Saturday, 7 January 2012

Thoughts on Nikon D4 and Canon 1DX

I have watched Nikon and Canon compete with each other since the mid eighties when I first became interested in SLR cameras. With the EOS system Canon took away a lot of the pro photographers away from Nikon - primarily due to better autofocus, and later on, better digital sensor technology. But with the D3 Nikon fought back and at present there is more of a level playing field.

It seems that in recent years Nikon and Canon have been anticipating each others developments. The 1DX and D4 have both adopted a number of features that are uncannily similar to each other.

I personally have not used the D3 series cameras - I own a D300. I also own a Canon 1DS - which is quite primitive compared to today’s DSLRs. So my thoughts here are more armchair speculation than real life experience with these products.

With the 1DX, Canon had to prove their autofocus credentials which were dented by 1D Mark III and did not recover fully with 1D Mark IV. They seem to have gone for an all out redesign of the system - giving up their original 45 point design that they introduced way back with EOS-3 I think. They have even copied Nikon’s AF layout - presumably this says something about the effectiveness of Nikon’s design.

Canon have also completely reworked their mirror technology - replacing the one they introduced originally in the EOS 1V in 2000. The new method seems to be much faster than any prior implementation - 60ms blackout compared with 74ms blackout in D3 and 80ms blackout in 1D Mark IV. Shaving off 20 ms is a major achievement I think. Nikon seems to be still tweaking their F5 based design introduced in 1996. This system probably needs a major redesign to reach Canon’s level of efficiency. An alternative approach for Nikon would be to develop the technology they used in the Nikon 1 series cameras as this would remove the mirror speed barrier - but I think the on-sensor phase detection technology is not yet mature enough for a pro DSLR and may have some fundamental limitation (such as requirement to interpolate missing pixels or low light sensitivity of the AF pixels or the degree of de-focus that can be detected - Nikon is tight lipped about any limitations) that may mean that it will never match the mirror based approach.

The mirror blackout time is important as it increases the time available for the AF system to work. So at 10 fps, the Canon 1DX has more time available to track focus between shots, although that may be offset slightly by the Canon’s longer release lag of 55ms compared with Nikon D4’s 42ms. (Canon do say that the release lag can be reduced down to 36ms but I suspect this will work only when the max aperture is used so that the lens stop down time is eliminated.) Canon have also added more cross-type sensors - as well as improving their light sensitivity. So from a spec point of view, Canon’s AF system appears to have an advantage.

Of course, it remains to be seen whether the 1DX performance in the field lives up to its specifications. Where the D4 has trumped the 1DX is in the ease of AF mode selection which no longer requires one to take the eye away from the viewfinder. Surely this is a major usability breakthrough - one that Canon must wish they had thought of. Canon is still relying on menus to configure AF mode selection. Canon’s AF options also appear far more complex - almost as if their auto modes are not as good as Nikon’s so the user must manually tweak the system.

On the sensor design front also the Canon system appears to be more efficient on paper - at least if one can infer this from the highest normal ISO rating of 51200 on the 1DX. Generally though my impression is that Canon rely more on noise suppression in software than Nikon - so it may well be that the higher ISOs in the 1DX use greater noise suppression - with potential side effects. But this is just speculation - the actual results will come out pretty soon. If Canon have indeed created a sensor that produces cleaner high ISO images than Nikon’s then it would be major advantage for Canon.

For Nikon the challenge must have been to maintain or surpass the D3S performance while increasing the number of pixels. I am glad that they stuck with ISO 12800 as the highest normal ISO - I hope this is because Nikon want a certain level of quality in the normal ISO range.

It seems that due to new laws in Japan, Nikon has reduced the battery capacity in D4. Also, Nikon's D4 design appears to be more power saving - for example less number of CPUs than the Canon 1DX, which uses two DIGIC 5 and one DIGIC 4 processors compared to a single EXPEED 3 processor in the D4. Given the restrictions on battery capacity - which presumably impacts the Canon 1DX as well - it will be interesting to see if the 1DX battery performance is as good as Nikon D4's. There is obviously a trade-off somewhere between meeting ever higher computing requirements and the power source available to support this requirement. Photographers do appreciate longer battery life - Nikon appears to work harder in this area traditionally than Canon.

My final thoughts are with regards to video implementation. Canon obviously have an edge here as they have been in the video business for long. But this advantage is also a disadvantage as it means Canon have more incentive to protect their video business - including the new Cinema EOS system. Nikon on the other hand have no competing video business - so they can go for an all out video implementation on their DSLRs. And they seem to have given it their best shot. The 2.7 crop mode which gives native 1080 resolution is a very clever move - I believe the Panasonic GH2 has a similar feature. Another feature that is desired by many is uncompressed 4:2:2 output via HDMI. The real test though will be the quality of the video capture - if Nikon has been able to achieve the same quality as Canon then they have a winner.

p.s. I previously put together some material on the Nikon mirror system here - http://www.cameragossip.com/nikon/nikon-f5-and-d3-shared-genes