Digital Dynamic Range and Exposure Latitude

Digital Dynamic Range and Exposure Latitude

Well, I stopped by here for the first time in a year or so.
I see not much has changed.

But in case anyone is interested in some data and images
that illustrate what digital and film can do regarding
dynamic range and exposure latitude, I offer the following.

A big problem with film versus digital and what each
can deliver is that published data are different for the
two technologies. For example, film's characteristic curve,
which looks great, is measured with a 48-micron diameter
spot (that would be about 60+ 5-micron DSLR pixels averaged).

So I tried to measure each with close to the same sampling.
For example, 4000 ppi scanning is a 6.3 micron spot
size, similar to DSLR pixels (in practice the scanner resolution
is not that good, so the result is slightly larger pixels, which
would be closer to a high end DSLR).

Here is one comparison, with images:
Figure 5 compares digital 16-bit, digital jpeg,
print film and slide film:

Dynamic Range and Transfer Functions of Digital Images
and Comparison to Film
http://www.clarkvision.com/imagedetail/dynamicrange2

Then see Figure 8B for characteristic curves of 16-bit digital,
print and slide film. Note how the film loses it in the shadows,
just as the images in Figure 5 show.

What the curves and images don't show is on the print film
you can recover further into the highlights with some different
scanning techniques that I developed after I did this page.
Trying to recover those highlights on film does result in
increased noise and color shifts but at least you can go higher
than digital.

Next look at:
The Exposure Latitude of a Digital Camera
and Comparison to Film
http://www.clarkvision.com/imagedeta...ure_latitude-1

Figure 1 shows the noise analysis of print film and digital. I still
have more points to measure on the film, but I got some of the
high end ones to show how far up the film goes (again with color
shifts).

The images show digital can produce very good images over the
range of +2 to -5 stops and show image detail over 15 stops.
Other digital cameras and meters may shift this up or down,
but should have a similar range.
Yes this can be done with a 12-bit system because you have many
pixels where your eye averages the noise, much like it averages
film grain.

The bottom line is that digital produces better shadow detail,
and print film goes much higher in the highlights. One just
needs to learn to expose the two differently. But their total
range is similar. Slide film is less than both.

But there is more to the story. To achieve the large dynamic range
with digital, one must use raw data and do a good 16-bit conversion
(or 32-bit) with software that can deliver those results.
Not all raw converters do a good job in this regard.
See:

Digital Camera Raw Converter Shadow Detail and Image Editor
Limitations: Factors in Getting Shadow Detail in Images
http://www.clarkvision.com/imagedeta....shadow.detail

I'll close with a high dynamic range image done with a single
exposure and one raw conversion. This was a tough image to acquire
and I exposed it to not over-expose the area around the sun. The
foreground was very dark in the converted image and the dynamic
range had to be compressed a lot to show the range of the scene:
http://www.clarkvision.com/galleries...867.c-800.html

So, good digital cameras with large pixels have a huge dynamic range,
comparable to print film, but shifted to the low end, and digital
has much higher dynamic range than slide film. Note also that DSLRs
will maintain high dynamic range to higher ISOs than film.

Roger

Mr. Clark:
I usually just lurk, but I wanted to say "Thanks" for your clear and
detailed explanation, plus the great examples. I appreciate you taking
the time to work through these examples, and to publish your results.
It clearly took significant time and effort - and I benefitted from
it.

While my focus is much less on the technical details at this level,
and more on the creative process and results, having this information
available from a non-vendor helps me make better choices about the
tools and techniques I choose to use. You confirmed and clarified
things I've been learning the hard way after switching from
transparency film to digital.

We are fortunate to have both the technical and the artistic
tempraments available to us on the web! What a great combination!

Thanks and best regards,
Art

PS - If you ever have the time, I'd really value a comparison of the
Open Source UFRaw converter against the Photoshop ACR CS4 converter.
With many of my friends and colleagues using it, as well as myself on
my field gear since Photoshop licenses are not cheap, I'd really like
to know the technical differences and capabilities.




On Fri, 07 Aug 2009 23:56:27 -0600, "Roger N. Clark (change username
to rnclark)" wrote:

Digital Dynamic Range and Exposure Latitude

Well, I stopped by here for the first time in a year or so.
I see not much has changed.

But in case anyone is interested in some data and images
that illustrate what digital and film can do regarding
dynamic range and exposure latitude, I offer the following.

