DSLR Training - Reviews

5D Mk III

Canon announced the new 5D MkIII on the 2nd March 2012. I am not going to go through all the new features on the camera, which you can see under the product news.

This document is to see how well the camera performed in the field and with our inhouse tests.

TEST RESULTS

We looked at the camera’s AF system, how well it tracks and focuses. The ISO, we compared it against the 5D MkII, 7D and 1D MkIV, and lastly the resolution. So let’s look first at the AF system.

AF FOCUS

Canon has had some bad press about their AF systems in the 1D MkIII, improved with the 7D and 1D MkIV but did not win to many critics over. They really stepped up their game with the 5D MkIII. When you look at the new specification they increased the dual cross AF sensors, they increased the cross type AF points and spread them further out towards the end focusing points and increased the total AF to 61. Now what is important to note here is that the dual cross AF sensors only work with lenses with a maximum aperture of f2.8 and faster. The 41cross type AF sensors work with lenses with a maximum aperture of f4 and faster. If you use a lens with a maximum aperture of f5.6 then you have 21 cross type AF point all grouped in the centre. See images below:

AF-available f2-8 colour with legend.jpg

AF-points available.jpg AF available f2.8 colour AF available f4 colour AF available f5.6 colour


The reason I am showing this is because it is important to know where the most sensitive and accurate AF points are. From our tests, you will see that we used lenses with all three maximum apertures. Remember, as you change lenses from f2.8 to f4, the AF points that was dual cross AF points now becomes cross type AF points and when you change to a lens with f5.6 as its maximum aperture, cross type AF Points change to vertical sensitive AF points (sensitive to horizontal lines).

We use two AF tests, one for static subjects and one for moving subjects. Lets look at the static subject, and what we wanted to see is if the focus point focuses at the point where you aim the focus point and how does the different assist focusing points influence the main focusing point. For the static test we printed a test chart from Norman Koren at http://www.normankoren.com. The chart is really to test lens resolution but we thought that this gives you a very nice chart with clear black and white lines that will create ample contrast to focus on. We used three lenses, namely EF 100 f2.8L IS USM Macro lens, EF 24-105 f4L IS USM lens and EF 100-400 4.5-5.6L IS USM all at maximum aperture and at 100mm focal length.

a.The camera settings was as follow:
- Quality: Jpeg L
- Color space: Adobe sRGB
- Picture Styles: Standard

b.The images were viewed in a beta DPP version. The thumbnail will show the image plus the focusing point used and the enlarged view is a portion of the Actual Pixels (100%) of the Jpeg image where the focusing point was aimed at. Both Auto Focus modes were used namely One Shot and AI Servo

We are not going to include all the test images, there is just to many, but you will get an idea of how the test was conducted and we will show the problem areas.

For the first test we used the EF 100 f2.8L IS USM Macro lens and selected “Manual point selection”, “Case 1” and “Only cross-type AF points . See image below:

Manual point AF and case

We started with the centre focusing point and selected other focusing points in the following sequence (this sequence was used in all the tests), centre, far left, far right one in, middle top and middle bottom. See image below:


The camera focused perfectly on the spot without any hesitation or hunting (in both One Shot Focus and AI Servo Focus). We did notice that the Top and Bottom focus points tended to have a slight front focus whilst the middle focusing points (Centre, far left and far right) were spot on. See actual pixels image of centre, top and bottom below:

4004 AP top bottom.jpg

The next set of tests we then expanded the focusing points by selecting the following options, even though “only cross-type AF points” were selected the camera will use non cross-type AF points to assist with the focus process, see image below:

_VDB2053 manual point AF expanded 4 and case1 Selectable AF.jpg

We were very keen to do the expanded AF point test. Out of experience with earlier models (1D MkIII, 1D MkIV and 7D) the camera will focus with the closest expanded focusing point ignoring the selected focusing point. This was very irritating, you think you are focusing on the eyes and then another focusing point is used and focuses on the nose, making the eyes “soft”, especially if you use long focal length lenses and are close to your subject.

The results were very pleasing!! Every test shot, the selected focusing point was used to achieve focus. The camera focused perfectly and never selected an expanded AF point (in both One Shot Focus and AI Servo Focus).

