April 2, 2012

even with high ISO settings, you still need great light

Still having fun with the new Canon 5D Mark III (B&H), I met up with Elmira again in New York. Elmira is the model I used in my initial tests of the Canon 5D Mark III high-ISO performance. Being a delightful model to work with, I decided to use her again as a subject.

New York was cold on this day, so shooting indoors just seemed a lot more attractive. We went to Grand Central Station – a grandiose building, but with light levels quite low. Low enough that I was glad that I brought the Canon EF 35mm f1.4L (B&H) along.

camera settings: 1/100 @ f/2 @ 3200 ISO
Even with a high ISO like that, I had to use a fast aperture.

An approach that I strongly believe in though, is that “using the available light” is not random decision. It needs consideration of what your light is actually like, and whether it is flattering. What I did here was to pull Elmira towards a light source, so that the light would come in from an angle over her shoulder …

With the fluorescent light coming through the frosted glass on the door, I knew I had a large light source there .. and I used it. This way, I purposely controlled the available light falling on her. It’s a conscious choice.

Hopefully with the medium-wide angle of 35mm, and the fast aperture, it lends the image a certain cinematic look.

I did crop the image slightly to remove an unwanted highlight in the top right-hand corner. If anyone wants to look at the photo close-up, along with the other images from the Canon 5D Mark III high-ISO test, the high-resolution JPG is available as a download.  The photo is the JPG generated from the RAW file, via DPP.  Therefore it is as close to an in-camera JPG as you will get, while still allowing me to adjust the WB of the RAW file.

camera settings: 1/80 @ f/1.6 @ 3200 ISO
Canon 5D Mark III (B&H);  Canon EF 35mm f1.4L (B&H)

The same idea was applied here in using the available light. I scouted around until I had a place where the light fell on her in a flattering way.

I did play with the image a bit, using the RadLab action set. (Therefore the high-rez file isn’t available, since it won’t be a representative image of what the camera can do.) You can order the RadLab or the Totally Rad action sets, via this affiliate link.

You can order a Canon 5D Mark III via this B&H affiliate link.


 

{ 12 comments. } Add a Comment

1 Stephen April 2, 2012 at 4:24 pm

For the first image, where the model is leaning against the office door with the fluorescent light, did you set the camera WB to fluorescent, tungsten, or some other value?

I get confused when there are mixed lighting sources. I usually set the WB to the dominant light source.

Reply

2 Neil vN April 2, 2012 at 4:43 pm

I had the camera set to Auto WB. However, it is immaterial since I am shooting in RAW.

Neil vN

Reply

3 Jerry April 3, 2012 at 3:45 am

So, Neil, are you just testing the Mark III a lot for the blog’s sake, or are you thinking about going back to Canon as your dominant camera again?

Reply

4 Neil vN April 3, 2012 at 3:48 am

Jerry .. nope, the camera is on loan from B&H for review purposes. And since this is one of the major releases in the camera world, I was very interested.

Neil vN

Reply

5 ugur April 3, 2012 at 3:51 am

Dear Neil,
Is the noise level related with ambient light for the same iso level? I mean that “high iso performance level photos” (first samples) are taken with great light (1/400, f4, 12800 iso), whereas your last samples are taken with less light (1/80, f1.6, 3200 iso) There are 3-4 stops differences i think. If you compare both of them according to the noise levels, is there any quality differences in raw files? It’s difficult to understand from jpegs. Thanks.

Reply

6 Neil vN April 3, 2012 at 4:03 am

Ugur, a direct comparison here isn’t easy. Not because of the differences in light levels – after all, correct exposure is correct exposure – but because the color spectrum is so different between the two scenarios. So the noise levels may appear different. For all the images I’ve shown here and in the previous article about the Canon 5D Mark III, I kept my exposure adjustment of the RAW file to zero. Nothing. Once you change the exposure, the noise appears differently.

With the photo in the previous article that was shot at 1/400 @ f4 @ 12800 ISO, there was enough light for a lower ISO .. but to check what the noise looks like (for that particular lighting situation), I decided to go to a higher ISO. It’s still a valid evaluation of what the high-ISO noise looks like.

I’ve seen criticism on the forums because that specific image on the previous page (1/400 @ f4 @ 12800 ISO), could’ve been shot with lower ISO … but I do think that is a trivial objection.

That there was more light should have little effect, especially compared to the spectrum of light we’re shooting under.

Neil vN

Reply

7 emopunk April 4, 2012 at 10:45 am

Hello Neil; when do you plan to share some impressions on the new D800/D800E as well? I think a lot of us are really waiting for that insight too.. Greetings!!

Reply

8 Neil vN April 4, 2012 at 11:05 am

emopunk … I don’t have the D800 (or D4) bodies yet, due to a hiccup with Nikon Professional Services. I thought I was an NPS member, but apparently wasn’t. Long story. So I am without those cameras at this point.

Neil vN

Reply

9 Duane Dale April 4, 2012 at 11:34 am

A major take-home for me from this post is: keep an eye out for natural-occurring softboxes, such as the frosted glass door with light coming through it. Not a surprise that this message comes from Neil, who has given us the “virtual softbox” in the form of bounce lighting from an on-camera flash, shaped by the black foamie thing.

Reply

10 emopunk April 4, 2012 at 4:17 pm

Hi Neil. Yes I read some of your posts about the troubles with Nikon NPS for the latest cameras. I hope they sort it out as soon for you. Regards!

Reply

11 Gene Fama June 24, 2012 at 11:06 am

Nice article! I think you have a typo in paragraph seven where you say “Canon 5D Mark II.”

Reply

12 Neil vN June 24, 2012 at 3:21 pm

Thanks, fixed now.

Neil vN

Reply

Leave a Comment

{ 1 trackback }

Previous post:

Next post: