Canon 5D Mk 4 Auto Focus

Like Nikon, Canon never do me any favors!

But I do feel that I must say to the world that it ought to give this camera a bit of a break.  It’s had a good mauling in a lot of places, usually by idiots and no-nothings, who keep comparing it to its big brother the 1DX Mk2 – a camera not without its very own set of unique foibles!

The Canon 5D Mk 4 is NOT designed to be a “poor mans” 1DX Mk2.

It’s hardly what I’d call ‘cheap’ in the old purchase price department for starters, so ‘poor mans’ and ‘budget’ are not terms I can easily associate with it.

There are lot’s of things I need to delve into further on this camera to give you guys a fuller picture of the cameras overall performance – most of which is going to involve Calumet or Canon lending me more lenses.

But I can say that I’ve formulated a solid opinion on the Canon 5D Mk 4 Auto Focus performance, and it’s turned out to be a lot better than I’d first imagined.

These are the style of shot that really tells you if your auto focus is working and up to the job:

But before you can start producing the shots you have to go through the tedious bit of testing the AF first.  It was while testing the overall sharpness and accuracy of the AF system that I came across a little problem.

When photographing the old ‘brick wall’ static target I found the system was front focusing by around 40 centimeters at 30 meters.  If I added +4 on the AF micro adjustment (using the 500mm don’t forget) then everything was razor sharp.

This didn’t seem right in my eyes – I’ve never felt the need to use micro adjust on Canon gear to achieve sharp focus on a static target – perhaps I’ve just been lucky!

But after testing this body with another 500mm L IS II, and Calumets lens on 3 other bodies, all tests revealed the same necessary +4 adjustment.

The difference is quite marked!

• Bare in mind that all these ‘static tests’ MUST be done with the aperture WIDE OPEN (in this case f4).
• I always use the high ISO capabilities of a camera to the maximum, which allows me the luxury of shooting at between f6 and f8 to maximise DoF and use a high enough shutter speed to stop the action.  Manual exposure with Auto-ISO is my usual method of shooting with long glass.  A noisy image that is razor-sharp will ALWAYS out-sell a low noise image full of motion blur!
• At f8 hardly any of the ‘poor sharpness’ (above left) is visible in the image because DoF is doubled from 40cms to over 80cms behind the plane of focus at this distance.
• If I was to swap out to a shorter lens then the required amount of micro adjustment would be less, and with a longer lens MORE!

However, when we come to photograph the likes of ‘Joey’ we have a BIG problem!

Adding positive micro-adjustment values is basically like adding BACK FOCUS – you are telling the system to focus BEHIND where it perceives sharp focus to be – in other words ‘focus further away’.

So with a head on closing target/subject the resulting AI Servo sequence of frames will all be back focused.  The camera will be focusing behind a subject that’s getting closer – it’ll never work!

What we need is the system to move the plane of focus AHEAD of the subject, so that when the shutter opens for the next frame, the subject and plane of focus are hopefully in the same place.  This is how PREDICTIVE AF works, and cameras like the 1DX Mks 1 & 2/Nikon D4/4S truly excel at it.

Dialing in an opposite value of -4, and using AF Case 4 settings with Zone AF for the AI Servo sequences of little Joey yielded good results, but the level of consistency was still below what I thought was possible.

And it certainly got even less consistent when I changed to Point Expansion or Point Surround AF modes.

But now I’ve settled on a custom setup that is NOT obtainable on any of the fixed AF cases; TS & ADT both at +2 together with -3 AF micro adjustment:

I’ve just uploaded a new video to my YouTube channel where I discuss the performance of the Canon 5D Mk 4 Auto Focus system, and go through A LOT of full resolution images.

Note, some of you may get bored and think I examine too many images – shame on you! There are 4 sequences, and each one represents around 4 seconds in real time and are a ‘buffers worth’ of shots.  So all those boring shots took less than 20 seconds to acquire –  I have to show all the shots in a sequence to illustrate the level of consistency, and I show 4 sequences to prove none of them are a fluke – I DO THE JOB RIGHT – unlike some other reviewers!

You can view it at full size by clicking the YouTube icon bottom right once you press ‘play’, but be warned it’s 36 minutes long!

I’m not finished with this camera just yet I don’t think; I must admit that I do quite like it!

Is the Canon 5D Mk 4 Auto Focus capable of better performance than that of the venerable old 5D Mk3 ?  Yes, it is.

Is the image quality better than the 5D Mk3 – oh you betcha it is, by a country mile and just like the 1DX Mk2 advantage over the 1DX.

Are the G/T algorithms (ISO), sensor and ADC output better – from what I can see that’s a ‘yes’ too; but then I’ve not done any dynamic range testing yet – kinda hard when the only lens you’ve got is a 500mm!

I’m getting used to the ‘touchy-feely’ screen now, but the fixed 7fps HS frame rate pisses me off a bit, I’d like to be able to drop it to 6 or 5 to the AF system even further.

Take my advice and don’t be impressed with the ‘Duel Pixel Raw’ feature – it’s CRAP. It does absolutely bugger-all apart from slow the camera down and produce massive files that are not worth the time or effort.  And you can only ‘work’ them in that clunky DPP software which is a total abomination to try and use!

And old UHS1 SD card tech? The camera would be better with a CF slot in conjunction with a CFast2 slot ‘a la’ 1DX Mk2 – in my opinion of course.

Become a patron from as little as $1 per month, and help me produce more free content. Patrons gain access to a variety of FREE rewards, discounts and bonuses.