A big problem with film versus digital and what each
can deliver is that published data are different for the
two technologies. For example, film's characteristic curve,
which looks great, is measured with a 48-micron diameter
spot (that would be about 60+ 5-micron DSLR pixels averaged).

So I tried to measure each with close to the same sampling.
For example, 4000 ppi scanning is a 6.3 micron spot
size, similar to DSLR pixels (in practice the scanner resolution
is not that good, so the result is slightly larger pixels, which
would be closer to a high end DSLR).

Here is one comparison, with images:
Figure 5 compares digital 16-bit, digital jpeg,
print film and slide film:

Dynamic Range and Transfer Functions of Digital Images
and Comparison to Film
http://www.clarkvision.com/imagedetail/dynamicrange2

Then see Figure 8B for characteristic curves of 16-bit digital,
print and slide film. Note how the film loses it in the shadows,
just as the images in Figure 5 show.

What the curves and images don't show is on the print film
you can recover further into the highlights with some different
scanning techniques that I developed after I did this page.
Trying to recover those highlights on film does result in
increased noise and color shifts but at least you can go higher
than digital.

Next look at:
The Exposure Latitude of a Digital Camera
and Comparison to Film
http://www.clarkvision.com/imagedeta...ure_latitude-1

Figure 1 shows the noise analysis of print film and digital. I still
have more points to measure on the film, but I got some of the
high end ones to show how far up the film goes (again with color
shifts).

The images show digital can produce very good images over the
range of +2 to -5 stops and show image detail over 15 stops.
Other digital cameras and meters may shift this up or down,
but should have a similar range.
Yes this can be done with a 12-bit system because you have many
pixels where your eye averages the noise, much like it averages
film grain.

The bottom line is that digital produces better shadow detail,
and print film goes much higher in the highlights. One just
needs to learn to expose the two differently. But their total
range is similar. Slide film is less than both.

But there is more to the story. To achieve the large dynamic range
with digital, one must use raw data and do a good 16-bit conversion
(or 32-bit) with software that can deliver those results.
Not all raw converters do a good job in this regard.
See:

Digital Camera Raw Converter Shadow Detail and Image Editor
Limitations: Factors in Getting Shadow Detail in Images
http://www.clarkvision.com/imagedeta....shadow.detail

I'll close with a high dynamic range image done with a single
exposure and one raw conversion. This was a tough image to acquire
and I exposed it to not over-expose the area around the sun. The
foreground was very dark in the converted image and the dynamic
range had to be compressed a lot to show the range of the scene:
http://www.clarkvision.com/galleries...867.c-800.html

So, good digital cameras with large pixels have a huge dynamic range,
comparable to print film, but shifted to the low end, and digital
has much higher dynamic range than slide film. Note also that DSLRs
will maintain high dynamic range to higher ISOs than film.

Roger

Roger N. Clark (change username to rnclark) wrote:

Digital Dynamic Range and Exposure Latitude


I must take time to read it properly, I've only skimmed it. It looks like
you put a lot of work in, thank you.

Pete

--
http://www.petezilla.co.uk

"Roger N. Clark (change username to rnclark)" wrote:
Digital Dynamic Range and Exposure Latitude

Well, I stopped by here for the first time in a year or so.
I see not much has changed.

But in case anyone is interested in some data and images
that illustrate what digital and film can do regarding
dynamic range and exposure latitude, I offer the following.

A big problem with film versus digital and what each
can deliver is that published data are different for the
two technologies. For example, film's characteristic curve,
which looks great, is measured with a 48-micron diameter
spot (that would be about 60+ 5-micron DSLR pixels averaged).

So I tried to measure each with close to the same sampling.
For example, 4000 ppi scanning is a 6.3 micron spot
size, similar to DSLR pixels (in practice the scanner resolution
is not that good, so the result is slightly larger pixels, which
would be closer to a high end DSLR).

Here is one comparison, with images:
Figure 5 compares digital 16-bit, digital jpeg,
print film and slide film:

Dynamic Range and Transfer Functions of Digital Images
and Comparison to Film
http://www.clarkvision.com/imagedetail/dynamicrange2

Then see Figure 8B for characteristic curves of 16-bit digital,
print and slide film. Note how the film loses it in the shadows,
just as the images in Figure 5 show.

What the curves and images don't show is on the print film
you can recover further into the highlights with some different
scanning techniques that I developed after I did this page.
Trying to recover those highlights on film does result in
increased noise and color shifts but at least you can go higher
than digital.

