Personally, I find that to be OK. At a low ISO the increase in noise is minimal. before the *ist-D I routinely shot with my old CP990 set to -1/3 exposure comp in order to prevent blown out highlights - the conversion process in that camera routinely put a good chunk of the pixels to full white. If something was lost in one of those bright areas, it was just lost. Far better (IMO) to have to pump up the brightness, even if that introduces a little noise, and not loose a critical detail.
That's not to say there should be no pure whites or blacks in an image. It's just that I prefer to make the decision about how the histogram gets clipped. and not have to live with the camera's decision.
- MCC
- - - - - - - - - - - - - - - - - - - -
Mark Cassino Photography
Kalamazoo, MI
www.markcassino.com
- - - - - - - - - - - - - - - - - - - -
----- Original Message ----- From: "Gonz" <[EMAIL PROTECTED]>
To: <pentax-discuss@pdml.net>
Sent: Wednesday, February 23, 2005 4:26 PM
Subject: Help! *istD problems (repost)
The second problem is that I seem to always be adjusting exposure upwards in the PS raw program, usually between .5 and 1.0 stops. What should a "good" histogram look like? My camera seems to like to underexpose, in my opinion. Is this normal because of the blown hilites problem? When I bring in the raw pics into PS raw, the histogram usually shows no pixel values above the halfway point, yet if I look at the histogram using the *istD lcd panel, it seems to show some luminance values, although very few of them, up to the max luminance value. Why is there a discrepancy? I am going to try the same thing with Pentax raw and see if the histogram matches PS raw to some degree. I know that PS camera raw breaks up the values into RGB components, while Pentax combines them into a single histogram, so there might be an apples-oranges thing going on here.
Anyhow, I would love to tap the collective PDML brains to see what I need to do here to solve some of these technical issues. Thanks ahead of time for any help you can give me.
rg
