Photoshop High Dynamic Range (HDR)Test Shot


Status
Not open for further replies.

marcwang

New Member
Jan 9, 2003
1,112
0
0
East Singapore
www.pbase.com
This is the first time I'm attempting to use the HDR feature in CS2.
The scene was taken at mid day, with a full intensity sun through the window. I'm quite impressed by the HDR feature, although there is a learning curve involved.

Pardon the quality... I save it at high compression.
original.jpg
 

Hi,

I was experimenting with it, too. Here's my take at the Punggol Beach at sunset. The final is a composite of 7 images (-2 to +2 ev)

#1 Over exposed
original_over.jpg


#2 Under exposed
original_under.jpg


#3 Final 7 image composite HDR
final.jpg
 

u mean CS 2 auto blends the files? or must go thru many steps to blend? got online tutorial? hmm..haven't go hunt for CS2.....
 

di0nysus said:
u mean CS 2 auto blends the files? or must go thru many steps to blend? got online tutorial? hmm..haven't go hunt for CS2.....

Kind of. It combines the multiple images and then lets you adjust the dynamic range (more highlights or more shadows) before converting it into normal 24/8 bit mode.
 

Interesting. Will auto-bracketing yield enough range for HDR to work with? I don't think I can afford to shoot 7 exposes for a single image, due to time and other constraints.
 

Wow, thanks for the info indeed. Like ibs, I wonder if auto-bracketing at -2EV 0EV +2EV will do the trick as well?
Time to upgrade! :)
 

hmm": I haven't tested this feature yet, should be able to work with a single raw file maybe?
 

ykia said:
I was experimenting with it, too. Here's my take at the Punggol Beach at sunset. The final is a composite of 7 images (-2 to +2 ev)

I just saved your "underexposed" picture and adjusted the characteristic curve - in less than a minute, I arrived at an image that is virtually indistinguishable from your 7-exposure-HDR picture.

The real difference should be in the noise level of the resulting image, which is hardly visible in scaled-down images. Would you mind posting some crops of the images (highlights, mid-range, and shadow areas) at full size? I'm curious how much difference photoshop's "HDR" really makes, compared to e.g. an image produced from a "raw" file (with more than 24 bit colour depth).
 

Fantastic...I wan the software...can download from web?
 

rainman said:
Fantastic...I wan the software...can download from web?
Can, buy online lor. hahaha
 

LittleWolf said:
I just saved your "underexposed" picture and adjusted the characteristic curve - in less than a minute, I arrived at an image that is virtually indistinguishable from your 7-exposure-HDR picture.

The real difference should be in the noise level of the resulting image, which is hardly visible in scaled-down images. Would you mind posting some crops of the images (highlights, mid-range, and shadow areas) at full size? I'm curious how much difference photoshop's "HDR" really makes, compared to e.g. an image produced from a "raw" file (with more than 24 bit colour depth).

As far as I know, the real advantage of HDR is in print. The difference will not be obvious on computer monitors because monitors are only capable of producing 24bit colour, whilst HDR imgaes can typically be 48bit. Please correct me if im wrong..
 

Shaff said:
As far as I know, the real advantage of HDR is in print. The difference will not be obvious on computer monitors because monitors are only capable of producing 24bit colour, whilst HDR imgaes can typically be 48bit. Please correct me if im wrong..

I don't know for sure, but I doubt that prints (or even monitors) would benefit from more than the usual 24 bit colour depth. Where additional precision becomes important is during processing: if previously imperceptible details in the shadows get revealed, so are previously imperceptible image artefacts.

I feel, however, that a lot of comparisons aren't fair. One can get a lot of details out of a single digital image file. A quick (and admittedly not very scientific) test on the raw output file of an entry-level DSLR indicated a dynamic range of somewhere between 8 and 9 stops.

But many comparisons are done using out-of-camera JPEG files and are thus fundamentally flawed: JPEG is designed to eliminate as much image information as possible from the file without causing perceptible artefacts. This also means that processing of a JPEG image that is suited to reveal previously imperceptible details will uncover compression artefacts - and the better the compression algorithm is, the worse will those artefacts be, irrespective on how good the image sensor in the camera is.
 

loupgarou said:
hmm": I haven't tested this feature yet, should be able to work with a single raw file maybe?

No it will not work with a single raw file. The program somehow knows they are from the same file... even if u rename. U need different real frames.
 

Wow! Since there is so much interest in this PS CS2 feature, I have copied and pasted the relevant section from the HELP CENTER.

- What does it do?
High Dynamic Range (HDR) images open up a world of possibilities because they can represent the entire dynamic range of the visible world. Because all the luminance values in a real-world scene are represented proportionately and stored in an HDR image, adjusting the exposure of an HDR image is like adjusting the exposure when photographing a scene in the real world. This capability lets you create blurs and other real-world lighting effects that look realistic. Currently, HDR images are used mostly in motion pictures, special effects, 3D work, and some high-end photography.

You can create an HDR image using multiple photographs, each captured at a different exposure. In Photoshop, the Merge To HDR command lets you create HDR images from multiple photographs. Because an HDR image contains brightness levels that far exceed the display capabilities of a standard 24‑bit monitor or the range of tones in a printed image, Photoshop lets you adjust the preview of the HDR image so it can be viewed on a computer monitor. Some Photoshop tools, adjustments, and filters can be used with HDR images. If you need to print the image or use Photoshop tools and filters that don’t work with HDR images, you can convert the HDR image to an 8‑ or 16‑bits-per-channel image.

- How do you do it?
Keep the following tips in mind when you take photos to be combined with the Merge To HDR command:
* Secure the camera to a tripod.
* Take enough photos to cover the full dynamic range of the scene. You can try taking at least five to seven photos, but you might need to take more exposures depending on the dynamic range of the scene. The minimum number of photos should be three.
* Vary the shutter speed to create different exposures. Changing the aperture changes the depth of field in each exposure and can produce lower-quality results. Changing the ISO or aperture may also cause noise or vignetting in the image.
* In general, don’t use your camera’s auto-bracket feature, because the exposure changes are usually too small.
* The exposure differences between the photos should be one or two EV (exposure value) steps apart (equivalent to about one or two f‑stops apart).
* Don’t vary the lighting; for instance, don’t use a flash in one exposure but not the next.
* Make sure that nothing is moving in the scene. Exposure Merge works only with differently exposed images of the identical scene.

Download the PS CS2 tryout and, well, give it a try!
 

marcwang said:
No it will not work with a single raw file. The program somehow knows they are from the same file... even if u rename. U need different real frames.


I don't know about that, I've been doing the digital blending technique with a single raw file (ie: I save it as 3 or 2 copies of PSD files with different exposure settings (ie: mid, highlights, shadows))

as usual, must experiment
 

loupgarou said:
I've been doing the digital blending technique with a single raw file (ie: I save it as 3 or 2 copies of PSD files with different exposure settings (ie: mid, highlights, shadows))

What does one gain from this procedure, compared to converting the raw data to a 48 bpp file and working on that? Is this simply a workaround for not having an image manipulation program that can handle more than 24 bpp?
 

Status
Not open for further replies.