Live view versus the cheating DSLR viewfinder

CANON may get some stick from their own users who reckon Live View is not much use, and Sony’s decision not to put it in the A700 may well be defended on the same basis. How wrong that is. Here’s one (plus another two) good reasons why Live View is your best friend in the studio.

DSLR viewfinders have to put up with 70 per cent of their light being diverted by mirrors and mirrors. No smoke, just mirrors. A semi-silvered main mirror and a couple of secondary mirrors magic all that light away to give it to the hungry autofocus module, leaving you very little to view with. To compensate, the makers fit very transparent focusing screens, which do not intercept the light – they bend it a little, scatter it a little, but don’t stop the image dead in its tracks the way a good dense groundglass used to.

Anyone who remembers vintage kit will know the prismatic screen of the Leicaflex, and the clear screen of the Contarex/Contaflex. These were unbelievably bright but did very little to help you judge depth of field. With any clear screen, deviating your eye a bit from the centre causes edge dimming or occlusion. What today’s ‘super spherical’ (load of balls…) or ‘super acute matte’ (ditto) do is to simulate the visual effect of a groundglass while transmitting nearly as much light as a prismatic screen or an oiled groundglass. That is an old 4 x 5 studio camera trick, oil the glass and you can see the focusing image from the slowest of lenses, but you stop being able to see into the corners and the depth of field no longer gets shown properly.

These modern screens fool you. You fit a lens which is f/1.2, f/2.8, or f/2. You view the scene, and it looks like so much is in focus. But your screen, the lens and the viewfinder optics are combining to create an aperture which equals f/4.5, f/5.6 or most commonly f/6.3. That’s why the image from an f/3.5-6.3 zoom does not appear to get two times darker as you zoom. The f/3.5 view is really being seen as if it was f/6.3 to start with. It is not that simple, as stopping the lens aperture down will affect brightness, and it will apparently change depth of field. It just doesn’t do so proportionately or correctly. It is rather as if the aperture behaved like a graded stop (as found in the Minolta/Sony 135mm ‘STF’ lens) with the periphery of the lens progressively shaded. Imagine a centre-filter from a superwide camera fitted next to the iris in your lens.

This is also why some cameras will tend to underexpose with some wide-angles. The exit pupil and the in-prism, through-screen metering just don’t agree at lengths under 20mm. Calibration to specific lenses, and firmware tables which tell the camera what to do, help. But stop-down metering can be way off, wides and teles may need different exposure bias, and all the time you are thinking that SLR viewfinder is WYSIWYG.

Cut to the chase. Live View on the Canon 40D. Here is a photograph taken with the beautiful 85mm f/1.2 USM II lens nearly wide open – I did some at f/1.2 but closing down to f/1.4 just made that tiny difference to the sharpness on the focused chess piece:

Now that’s the kind of shallow focus plane you can use for commercial stock shots and creative work!

But this is what you see through the viewfinder:

This is an actual photograph taken using my special rig-up, through the eyepiece of the Canon 40D. To check that the lens in use is not influencing the result, I took shots with the KM 7D’s special Loreo 35mm infinity focus lens set to both f/5.6 and f/22. To get the focus sharp, I bracketed the dioptre adjustment of the Canon finder. The view is identical, showing that lens aperture on the KM 7D used to shoot this image of the focusing screen does not affect the depth of field shown from the Canon’s lens. It also agrees exactly with what I see by eye through the finder.

Am I assassinating the 40D? No. It has one of the best screens and best finders around. Compared to mirror-prism entry level DSLRs, the screen gives a relatively accurate image. But it is still miles away from representing the true depth of field in the picture. And so is every single DSLR made.

Here is where Live View comes into its own in the studio. OK, you can take pix and review them, but with Live View you can adjust, preview, and above all not be fooled into thinking you have more depth of field wide open than you do. You can magnify the image 5X or 10X, you can focus with great accuracy, and you will be fully aware of what is sharp and what is not, and how the image is formed.

Here is the setup. The 3″ screen is great and you can see, already, that Live View represents the full aperture situation as it should. I don’t know why my shot sharpens up the apparent depth of field on the screen; I can assure you that it does reflect the actual image, and this is probably a result of brightness and contrast issues when rephotographing it, and the size to which it is shown.

Now magnify 5X:

This is useful. Hey, when I walk back to my camera which is shooting this pic, I see the image vibrate. I see how long it takes to settle. Let’s try 10X:

Now I see that the focus, carefully carried out using the AF point over the chess piece then switched to manual for Live View, is not quite perfect. You can use AF with Live View, with the mirror briefly flipping back and the Live View blacking out to allow a focus setting via the AF module. But I was find this very slightly inaccurate compared do manual focusing at 10X. So the focus gets tweaked:

At ten times view, every slight touch of the camera, the smallest shift as I stand on my wooden-boarded studio floor, sets the 10X image vibrating. I set the camera to self timer, even with Live View and no mirror to clack around and cause vibration. Ten seconds self timer. I hold my breath, watching the Live View, being sure that I am not causing those tiny vibrations I can see so clearly on the screen.

In the end I get my shot as shown at the top of this page, at f/1.4 (I could have opted for a different aperture, and used Live View with stopdown preview and/or Live Exposure Preview to assess to depth of field and brightness of the image properly). It is dead sharp, as it should be, where focused. There is no FF/BF. There is no tripod shake because I could see just how very bad the conditions were for this exposure, even at the 1/60th at f/1.4 the f/1.2 lens was generously allowing at ISO 100 with manually set tungsten WB. The lighting is from my Elinchrom Style 600/300 studio flash modelling lamps, and WB has been further adjusted when processing the .CR2 raw image using Adobe Photoshop CS2/ACR 4.2.

You can view a full 10 megapixel image of the final shot here. It might, indeed, benefit from a smaller aperture and also from some local sharpening (this is exported using zero sharpening and zero noise reduction). You may observe the slight purple and green biases given to objects in the foreground defocused zone and background defocused zone. This is common when using very wide aperture lenses on digital systems, and is a result of a convergent ray cone at the focal plane for close objects and a divergent ray cone for distant ones, and the angle at which image-forming rays strike the microlenses. If you want to see this emphasised, get a Sigma 150mm f/2.8 macro and use it wide open very close. The 85mm f/1.2 is not suffering from ‘CA’ to produce this effect, it’s entirely down to the structure of the sensor.

Is Live View useful? Should Sony have omitted it from the A700 when Nikon and Canon have both implemented it on their competitors?

I guess that is up to the user to decide. Now you know a little more about DSLR viewfinders, you will check your images after shooting, especially when aiming for those wide-open, shallow focus effects. But Live View could help you and after using the 40D I reckon it is a feature every comparable DSLR will one day offer.

– David Kilpatrick