Next look at:
The Exposure Latitude of a Digital Camera
and Comparison to Film
http://www.clarkvision.com/imagedeta...ure_latitude-1

Figure 1 shows the noise analysis of print film and digital. I still
have more points to measure on the film, but I got some of the
high end ones to show how far up the film goes (again with color
shifts).

The images show digital can produce very good images over the
range of +2 to -5 stops and show image detail over 15 stops.
Other digital cameras and meters may shift this up or down,
but should have a similar range.
Yes this can be done with a 12-bit system because you have many
pixels where your eye averages the noise, much like it averages
film grain.

The bottom line is that digital produces better shadow detail,
and print film goes much higher in the highlights. One just
needs to learn to expose the two differently. But their total
range is similar. Slide film is less than both.

But there is more to the story. To achieve the large dynamic range
with digital, one must use raw data and do a good 16-bit conversion
(or 32-bit) with software that can deliver those results.
Not all raw converters do a good job in this regard.
See:

Digital Camera Raw Converter Shadow Detail and Image Editor
Limitations: Factors in Getting Shadow Detail in Images
http://www.clarkvision.com/imagedeta....shadow.detail

I'll close with a high dynamic range image done with a single
exposure and one raw conversion. This was a tough image to acquire
and I exposed it to not over-expose the area around the sun. The
foreground was very dark in the converted image and the dynamic
range had to be compressed a lot to show the range of the scene:
http://www.clarkvision.com/galleries...867.c-800.html

So, good digital cameras with large pixels have a huge dynamic range,
comparable to print film, but shifted to the low end, and digital
has much higher dynamic range than slide film. Note also that DSLRs
will maintain high dynamic range to higher ISOs than film.

Thanks a lot, Roger, for the very informative explanation and links!

--
Chris Malcolm

On 11 Aug 2009 20:22:39 GMT, Chris Malcolm wrote:

"Roger N. Clark (change username to rnclark)" wrote:
Digital Dynamic Range and Exposure Latitude

Well, I stopped by here for the first time in a year or so.
I see not much has changed.

But in case anyone is interested in some data and images
that illustrate what digital and film can do regarding
dynamic range and exposure latitude, I offer the following.

A big problem with film versus digital and what each
can deliver is that published data are different for the
two technologies. For example, film's characteristic curve,
which looks great, is measured with a 48-micron diameter
spot (that would be about 60+ 5-micron DSLR pixels averaged).

So I tried to measure each with close to the same sampling.
For example, 4000 ppi scanning is a 6.3 micron spot
size, similar to DSLR pixels (in practice the scanner resolution
is not that good, so the result is slightly larger pixels, which
would be closer to a high end DSLR).

Here is one comparison, with images:
Figure 5 compares digital 16-bit, digital jpeg,
print film and slide film:

Dynamic Range and Transfer Functions of Digital Images
and Comparison to Film
http://www.clarkvision.com/imagedetail/dynamicrange2

Then see Figure 8B for characteristic curves of 16-bit digital,
print and slide film. Note how the film loses it in the shadows,
just as the images in Figure 5 show.

What the curves and images don't show is on the print film
you can recover further into the highlights with some different
scanning techniques that I developed after I did this page.
Trying to recover those highlights on film does result in
increased noise and color shifts but at least you can go higher
than digital.

Next look at:
The Exposure Latitude of a Digital Camera
and Comparison to Film
http://www.clarkvision.com/imagedeta...ure_latitude-1

Figure 1 shows the noise analysis of print film and digital. I still
have more points to measure on the film, but I got some of the
high end ones to show how far up the film goes (again with color
shifts).

The images show digital can produce very good images over the
range of +2 to -5 stops and show image detail over 15 stops.
Other digital cameras and meters may shift this up or down,
but should have a similar range.
Yes this can be done with a 12-bit system because you have many
pixels where your eye averages the noise, much like it averages
film grain.

The bottom line is that digital produces better shadow detail,
and print film goes much higher in the highlights. One just
needs to learn to expose the two differently. But their total
range is similar. Slide film is less than both.

