HDR techniques are still very much evolving and developing. Although first pioneered over 250 years ago, new digital equipment and software are allowing us to further develope techniques and extend the creative bounds of High Dynamic Range Imaging.
Some consider using just one image to create an HDR is false and the terms Faux HDR and Faux Bracketing are commonly used to describe single image HDR. Many also consider that single image HDR is a valuable tool and valid technique. My personal view is the latter and following is explanation and results explaining single image HDR technique. I don’t have all the answers nor promote myself as an “authority”, I do have my own practical experience and results, have researched articles and discussed this on forums. There is opposition to single image HDR which, for some, is based on unsubstantiated “facts”, a few of which I will also address here.
What ever one wishes to call it, I am not phased either way if one person considers an image as HDR where someone else does not due to the techniques applied. If the end result is a quality image then great. The intent of this article is to present some expressions of basic logic and some method application which may help some in the understanding of single image HDR processing.
The amount by which the Dynamic Range of an image is increased in the HDR process is not specifically defined. It does not need to be + or -2Ev or more to be HDR imaging, it can be subtle or more obvious.
Another misconception is that the HDR software requires images to have a large exposure offset range to effectively process the images into a single HDR. The reality is that the HDR software is equally as effective in subtle changes to images as it is for very high contrast images, even to toning down a few highlights by a small amount.
The two basic techniques for single image HDR are 1) to tone-map just the one original RAW image; and 2) to create a copy or copies of the original image and adjust the exposure up or down as required, then send these adjusted image(s) together with the original, for tone-mapping or fusion processing by the HDR software.
As previously mentioned some say this is cannot increase the dynamic range of an image. It has also been suggested that both the aforementioned processes produce the same result because they are both processing the same data. Really ??? …
The following image is a small valley a few minutes west of our home. The view is north-west just after sunset, so the light source is from the left in the image. The HDR effect is only subtle but I believe also improves the image.
Exposure adjusted copy of the above image:
This is the HDR from using method 1) above, tonemapping just the one original image:
Notice the light distribution is different from the image copies HDR above and the sky detail is noticeably less.
The Dynamic Range of a camera image sensor can be described as the greatest possible amplitude between light and dark details the sensor can record.
HDR photography, while generally accepted as mulit-shot, or bracketed shot photography, is simply defined as technique(s) applied to create a single image with a wider tonal range than is possible with a single exposure.
It follows from the definitions above that a High Dynamic Range image is one that extends beyond the Dynamic Range of a given camera sensor, but all digital cameras are not created equal. It logically follows also therefore, that what is defined as an HDR image can also vary by camera/sensor and scene being photographed, and that by definition some cameras capture HDR, or higher dynamic range images by default relative to other digital cameras. Follow? No? Don’t agree? Read on …
For example, while my current Nikon D7000 has a Dynamic Range of 13.9 Ev (DxOMark Lab results), another similar common camera, the Canon EOS 7D, has a Dynamic Range of 11.7 Ev. Potentially for the Canon 7D to capture an image of the same scene with the same Dynamic Range as the Nikon D7000, it would require bracketed shots to make up the extra 2.2 f/stop dynamic range differential – in theory at least.
The extent of increasing the Dynamic Range of any image is only defined as an increase, not any specific amount. It may be subtle for some scenes or dramatic for high contrast scenes, each is HDR none the less.
In the same valley as before but facing home with the light source behind me, is the 0Ev exposure of a bracketed pair of images, below that is the second bracketed image offset by -2Ev:
-2Ev bracketed sequence image:
This is the above image in Lightroom, showing the histogram:
Now in lightroom I created a virtual copy of the first, lighter image and adjusted the exposure, black clipping and fill light as seen here the histogram is also almost a spot on match to the previoius bracketed image:
The Lightroom export process creates a new .tiff file applying the edited values, so the original pixel values are no longer available to the HDR software (or even Lightroom for that matter) in the newly created .tiff image.
Here is the result of tonemapping the one image:
Resulting HDR from the bracketed pair of images:
HDR from the lighter image plus the adjusted virtual copy:
The latter two are very similar and both are subtly different to the single image tonemapped result most noticeable in the sky details. HDR can be as delicate as it can be pronounced.
This image is very clean with little noise so the skies are all quite clean but where the differences are most notable is in the following images, the left side is a crop of the bracketed image HDR, and the right side image is a crop of the image copies (single image) HDR:
The foliage movement has created more ghosting and halo effects than the HDR software can effectively handle resulting in loss of fine detail, not only in the highlights but across the image generally.
The next pair of images are zoomed to 2-1, showing the foliage and the lack of resulting detail in the bracketed image HDR on the left, compared to the image copies (single image) HDR which once again is on the right. Foliage misalignment between the bracketed images has blurred the fine details as is shown, the single image HDR is sharper on the finer foliage:
A few more examples -
Patong Beach Ski Train original:
Exposure adjusted copy:
Nai Yang Crown Hotel Original:
Nai Yang Crown Hotel -2Ev exposure adjusted copy:
Nai Yang Crown Hotel +2Ev exposure adjusted copy:
Final single image HDR, original exposure and two exposure adjusted copies of the same image:
Exposure adjusted copy:
HDR from one image:
HDR from the original plus exposure adjusted copy:
Wanganui River original:
Exposure adjusted copy:
HDR from adjusted image copies:
HDR from one image processing:
Single image HDR processing, is it indeed false HDR – Faux? Or does it have merit, is it also valid HDR technique and true – Vrai? Well you have seen some results here so you can obviously make your own mind up on that one.
This, for me, is another effective tool at my disposal. There is no substitute for a quality exposure, as always the basic rule is GIGO, Garbage In Garbage Out. We can and do apply many different processing techniques to achieve the vision of that scene as we saw it at the time.