The following test we then expanded the AF points even more, now using 8 assisting AF points rather than only 4 in the previous test. See images below:

_VDB2053 manual point AF expanded 8 and case1 Selectable AF.jpg

Once again flawless!! Canon has done a very good job here, so the obvious next test was to activate the AF point auto switching. So we moved the setting from 0 to 1 and selected the “expanding 4 points option”. See image below:

_VDB2053 manual point AF expanded 4 and case1 with AF Switch 1 Selectable AF.jpg

Centre, far right (one point in from edge), top and bottom used the selected AF point whilst the far left (one point in from edge) used the far left AF point to assist. See image below:

4032 wit AP.jpg

We then expanded the AF points to 8. See image below:

_VDB2053 manual point AF expanded 8 and case1 with AF Switch 1 Selectable AF.jpg

Centre, top and bottom used the selected AF point whilst the far left (one point in from edge) used the far left AF point to assist, the same as what happened above when the AF assist points were expanded to 4. Far right (one point in from edge) used the two-assist focusing points below the selected AF point (the same behavior we use to see in the 1D MkIII, 1D MkIV and 7D). This is not a problem as we have told the camera to switch AF points, so the camera is doing exactly what we instructed the camera to do, and that is if you can’t achieve focus with the selected focusing point use the assist AF points to achieve focus. See image below, arrow indicates selected AF point:

4038 wit AP.jpg

In the next test we moved the AF point auto switching to +2 and AF expansion back to 4 assisting points. See image below:

_VDB2053 manual point AF expanded 4 and case1 with AF Switch 2 Selectable AF.jpg

The results were exactly the same as when the AF point auto switching was on +1, and when we expanded the assisting points to 8 the result was also the same as when we had AF point auto switching on +1 with 8 assisting points, so no impact switching between +1 and +2. Obviously this will have a major impact on moving subjects. What did impress us is that it seems Canon has moved away from the philosophy that the AF points must AF on the closest subject, and now is employing the idea that the selected AF point takes priority and the surrounding AF points assist with the focus.

We used both the EF 24-105 f4L IS USM and EF 100-400 f4.5-5.6L IS USM lenses and the results was exactly the same as with the EF 100 f2.8L IS USM Macro.

When we tested the next two AF areas, namely Zone and all 61 points the camera behaved like his previous predecessors. Most of the time the AF system selected the closest focusing point. See images below:

_VDB2053-2 and 3.jpg

4027 zone and auto.jpg

CONCLUSION ON STATIC SUBJECT FOCUS

The AF system really impressed us!! When you use the manually selected AF points, weather it is with expansion or not the AF is instantaneous and without any hunting. The 5D MkIII did much better than what the 7D and 1D MkIV did in the same tests, and that is because of the shift in mindset from Canon. I have already mentioned this but is worth repeating, the 5D MkIII’s AF system put priority on the AF point that is selected, especially when you have assisting points on, and “AF point auto switching” set to “0”. For the first time I will be very comfortable using AF assist points!!!

For the second test, a moving subject, to test the tracking ability of the focusing system, we came up with a test to minimize “operator” error. So many people say that a camera does not focus properly and lots of time it is the photographer and not the camera. We printed black and white diagonal lines on a canvas that was in turn glued on a board. The board was then mounted on the bulbar of our Landrover. My wife and a friend then drove at 40km/h towards me whilst I will photograph the board. The first image was taken from around 70m from the camera and the last image will be around 6m from the camera. The tests were done with three lenses namely, EF 600 f4.0L IS USM, EF 800 f5.6L IS USM and the EF100-400 f4-f5.6 L IS USM lens at 400mm. All three lenses was used on their maximum aperture. The camera was supported on a Gitzo 1648 Carbon Fiber tripod with a Wimberley. The camera settings was as follow:

a.- Quality: Jpeg L
- Color space: Adobe sRGB
- Picture Styles: Standard

b.- Focusing mode AI Servo and Drive Mode 6fps..
- The centre focusing point was used with the different expansion options
- Both Case 1 and Case 3 were tested
- ISO was set to Auto
- Tv mode and shutter speed set at 1/2000 sec

c.The images were viewed in a Beta DPP. The thumbnail will show the image plus the focusing point used and the enlarged view is a portion of the Actual Pixels (100%) of the Jpeg image where the focusing point was aimed at.

CONCLUSION ON MOVING SUBJECT FOCUS

See below an example of the test, once again we are not going to show all the results because we took over 1000 images.

