This needs to be repeated. Yes I’ve blogged about this for nearly 2 years now, but I can’t drive home this point strongly enough…
If you are trying to follow someone’s HDR tutorial, you will DEFINITELY get different results than they got if you are not shooting with the exact same sensor.
A single shot on a D3x is like taking two separate shots with an older Canon Rebel about 1 EV stops apart and then blending them in post. So if a D3x shooter says you only need 3aeb with frames at +/-2EV, you actually need 6aeb at +/-1EV with your older Canon Rebel to get the same gorgeous post production results. How do I know this? I’ve shot with both!!!
So you can’t follow tutorials of someone shooting with a D3x if you only have an older Canon Rebel. Here’s a chart showing some recent DSLRs and their sensor rankings.
What is extremely important to understand is that these numbers show ONLY the absolutes. Even though a camera may have light show up for 10+ stops does not mean the data is clean throughout the entire histogram. There is noise in shadows on all of these models. And the noise print is dramatically different.
If you are shooting with even a brand new APS-C sensor, it more than likely will not have shadows as clean as even a 3yr old model with a full frame sensor. Even at 100 ISO on both models. Its just physics.
You can yank FILL LIGHT up to 70 on a Nikon D3 shot and the same scene shot with a Canon 60D might only be able to have FILL LIGHT pulled up to 30 or 35 before breaking down badly. So the 60D shot never gets to be lightened up as much as the D3 shot. Again, its mostly physics. An APS-C sensor is simply much smaller in size than a full frame sensor.
For example, my Canon 60D is a 2010 camera, but its APS-C sensor can not hang at all even at 100 ISO with my 4 year old full frame Nikon D3 sensor shooting at 200 ISO. And my old and crusty full frame 5D from many years back has better shadow data than the 60D as well (100 ISO on both).
When bracketing you only get extremely clean data in the center/sweet spot of each capture. A camera like a Nikon D3x gives you more clean data in the middle part of the histogram than even a Canon 5D Mark II. So if you have to bracket a huge dynamic range scene with 11 frames spaced 1EV apart with a D3x, you might want to consider bracketing only 2/3rds steps between frames with a 5D Mark II (and increase the number of shots to cover the same dynamic range). With an APS-C sensor you will often need to go down to 1/3rd EV jumps between frames and take far more shots to cover the same dynamic range in order to caputure the same clean data throughout that range. Basically, the sliver of perfect data of each sensor (per capture) decreases as you move down the scale of the chart provided.
I know this because I’ve delivered nearly 15,000 commercial images that were originally fully bracketed series of 7-15 shots per image. And I’ve used dozens of cameras from various brands like Pentax, Nikon, Sigma and the worst of dynamic range brands… Canon. So I’ve done a ridiculous amount testing to try and find the holy grail of HDR capture. In the end, there is no one camera sensor or system that is perfect, but what I’ve found are general rules that allow you to make sure you get the entire dynamic range of a scene nailed with what system you are currently using.
