HDR images

[Jim Benton]

Seems rather a statement of fact to me,

If you think the statement that his article was full of inaccuracies, and false information was a fact, please tell us all what these are as I can't see anything wrong and I don't want to go away with the wrong information.

however this statement, "Your rather puerile head in the sand approach does not... "

Oh, come on!

Putting someone in your kill file is certainly a 'head in the sand' approach. Maybe a sensible thing to do in some cases but then taking pot-shots at the person you've kill filed is childish in the extreme.

If someone has told Bifurcator about various errors, why did neither the informant or bifucator expose them rather than just making vague, unsupported accusations?

 

[Dao]

Got this information from HdrSoft, the maker of Photomatirx.

----------------------------------------------------------------------------
Direct Quote:

What is an HDR image?


The Dynamic Range of real-world scenes can be quite high -- ratios of 100,000:1 are common in the natural world. An HDR (High Dynamic Range) image stores pixel values that span the whole tonal range of real-world scenes. Therefore, an HDR image is encoded in a format that allows the largest range of values, e.g. floating-point values stored with 32 bits per color channel.


Another characteristics of an HDR image is that it stores linear values. This means that the value of a pixel from an HDR image is proportional to the amount of light measured by the camera. In this sense, HDR images are scene-referred, representing the original light values captured for the scene.


Whether an image may be considered High or Low Dynamic Range depends on several factors. Most often, the distinction is made depending on the number of bits per color channel that the digitized image can hold. However, the number of bits itself may be a misleading indication of the real dynamic range that the image reproduces -- converting a Low Dynamic Range image to a higher bit depth does not change its dynamic range, of course.

· 8-bit images (i.e. 24 bits per pixel for a color image) are considered Low Dynamic Range.

· 16-bit images (i.e. 48 bits per pixel for a color image) resulting from RAW conversion are still considered Low Dynamic Range, even though the range of values they can encode is much higher than for 8-bit images (65536 versus 256). Converting a RAW file involves applying a tonal curve that compresses the dynamic range of the RAW data so that the converted image shows correctly on low dynamic range monitors. The need to adapt the output image file to the dynamic range of the display is the factor that dictates how much the dynamic range is compressed, not the output bit-depth. By using 16 instead of 8 bits, you will gain precision but you will not gain dynamic range.

· 32-bit images (i.e. 96 bits per pixel for a color image) are considered High Dynamic Range. Unlike 8- and 16-bit images which can take a finite number of values, 32-bit images are coded using floating point numbers, which means the values they can take is unlimited. It is important to note, though, that storing an image in a 32-bit HDR format is a necessary condition for an HDR image but not a sufficient one. When an image comes from a single capture with a standard camera, it will remain a Low Dynamic Range image, regardless of the format used to store it.



There are various formats available to store HDR images, such as Radiance RGBE (.hdr) and OpenEXR (.exr) among the most commonly used. See Greg Ward's HDR Image Encodings page for an excellent overview of HDR formats.
----------------------------------------------------------------------------


Being a beginner, and after I read the information I got from Wikipedia and HdrSoft, I tend to lean more on Helen and Bifurcator side. Moglex, don't get me wrong, I do not mean you are wrong, and I DO really appericate you bring this subject up so that I can do more reading. It's just that the information I read match better with what Bifurcator said. Maybe I just dumb and not quite get some of the information you provide.

 

[Dao]

BTW, the FAQ from HDRSoft has some good info on HDR images for Photography.

http://www.hdrsoft.com/resources/dri.html

 

[Moglex]

Being a beginner, and after I read the information I got from Wikipedia and HdrSoft, I tend to lean more on Helen and Bifurcator side. Moglex, don't get me wrong, I do not mean you are wrong, and I DO really appericate you bring this subject up so that I can do more reading. It's just that the information I read match better with what Bifurcator said. Maybe I just dumb and not quite get some of the information you provide.

Don't worry, you're not dumb! This is really pretty complicated stuff. I'm always prepared to admit that it might be me that has it wrong - all I ask is that people point out particular errors of fact or logic rather than just saying "I'm right and you're wrong" or putting up a smokescreen by concentrating on off topic factors.