 
  9924.png

 

 

 

 

 

 

 

 

The AF system was set up as follow:

Lens drive-AI servo 1 image-AI servo 2 image

With the different expansion and case options:

_VDB2053 manual point AF expanded 4 and case1 Selectable AF.jpg

_VDB2053 manual point AF expanded 4 and case3 with AF Switch 1 Selectable AF.jpg

The results then was checked in the beta version of DPP we had as well as in Photoshop CS 5 Bridge at actual pixels. See below for one set of results:

Servo Test.jpg

There was hardly any difference between Case1 and Case 3, so we combined the information, just for you to see how the AF system coped with the different lenses and maximum apertures:

With the 600 f4.0L IS USM lens: 80.5% accurate
With the 800 f5.6L IS USM lens: 72.6% accurate
With the 100 – 400 f4.0 -5.6L IS USM lens: 72.9% accurate

All the inaccuracy was usually within the first 15-20 images, and that was from around 100m down to about 55m. Below that that all the images were in focus except 1 image with every test, and that image was when the vehicle started to decelerate. So in summary, the camera tracked exceptionally well, 99% plus accuracy between 8m to 55m, and struggled between 55m -100m. This is not uncommon, all our previous tests we had very similar results regarding subjects further than 50m.

The above numbers is meaningless if you do not compare them so we repeated the test with the 1D MkIV under the exact same conditions with the same lenses. The 1D MkIV AF was setup as follow:

Custom Function set-up:
- CFnIII: 2 the tracking sensitivity was set on Medium Fast
- CFnIII: 3-0
- CFnIII: 4-0
- CFnIII: 5-0
- CFnIII: 8-0

With the 600 f4.0L IS USM lens: 84.8% accurate
With the 800 f5.6L IS USM lens: 75.8% accurate
With the 100 – 400 f4.0 -5.6L IS USM lens: 51.0% accurate

Once again, most of the inaccuracy was when the car was far from the camera, 100m down to 70m, from 70m down to 8m most of the images were in focus. The 1D MkIV struggled with the 100-400 lens, the tracking was allover, but once the car was about 35m from the camera the tracking became more consistent. Is the increased accuracy due to the cropping factor? Both cameras tracking was flawless below 50m with the 600 and 800mm lenses, the 5D MkIII was much better with 100-400 lens than what the 1D MkIV were.

In the field we found the 5D MKIII’s AF tracking ability outstanding. The AF was every time precise and very fast. The 5D MkIII out focused the 1D MKIV with ease. We had several opportunities to photograph Egyptian Geese and Grey Herons in flight, weather you used the 600 f4L IS USM lens with or without 1.4 II converter the tracking was exceptionally good. See images below:

Geese.jpg

 

Action shots.jpg

We also took the camera to a cage fighting night to see how well it will focus in low light conditions. The 5D MkIII again did not disappoint. The light conditions was very low, an average reading gave you 1/90sec, f2.8 at ISO 6400, obviously far to low for action so we had to use flash. The camera was set on Manual mode at 1/1500sec, f2.8 and ISO 6400 and we used the new Speedlite 600EX-RT. We used two lenses namely, EF 70-200 f2.8L IS II USM and 24-105 f4L IS USM. The AF system was setup as follow:

_VDB2053 manual point AF expanded 4 and case3 with AF Switch 1 Selectable AF.jpg

 

Cage fighting.jpg

ISO

Studio test

Below is images photographed in the studio. The tests were done as follow:

a.Camera was mounted on a EF 100 f2.8L IS USM Macro lens which in turn is mounted on a sturdy tripod and ball head.

b.We used Mirror Lock-up and a cable release to trigger the shutter.

c.The camera settings was as follow:
- Quality: RAW
- Color space: Adobe RGB
- Picture Styles: Neutral (Zero sharpening)

d.We photographed a chart that we got in the back of one of Scott Kelby’s books. The reason for using this chart is that it has nice shades of grey to black. Remember noise is easily detected in the darker areas.

e.The images were viewed in LR4 and any Noise reduction was removed and no sharpening was applied. The view you are going to look at is a portion of the Actual Pixels (100%) of the RAW image.

f.Out of experience, the moment you notice any noise in the studio shot the image will be very noisy in the field.

g.We also have showed the ISO from ISO 400 and above as all the cameras that were used give excellent low noise images from ISO 400 and below.

Thumbnail ISO.jpg Thumbnail ISO.jpg Thumbnail ISO.jpg Thumbnail ISO.jpg Thumbnail ISO.jpg
ISO 400 ISO 800 ISO 1600 ISO 3200 ISO 6400
Thumbnail ISO.jpg Thumbnail ISO.jpg Thumbnail ISO.jpg Thumbnail ISO.jpg  
ISO 12800 ISO 25600 ISO 51200 ISO 102400  

Again these image don’t tell you too much unless you do a comparison, see the comparison below at ISO 3200. We converted the RAW images using LR 4 without any Noise reduction or sharpening, and resized all the images to 40cm x 26.67cm at 300dpi just so that all the images are the same size. We then cropped an actual pixels view of each image, see below:

ISO Comparision.jpg

Immediately you can see that the 5D MkIII wins this contest hands down.


