more low-light details from the Mark III ?

[Gregory]

a friend shoots exclusively in mid- and large-format black & white. he will repeatedly visit the same scenery locations over a period of years, only shooting one or two images and only shooting when the scenery is special enough to warrant the picture. he's patient and intelligent and his photos are delightful.

the dynamic range of his photos is astounding, at least compared to what I can get from my 350D. some of his photos from within a recent film-shooting location show great detail from the dark corners of the wooden floor to the outlines of the lit light bulbs in the lamps hanging from the ceiling. I can't do this with my 350D. I can either have detail in the dark and blow the highlights or have detail in the highlights and black out the dark areas.

is this a limitation of digital photography, or of my 350D? will the Mark III (I have one on order) be better with the extra bits it provides?

I know very little about the technology but I'm conjecturing that with digital sensors, the compression (or sampling or sensitivity) of the light across the range from black to white is not even, that the compression is more severe (i.e., fewer levels) at the lower and upper regions so that less detail is retained in these regions than in the middle region. I know that in my bird photos, detail is quite often very murky in the darker areas making it very difficult to get 'acceptable' photos of dark and black feathered birds without 'shooting to the right' and losing details in the brighter areas of the photos.

in my quest to get more detail and a greater range of light in my photos, any enlightenment would be appreciated.

regards,
Gregory

[larsrc]

a friend shoots exclusively in mid- and large-format black & white. he will repeatedly visit the same scenery locations over a period of years, only shooting one or two images and only shooting when the scenery is special enough to warrant the picture. he's patient and intelligent and his photos are delightful.

the dynamic range of his photos is astounding, at least compared to what I can get from my 350D. some of his photos from within a recent film-shooting location show great detail from the dark corners of the wooden floor to the outlines of the lit light bulbs in the lamps hanging from the ceiling. I can't do this with my 350D. I can either have detail in the dark and blow the highlights or have detail in the highlights and black out the dark areas.

is this a limitation of digital photography, or of my 350D? will the Mark III (I have one on order) be better with the extra bits it provides?

I know very little about the technology but I'm conjecturing that with digital sensors, the compression (or sampling or sensitivity) of the light across the range from black to white is not even, that the compression is more severe (i.e., fewer levels) at the lower and upper regions so that less detail is retained in these regions than in the middle region. I know that in my bird photos, detail is quite often very murky in the darker areas making it very difficult to get 'acceptable' photos of dark and black feathered birds without 'shooting to the right' and losing details in the brighter areas of the photos.

in my quest to get more detail and a greater range of light in my photos, any enlightenment would be appreciated.

regards,
Gregory
[a href=\"index.php?act=findpost&pid=130619\"][{POST_SNAPBACK}][/a]

You're halfway right.  The number of levels differs across the range, but most are in the upper levels.  In fact, the f-stop just before blow-out has half the information levels, the next f-stop down has half of the rest, etc.  The shadows end up with very few levels, and they tend to get lost in the noise eventually.  However, the tricky thing is that there is a sharp cut-off at the top, with practically no "shoulder".  You can see this clearly if you compare pictures with the sun in the frame between digital and film, in digital either the sun will have a halo of blown-out or else the subjects will be left in the dark.  While the Mark III will help you somewhat, it's not going to be a panacea for this problem.  There are various HDR techniques that can help, but they come with their own limitations, mainly of having to take more than one shot of the subject and having to merge them afterwards.

-Lars

[Gregory]

thank you Lars. that was very educational.

is this a characteristic of all digital sensors or just CMOS sensors?

having very few levels in the darker region of the light range is very discouraging. no wonder film looks so much better.

HDR looks interesting but it would require stationary subjects and usually a tripod; not very suitable for wildlife photography. and it's more work that it should be.

regards,
Gregory

[larsrc]

thank you Lars. that was very educational.

is this a characteristic of all digital sensors or just CMOS sensors?

having very few levels in the darker region of the light range is very discouraging. no wonder film looks so much better.

HDR looks interesting but it would require stationary subjects and usually a tripod; not very suitable for wildlife photography. and it's more work that it should be.

regards,
Gregory
[a href=\"index.php?act=findpost&pid=130626\"][{POST_SNAPBACK}][/a]

That's inherent to all current digital cameras.  There was some news about half a year back (I think) about a different sensor with much better DR [1], but it was only at the research stage.  However, you do have a pretty good range with an SLR, and mkIII should give you better than 350D.  Good luck with it!

-Lars

[1] Instead of letting each sensor "fill up" and then "measure the fullness", it would let each sensor count the number of times it gets to a given (low) level of fullness.  If it can count fast enough (order of 100.000 counts per sensor per exposure), you can get more bits -- while the distribution of levels would be similar to current sensors, the upper limit is not on how much charge a sensor can contain, but on how fast it can count them, thus changing the upper limit (blow-out limit) from being determined by the sensor size to being determined by speed of the logic, something that's probably easier to improve.  Sounded promising, but I'm not holding my breath.