Firstly I'd like to address a factual error in what you've quoted. When they say: "32-bit images are coded using floating point numbers, which means the values they can take is unlimited.", this is quite simply completely incorrect. A 32 bit number can store approximately 4.3 billion values. It can, however, store values (much) bigger than 4.3 billion but it is still limited to ~4.3 billion values.

I think the confusion (yours, mine, Helen's, Bifurcators?) stems from this (or something like it): "However, the number of bits itself may be a misleading indication of the real dynamic range that the image reproduces -- converting a Low Dynamic Range image to a higher bit depth does not change its dynamic range, of course."

This is perfectly correct, but you have to be careful to note that it only works one way:

The number of bits is not necessarily an indication of the dynamic range - true.

A higher dynamic range does not necessarily require a greater minimum number of bits than a low dynamic range - false.

In the earlier post I tried to very carefully explain why it is the number of bits (or, to be more accurate, the number of significant bits) that determine the dynamic range of a system (and even provided a link to a source that stated exactly that). This is somthing that Helen B specifically denied.

It would be very helpful if you could explain exactly why you tend to lean towards her version of DR rather than mine (and the source I quoted). It will either help me understand where I'm wrong or help me explain why I'm not.

 

[tirediron]

I'm so glad I can't read the above message. The ignore list is a wonderful invention. I'm told however that it's full of accusations, inaccuracies, and false information - so let the reader beware.

Hmmm, "I'm glad", "I'm told"; it seems reasonable to assume that those are statements of fact with respect to Bifurcator's feelings. Whether you, I, or anyone agrees or disagrees with them is irrelevant. There is no attack made on anyone; perhaps a subtle insinuation at most. Entirely different however, than telling someone they are acting childishly. You can hardly disagree Moglex's comment is aimed directly at someone. Again, concurrence or not with the statement has no bearing on the situation.

I'm not denying Moglex seems to have a great deal of knowledge on the matter, however other people whom I know to be very knowledgable disagree with him. One side is right, one side is wrong. I see no reason why Moglex needs to resort to insulting remarks.

 

[Jim Benton]

LOL.

So if Moglex had said that he'd been told that Bifurcator was being childish that would have been a statement of fact and OK?

I don't see why Bifurcator's being rude about Moglex's post is acceptable just because he mentions a third party.

The real problem, though is that Birfurcator, Moglex and Dao have all made posts about the subject in hand whereas all you're doing is talking about posts, a job I would have though would be much better left to the moderators who could have a quiet word with Bifurcator and/or Moglex if they think they've overstepped the bounds.

It seems as if I'm now also guilty of posting about posts rather than the subject but I really want to know if anyone has concrete facts or arguments to counter those of Moglex as it's a very interesting subject and I'd like to go away with the correct information and I don't want to see this thread stopped in its tracks like yesterdays.

 

[Bifurcator]

Got this information from HdrSoft, the maker of Photomatirx.

----------------------------------------------------------------------------
Direct Quote:

· 16-bit images (i.e. 48 bits per pixel for a color image) resulting from RAW conversion are still considered Low Dynamic Range, even though the range of values they can encode is much higher than for 8-bit images (65536 versus 256).

Range values here are the number of steps - not the height of the staircase. 65534 steps (and the two floors )

As should be made obvious from the blue sentence below quoted from the same document.

Converting a RAW file involves applying a tonal curve that compresses the dynamic range of the RAW data so that the converted image shows correctly on low dynamic range monitors. The need to adapt the output image file to the dynamic range of the display is the factor that dictates how much the dynamic range is compressed, not the output bit-depth. By using 16 instead of 8 bits, you will gain precision but you will not gain dynamic range.

· 32-bit images (i.e. 96 bits per pixel for a color image) are considered High Dynamic Range. Unlike 8- and 16-bit images which can take a finite number of values, 32-bit images are coded using floating point numbers, which means the values they can take is unlimited. It is important to note, though, that storing an image in a 32-bit HDR format is a necessary condition for an HDR image but not a sufficient one. When an image comes from a single capture with a standard camera, it will remain a Low Dynamic Range image, regardless of the format used to store it.