Field test

Below is images photographed in the field, it usually tells a complete different story than the studio images. The camera settings and process of the images were as follow:

a.The camera settings was as follow:
- Quality: RAW
- Color space: Adobe RGB
- Picture Styles: Neutral (Zero sharpening)

b.The images were viewed in LR 4 and any Noise reduction was removed. The view you are going to look at is a portion of the Actual Pixels (100%) of the RAW image.

ISO Comp Field.jpg

CONCLUSION ON ISO

Over and above testing the AF system at the cage fighting we also compared photographing in Jpeg with high ISO noise reduction on and in RAW removing the noise ourselves in LR 4. The camera was set-up as follow:

The camera settings for Jpeg:
- Quality: Jpeg L
- Color space: Adobe sRGB
- Picture Styles: Standard
- High ISO speed noise reduction: On standard

20120330-210A3358-3.jpg


The camera settings for RAW:
- Quality: RAW
- Color space: Adobe RGB
- Picture Styles: Neutral
- High ISO speed noise reduction: Disabled

Images were processed in LR 4. Noise was remove using the Detail tab with the following settings:
Luminance: 20
Detail: 100
Contrast: 0
Color: 25
Detail: 50
Default sharpening was applied and no additional sharpening was applied during export.

20120330-_10A3565-1.jpg

Both sets of images was photographed with the EF 70-200 f2.8L IS II USM lens at 1/1000, f2.8 at ISO 6400. We think that the noise removed by the camera in the Jpeg image is slightly just too much and makes the skin a bit to smooth. In hindsight we should have put the noise reduction on “low”. The noise is well controlled and images up to ISO 3200 have exceptional detail and very little noise in. As you can see from the above ISO 6400 images, and the images shown in the AF section, the results are very good and will make a lot of sport photographers very happy. In our field tests we hardly shot below ISO 800, and all the images were good enough to use without removing any noise. See example below:

20120327-_10A1797-1.jpg

RESOLUTION

The resolution of the 5D MkIII is exceptional!! The amount of detail you get in your images are suburb, even on high ISO’s like ISO 6400. The image below was shot on ISO 400 with a 1.4II convertor on:

20120326-_10A1063-1.jpg

 

Resolution comp ISO 100 and ISO 3200.jpg

The image above just shows you how well the resolution is still at ISO 3200. Noise was removed from the ISO 3200 in LR4’s detail tab. The settings was as follow:

Luminance: 17
Detail: 100
Contrast: 0
Color: 25
Detail: 50

GENERAL:

There were two other functions that we really liked about the 5D MkIII and that were assigning One ShotÛ AI Servo switching to the new larger DOF preview button. The position of the button is very handy and makes switching between AF modes very easy. Without taking your eye of the subject you can switch between AF modes and recompose your image. The image below show such an example, to be able to get the Hippo’s head in, we switched to One Shot mode by depressing the DOF preview button, focused on the Heron and recomposed:

20120328-_10A2227.jpg

The second function that we like is the compare and rating of your images on the LCD screen. The rating is compatible in LR and PS.

Compare images.jpg

FINAL CONCLUSION:

Well let’s start with what is negative about the camera, not to many things? I would have liked to see the frame rate at 8 fps. We did miss some action shots, due to the lower frame rate. The buffer “maximum burst” is also a little low for those that want to photograph action in RAW. The camera we used had a maximum burst of 13 frames, I see that the technical specifications says 18. The firmware on our test camera was not the latest firmware and that could have caused the lower buffer, but 18 is also not enough.

On the positive note, Canon has built the best camera I have ever shot with. The AF system is the best AF system in a Canon camera I have seen. The tracking is suburb, the speed with what AF is achieved on a subject is very very good. The AF menu is well laid out, and with the Info button help facility, anybody will be able to customize the AF system. I think the highlight for me in the AF system is that the emphasize has clearly shifted to making the selected AF point the priority and surrounding/assisting AF points becomes exactly what they should be doing, assisting.

The 5D MKIII has also achieved this very fine balance between very good resolution (image detail) and noise. With the 5D MKIII you will be able to capture exceptional high detail images at ISO’s as high as ISO 3200. The camera will suite most styles of photography, but the 5D MKIII must be a wedding photographers dream come true!!!