But there is more to the story. To achieve the large dynamic range
with digital, one must use raw data and do a good 16-bit conversion
(or 32-bit) with software that can deliver those results.
Not all raw converters do a good job in this regard.
See:

Digital Camera Raw Converter Shadow Detail and Image Editor
Limitations: Factors in Getting Shadow Detail in Images
http://www.clarkvision.com/imagedeta....shadow.detail

I'll close with a high dynamic range image done with a single
exposure and one raw conversion. This was a tough image to acquire
and I exposed it to not over-expose the area around the sun. The
foreground was very dark in the converted image and the dynamic
range had to be compressed a lot to show the range of the scene:
http://www.clarkvision.com/galleries...867.c-800.html

So, good digital cameras with large pixels have a huge dynamic range,
comparable to print film, but shifted to the low end, and digital
has much higher dynamic range than slide film. Note also that DSLRs
will maintain high dynamic range to higher ISOs than film.

Thanks a lot, Roger, for the very informative explanation and links!

And of course, that someone who has never tested these things for
themselves would find out that "x-spurt" Roger is in total error on 80% of
the crap he leaves laying around on the net. But go ahead, believe it if
you don't test it for yourself. Fools are like that.

The Believe Crap wrote:
On 11 Aug 2009 20:22:39 GMT, Chris Malcolm wrote:

"Roger N. Clark (change username to rnclark)" wrote:
Digital Dynamic Range and Exposure Latitude

Well, I stopped by here for the first time in a year or so.
I see not much has changed.

But in case anyone is interested in some data and images
that illustrate what digital and film can do regarding
dynamic range and exposure latitude, I offer the following.

[Excellent explanation snipped.]

Thanks a lot, Roger, for the very informative explanation and links!

And of course, that someone who has never tested these things for
themselves would find out that "x-spurt" Roger is in total error on 80% of
the crap he leaves laying around on the net. But go ahead, believe it if
you don't test it for yourself. Fools are like that.

I don't have the equipment, discipline, and probably the education to
do the comprehensive testing that Roger does. But as a retired
scientist with some optical expertise I like to do what testing I
can. In this case, as is usual with Roger's posts, I find that his
test results are consonant not only with my own much lesser efforts
but those of others who have gone further and whom I have learned to
trust.

I have a tendency to suppose that insolent people with bad grammar and
spelling who don't supply evidence or argument to back up their
vituperative postings are just angry fools, but I realise that's
stereotyping and I'm trying to avoid that. So in case you do know
what you're talking about I invite you to post here the evidence and
arguments that show that Roger's claims about dynamic range are
mistaken. I'm sure I won't be alone in finding that interesting and
well worth studying.

--
Chris Malcolm

On 12 Aug 2009 11:14:03 GMT, Chris Malcolm wrote:

invite you to post here the evidence and
arguments that show that Roger's claims about dynamic range are
mistaken.

Already been done, dozens of times. What rock are you living under, troll?

In article , Good Grief
wrote:

On 12 Aug 2009 11:14:03 GMT, Chris Malcolm wrote:

invite you to post here the evidence and
arguments that show that Roger's claims about dynamic range are
mistaken.

Already been done, dozens of times. What rock are you living under, troll?

Should that not be a bridge?

I'm with Chris, I'm also a newbie who does not yet understand why
vituperation is essential in every post here.

Could you provide a list of 24 article ID's (min value of "dozens")
where some or all of Roger's claims have been logically and
scientifically shown to have been mistaken? Roger's maths and
methodology look OK to me.

If you don't have the list, give me something I can Google, preferably
articles that contain something more grown up than insult flinging if
at all possible.

Oh, I don't suppose you have a real name? Hiding behind "Good Grief" is
more grief than good when expecting readers to pay attention is it not?

--
To de-mung my e-mail address:- fsnospam$elliott$$
PGP Fingerprint: 1A96 3CF7 637F 896B C810 E199 7E5C A9E4 8E59 E248

Good Grief wrote:
On 12 Aug 2009 11:14:03 GMT, Chris Malcolm wrote:

invite you to post here the evidence and
arguments that show that Roger's claims about dynamic range are
mistaken.

Already been done, dozens of times. What rock are you living under, troll?

I asked for evidence and argument.

--
Chris Malcolm

The Believe Crap wrote:
And of course, that someone who has never tested these things for
themselves would find out that "x-spurt" Roger is in total error on 80% of
the crap he leaves laying around on the net. But go ahead, believe it if
you don't test it for yourself. Fools are like that.

I probably argue technical details and their
interpretation with Roger Clark more often than any
other individual. Roger's work is right at least 98% of
the time. He is human, he does make mistakes and like
most humans he gets defensive. But the *facts* are that
Roger produces more, and better, information than just
about anyone.

--
Floyd L. Davidson http://www.apaflo.com/floyd_davidson
Ukpeagvik (Barrow, Alaska)