As you can see number of steps and staircase height are two different things - as obviously the 32bit HDR staircase isn't infinitely high.

There are various formats available to store HDR images, such as Radiance RGBE (.hdr) and OpenEXR (.exr) among the most commonly used. See Greg Ward's HDR Image Encodings page for an excellent overview of HDR formats.
----------------------------------------------------------------------------


Being a beginner, and after I read the information I got from Wikipedia and HdrSoft, I tend to lean more on Helen and Bifurcator side. Moglex, don't get me wrong, I do not mean you are wrong, and I DO really appericate you bring this subject up so that I can do more reading. It's just that the information I read match better with what Bifurcator said. Maybe I just dumb and not quite get some of the information you provide.

But the way some documents are worded does indeed make it hard to understand and the number of different formats (HDR-Half, HDR, LDR in all the various bit depths) doesn't help. Like the document dklod quoted from luminous landscape used confusing terminology and could be construed to mean either higher stairs or more steps depending what assumptions you brought to it. The above document is much more clear and states precisely "By using 16 instead of 8 bits, you will gain precision but you will not gain dynamic range." meaning more steps, same stair height. HDR is higher stairs and infinite steps.

 

[Moglex]

Range values here are the number of steps - not the height of the staircase. 65534 steps (and the two floors )

Actually, 65,535 steps. Neither floor raises the value and the number of 'steps from 0 tp 2^n is always 2^n - 1.

As should be made obvious from the blue sentence below quoted from the same document.

{quoted: 32-bit images are coded using floating point numbers, which means the values they can take is unlimited}.

Amazingly, Bifurcator has picked the one part of the quoted documant that is absolutely and unequivocaly incorrect.

As anyone who has done computer science at even the most basic level should be able to assure you: No number represented by a finite number of bits can possibly take an unlimited number of values.

That is so fundamental it beggars belief that someone who claims to have a computer science degree, as Bifurcator has, could possibly make such a mistake.

As you can see number of steps and staircase height are two different things

Blindingly obvious which is why you cannot actually see that in the quoted text.

What they've actually said is that becase the imaging device cannot handle more that 8 bits per channel even if you use 16 bits it will be converted with a dynamic range suitable for the output.

as obviously the 32bit HDR staircase isn't infinitely high.

And neither does it have an infinite number of steps!

The above document is much more clear and states precisely "By using 16 instead of 8 bits, you will gain precision but you will not gain dynamic range." meaning more steps, same stair height.

The document also says: "The need to adapt the output image file to the dynamic range of the display is the factor that dictates how much the dynamic range is compressed". This is vital. As someone quoted in a document earlier, more bits does not necessarily mean greater dynamic range. In the case of a conversion from raw to an output intended for a display device it won't because the display device could not handle the range.

That does not mean that you cannot get higher dynamic range by creating a 16 bits/channel image.

HDR is higher stairs and infinite steps.

No it isn't.

Anyone who doesn't believe me just ask any reasonably competant computer science student. No computer has ever been built that is capable of holding a number than can take an infinite range of values (steps).

 

[LaFoto]

Oh well, I take it you better like all the HDR-threads to be closed. So be it. Sigh...

 

[Arch]

yea guys, rather than see this thread locked, i would appreciate it if you could try and stick to the technical facts (which im sure we can all find intriguing) rather than argue on a personal level... so from now on please stop with the 'MRsmith said this.. which is wrong'... just expalin your side and your findings and if people don't buy it then its thier problem.

You may continue (and please do) if you can abide by this request. Thankyou.

 

[photographyaddict]

There are some really great tutorials on YouTube.

[ame]http://www.youtube.com/watch?v=CVuDbcAfN_I&feature=related[/ame]

(This one involves CS3)

I think the hardest part is getting the bracketing right with the initial photography. Have to make sure you get the best range of lighting.

 

[prodigy2k7]

I love these HDR debates. You are all wrong, how about that

 

[Moglex]

I love these HDR debates. You are all wrong, how about that

That wouldn't surprise me in the least.