One of the best things about digital cameras is that they allow a lot of
experimentation due to the lack of processing costs and the instantaneous
availability of results. It becomes extremely easy for the user to test out
ideas, and technical exercises are easily carried out.
These factors led me very quickly to stumble upon a novel method of getting
sharper hand held photos from my EOS 10D, this method is not limited to this
particular model or brand of camera however and can in fact be used with any
camera that can take a rapid series of continuous frames.
In fact, this particular technique has always been possible, however I have
never heard it mentioned anywhere despite reading avidly about photography for a
number of years. If you have seen this before, please accept my apologies. I do
believe that it is far more convenient with a digital camera, not only because of
film costs but also because with digital it is very quick and easy to identify
the sharpest image, as will become clear later.
The technique itself is very simple: simply compose your image in the
viewfinder and take a series of consecutive frames – as many as your camera
will allow or is economical. On my 10D I take a burst of nine frames. Now while
this may sound simple to the point of being totally obvious, there is a twist. My
handholding technique is fairly poor, I can barely use the 1/focal length rule
and have in fact resorted to shooting a least a stop higher than this
“rule” in the past after discovering that any less would sometimes
result in soft photos in my case, as clearly my technique is lacking.
However after trying out the method outlined above I have not only been able
to get consistently great results at the 1/focal length setting, but in fact can
also get very good results two or more stops slower, which is comparable to the
results people have been getting with IS lenses – but without IS and with
all types of lenses, including the shorter focal lengths that IS isn’t
designed to cater for. As an aside I have found that the 1/focal length rule
doesn’t seem to be affected by the DSLR cropping factor, in my case
As I mentioned earlier, with digital files it is extremely easy to quickly
identify the sharpest frame in a burst: simply look for the largest file size in
the sequence. In a series of JPEG’s taken in sequence the one with the
largest file size is invariably the sharpest. This must be down to the way that
the image compression works, whatever it is I have found this rule to be
consistently true. If you shoot RAW it is worth extracting the embedded
JPEG’s in order to determine the sharpest file before doing any RAW
I believe that there is nothing more to this trick than simply maximising your
odds of getting a sharper image, and by taking 9 consecutive frames you are
simply increasing your chances 9 times. Having said that, I do find that firing a
continuous burst is more effective than taking 9 individual frames, probably
because I relax more when firing a burst and the shutter is only depressed once
which also helps to minimise movement. Besides, firing a burst is obviously a
faster and more convenient way of getting the 9 shots.
To demonstrate the technique I have done a test, which is detailed below. Each
sequence consists of a shot first on taken a tripod, then as a single shot hand
held using my best technique, and then finally a burst of nine was taken and the
best frame from it was compared to the two control shots. In the images below you
may examine the results, in each case the first image on the left is the tripod
shot, the next is the single hand held shot and the third image is the best shot
from the burst of nine.
The camera was set to manual and auto white balance, the images below are 100%
crops taken from the centre of the image. No processing whatsoever was done to
the images, this is the output straight from the camera, shot in JPEG mode.
First up was my 17-40 F4L, set to 17mm:
1/15 @ F11: At this focal length and shutter speed there is very little
difference between any of the shots.
1/8 @ F16: Again not much in it, shorter focal lengths are pretty
hand-holdable it seems.
1/4 @ F22: Diffraction has softened the lens itself at this point, but the
shot from the 9 frame burst is visibly much better than the single hand held one,
although not quite as good as the tripod mounted shot but that is hardly a
surprise. This is definitely a usable shot, which is only slightly softer than
the one from the tripod and is pretty remarkable for this shutter speed,
especially when you consider that with the DSLR crop the effective focal length
is actually 27.2mm.
Next I zoomed to a focal length of 28mm:
1/30 @ F8: Here you can see what I mean about my hand-holding technique, the
single hand-held shot is clearly soft and in a real world situation I would have
been disappointed with the sharpness of the image. The shot from the burst though
is pretty indistinguishable from the tripod mounted one.
1/15 @ F11: I think this illustrates my point about probabilities, the single
hand-held shot at this shutter speed is actually sharper than the one taken at
1/30. This isn't surprising really when you consider that hand-holding is a hit
and miss affair even at the best of times. The shot from the burst is sharper
though, and pretty hard to distinguish from the tripod mounted one.
1/8 @ F16: Again, the tripod mounted shot is obviously the clear winner, with
the single hand-held shot being very soft. The shot from the burst is usable
though, albeit visibly softer than the tripod shot.
Next I fitted my 50 1.8 to the camera:
1/45 @ F8: Not much in it, but the tripod mounted shot has a slight edge,
followed by the shot from the burst and then the single hand-held shot. All are
Similar results to the last sequence but this time the single hand held shot
is slightly softer still.
1/10 @ F16: Here again the benefits of this technique are pretty obvious, the
image from the burst is slightly softer than the tripod mounted shot but still
perfectly usable, and streets ahead of the single hand-held effort. Again,
considering the effective focal length of 85mm and the shutter speed this is a
pretty good result.
Finally I decided to mount my 300 F4 L, which sadly does not have IS, and put
it to the test:
1/250 @ F4.5: Again not much in it at this point, and the result is similar to
others in that the tripod mounted shot is sharpest, followed closely by the one
from the burst and the single hand-held shot is the softest.
1/125 @ F6.7: Here the results are more conclusive, the single hand-held shot
is soft while the one from the burst is almost as sharp as our tripod control
1/60 @ F9.5: Again a very conclusive result, the tripod mounted shot is the
sharpest but the one from the burst is almost as good. The single shot is awful.
Again considering the effective focal length of 480mm this is a very good
Of course this method has many limitations, and isn’t as convenient as
having an Image Stabilised lens, but I do think it can have it’s uses and
may help you to get the shot in a pinch. It is obviously only really suited to
static subjects, and can only realistically be for occasional use unless one has
an awful lot of storage capacity. On my particular camera model there is no way
of knowing the size of particular files and I don’t find the LCD useful for
judging critical sharpness, which means that the images need to be compared on a
computer in order to sort out the sharp from the soft. On a long trip I can
imagine that it would be very easy to accumulate hundreds if not thousands of
images if one were to rely too much on this method, whereas with an Image
Stabilised lens this would not be an issue.
I am also curious to see whether these results could be improved if used in
conjunction with an IS lens, however since I do not own one I have not been able
to put this idea to the test. Maybe someone else could carry out such a test and
With the excellent high ISO performance of current DSLRs, this technique does
offer the ability to get hand holdable images in very low light, or can also
offer the possibility of keeping the ISO setting lower rather than higher in
marginal light and give better quality results.
As an example of this technique put to some real world use rather than
shooting the block on the other side of my road, this is an image taken in Exeter
Cathedral with my 17-45 set to 35mm (effective focal length 55mm) and taken at
1/20 second at F5.6 and ISO 800. The full size file is pin sharp.