HDR question.

[Bobobird]

I wonder if there is a way to determine the amount of EV to use?

Or is there some procedure that can be used to determine the amount of EV to use?

Surely it cannot be a random decision or can it?

Thanks.

[victor]

Bobo,

I use -2,0 and +2 on the main. I will sometimes take an extra 2 exposures @ -3 and +3 for church interiors with a mix of natural light and shadows.

Look at the following sites hdr tutorial www.stuckincustoms.com

Regards

David

[Tringa]

I don't know if there is a way to determine the ev steps needed for a particular scene, other than experience and the general guide that the greater the range of light in the shot the more images you will need.

Because it is available on the camera I tend to do five shots +2,+1, 0, -1, and -2.

Dave

[Bobobird]

Thanks David and Dave.

Will check out the site and see how it goes tomorrow.

5 shots seems an interesting concept. Will do that too.

[FrankMi]

Hi Bobo, the main reason most folks use five exposures is that most cameras have 1 EV increments for Auto Bracketing and shooting -1EV to +1EV really doesn't buy you enough to be worth while. I don't have Auto Bracketing on my D3100 and I have found that -2EV, normal, and +2EV works well for all but the most extreme dynamic range scenes. For shooting directly into the sun near sunset, indoor shots with a bright sun blanking out the windows, and situations where it is very difficult to deternine what a normal exposure would be, I use -4EV, -2EV, normal, +2EV and +4EV. There is so much overlap in the detail you can recover that unless the camera will only do 1EV increments, it's not really needed.

HDR Software like Photomatix Pro doesn't care how many shots you include (I've tested with up to 9 images) so long as they are at least 1EV apart.

[Dizzy]

I've been using LR/Enfuse, and while it doesn't create the HDR effect, it blends images
of various exposure into a nice composite.

Shooting on Manual, I generally take one shot exposed correctly for the brightest item
in the frame, then 2-3 exposures on each side of that in 2/3 of a stop increments. From
those I'll take the best 5-6 of the lot and combine in LR/Enfuse. It's not always a blend
of +2, +1, 0, -1, -2, and it could be that -2, -2/3, 0, +2/3 gives the best result. Cover
the exposure spectrum well, and then try blending different combinations until the desired
result is achieved.

This Fall colors image was the result of blending 4 frames, with exposures of 105mm @
f/5.6 and 1/160, f/5.6 and 1/60, f/8 and 1/100, f/10 and 1/50 (hellofa mixture..lol) Unfortunately,
it was also breezy, and the trees were blowing around between exposures so it doesn't sharpen well.



Mike

[Colin Southern]

The short answer is "however many it takes to get you from clean shadow detail to good highlight detail".

Most modern cameras are already capturing around 12 stops of dynamic range (which incidentally also means that in many cases people don't need to use HDR techniques), but if you're capturing 12 stops anyway, then going 1EV either side of that isn't going to make much difference (to get an idea of how little the change is, stop your lens down 1 stop from wide open, and then press your DoF preview button whilst looking through the viewfinder - it'll cut the light back by 1/2, but you'll hardly be able to notice the difference.

So the normal mantra is "2 stops" for the bracket, but even that's going to depend on the DR you need to capture, and how many shots your camera can do in 1 bracket. I normally shoot at 2 stop increments, but then again, I can do 7 shots in a bracket.

[herbert]

You can use your camera histogram. Shoot enough brackets so that at one end you have no blocked shadows (peak only visible above half way on the histogram) and at the other end you have no blown highlights (peak below half way on the histogram). Two stops apart should be fine for all the frames in between.

Obviously a tripod will really help too.

[pwnage101]

I say overexpose until you can see all the shadow details, then underexpose until you can see all the highlights. You should be able to use those two images, no matter how much they differ. Worst case scenario your camera has 9 stops of dynamic range (DR) so the maximal range you could get with two shots should be around 14 stops.

The Nikon d7000 and the Pentax K5 get 14 stops with one shot! If you have one of these cameras, use HDR sparingly.

[Inkanyezi]

There is actually a way to determine the dynamic range of a scene and to evaluate whether you need to blend two images into HDR or not. You would seldom need more than two images, as even the worst case low dynamic range of a camera would be at least eight stops, which may overlap into 14 stops for a HDR image. And as pointed out, there are cameras that are able to capture so much DR in one shot.

To determine exposure and scene dynamic range, spot metering is a method that has been used for well over half a century, and many cameras have spot metering included in their exposure modes. You need to find out how much you may compensate to plus, in order to render a highlight close to clipping. On many cameras, this is two steps, but it can be more, up to three full steps. So the darker of the two exposures is taken with a spot reading on the brightest highlight compensated by your personal plus value, and the second exposure should be at least four stops more. For a camera with low dynamic range, six stops more is the limit, but cameras with higher DR may allow more stops. With a camera that has eight stops DR, two images with four stops difference give you twelve steps of dynamic range, and with six stops difference you get fourteen. Mostly this is sufficient, and Pentax K5 users need not bother about the second image.

You may use spot metering to determine the dynamic range of a scene, by comparing shutter times the camera would choose for the brighetst highlight and darkest shadow area.

It might seem odd to compensate to plus for the darker image, but if you think a bit deeper about it, it is a spot reading of the brightest area that shall be rendered in the image. This area should have an exposure maybe three stops brighter than an average reading; hence compensation when you spot meter the highlight is set to plus for the "dark" image, which does not burn out the hightlights.