write up which was very through and well written. The push and pull is all right on. I also love the fact that you get a little more dynamic range. I think the biggest take out of all this is that Canon has cracked the fix pattern noise issue on the 1DC and delivery noise patterns that look and feel like film grain and that there is incredible low noise.
Hi Shane,
thank you for your articles.
I didn’t understood your reply tho Josh’s article, in wich he states that ISOs 160x looses about 3/10 stops of Dynamic Range.
However in my reply to his article, I told that I can’t perform a latitude-emulsion test my T3i, because I don’t have a color chart neither a Zone Scale.
I could only test emulsion on my T3i (Magic Lantern allows me to dial ISOs 160x).
So I suggest you to perform some test on Canon DSLRs to check if those 160x ISOs looses 3/10 stops of d.r.
Carlo
Shane,
I consider the brent’s point very important. Shall not be forgotten. Native ISOs, loss of DR …
Thanks for sharing knowledge.
Great stuff. If only we were given the time on our canon 5D we are using on a US/UK reality series I am working on. Looking forward to next weeks tests.
Gavin Boyd. Thanks so much for the comment and support.
[...] http://www.vimeo.com/55472434Ce test assez simple donne une idée assez précise de la qualité d’image de cet EOS, qui filme en 4K, ainsi que de sa sensibilité en basses lumières. Et là, c’est assez bluffant à première vue, même à des niveaux d’iso très élevés, on ne sent quasiment pas de grain. A lire en complément, l’article très détaillé que Shane HURLBUT consacre à détailler son test sur son blog. [...]
Thanks for sharing this. I’ve got a couple of questions.
1. Does white balance matter on the noise tests? (I’m guessing no.)
2. I see that you did at least two sets of noise tests on the 1DC for two different picture settings. I’m guessing I need to do tests on every picture setting I may want to use. What’s your position on that?
3. How do you white balance for the skin tone tests?
4. How do you set exposure for the skin tone tests? Do you stop down, change the lighting or both? (I assume you keep the same shutter angle.)
Thanks again. It was a pleasure meeting you in Austin at the Masters in Motion workshop.
Robert Shaver. Thanks so much for the kind words and support. Great meeting you as well Masters in Motion. It was a great event with so many passionate filmmakers.
1. The white balance doesn’t matter in noise tests.
2. We did noise tests for Canon Log, and the Neutral picture style because both settings have different sensitivities and noise.
3. The skin tone test was white balanced by eye to what looked the best.
4. I used a light meter to keep the exposure consistent through every Iso. Keeping the shutter angle the same, but dialing down the lighting, and T-stop on the lens to compensate for the increased sensitivity in the higher Iso’s.
As usual Shane .. You have given great info and details ..
Can’t wait next part ..
Mosalam. Thanks so much for the kind words and support.
Shane,
Thank you for this information it will be very helpful!
Q: Will using different lenses effect the noise when shooting real world?
Bill
Bill Hamell, you are very welcome. No, the lens has nothing to do with the noise level. That is all sensor.
Hi Shane, really hope you can advice me on this. I’ve tried this out with my new full frame Sony A99 and I hope I’m doing it right…
The A99 has monochromatic noise so it’s very difficult to see where its native ISO are. I’ve recorded every ISO and played it back on my 21′ monitor. Unless it’s the higher ISO at 4000 above, it just looks black. Which is good right?
What’s your take on this? Shall I upload the video file for you?
Hi, Derek my assistant will talk you through this, it requires use of after effects to get the noise to read. Thanks for the support
Thanks Shane, I know you’re in my country Malaysia right about now. Regret I can’t attend but I’ve seen your presentation at NAB. On behalf of us Malaysians, thanks for making the 19 hour flight round the world and still have time to answer blog comments :)
Hi Baron. I had the best luck with the Exposure effect in Adobe After Effects CS6. I would dial in the exposure and offset until it would bring out the noise and colorize the grain. Then I would apply that same adjustment layer to all the ISO’s to compare the noise.
Hi Derek, so that’s how you did it. I’ll go try it out with AE then. Thanks!
Derek, which iso did you use for the benchmark and applied to all others? Thanks
Hi Shane, Derek & Laura,
Did you you saturate the video of the test to emphasise the noise?
When I do this test with my Nikon D800, is just black with very little noise, even at ISO 6400.
Only if I really saturate the image I begin to see the noise.
Thank you very much for all the great content on this webiste !
Kind regards,
Marien
Marien. Thanks for the kind words and support. We used the Exposure effect in Adobe After Effects Cs6 to bring out the noise. Dial in the Exposure level and offset to bring it out, and then apply it to all the Iso’s for coparison.
I will buy the Canon 1DC for sure! Thank you Shane for all the info.
Juan, your so welcome. It is amazing cinematic camera.
I’m waiting impatiently for it to arrive; I have some projects lined up in which I want to use the camera, but I will shot some tests before production as you wrote here.
[...] See on http://www.hurlbutvisuals.com [...]
I didn’t see anywhere where you discuss shutter speed during this test or if it matters…..??
Lawrence Rose. All of the tests we’re shot at a 1/50th shutter. With the exception of the high speed tests that will be shown in part 4. At 24fps a 1/50th shutter is the equivalent of a 180 degree shutter on a film camera.
If I did this right my canon7d also is best at 160,320,640,1250. Testing at. 1080/30. 1/50 @ 5.6
Neutral I tested all ISO at +7.0 exposure in AE Could it be inherent to canon sensors?
David, yea I think it is
Hi Shane,
Thanks for this great information about tests.
Does this test tell us that the Neutral Picture Style is actually cleaner than the Canon Log? Of course the Log will have more latitude.
Thank you very much!
Eric, yes it does, much cleaner, but yes, less latitude