I'll be the first to admit that the digital color stuff still gives me a hard time. I'm not resistant when it comes to modern digital technology (and converted from the typewriter to the word processor very quickly), but all of the parameters in digital photography can be a bit overwhelming after decades in the analog world. I've been leaning toward shooting digital as though it were color slide film, and that's why the -1 exposure comp. The lighting conditions were a bit extreme (shade with some blown out sunlit areas, and theater, with some blown out highlights on the actor's face).
I'll try to get my lingo right. ;-) Old habits die hard. I also still think of my Pentax lenses as "x mm equivalent", which I need to stop doing. A 25 1.4 isn't all that much like a 50 1.4, even if it is on an Olympus E-1. The field depth is astronomical. Jeffery On Nov 18, 2010, at 4:08 PM, Godfrey DiGiorgi wrote: > On Thu, Nov 18, 2010 at 12:10 PM, Jeffery Smith <jsmith...@bellsouth.net> > wrote: >> Photoshop allows a large measure of filter "effects" in converting color to >> B&W, using slider tools that show the preview of the filter's effects. > > "Filter effects" isn't the right expression, nor is "converting color > to B&W" when you're talking about digital capture to producing B&W > results. It's somewhat sloppy language that has somehow become the > standard. > > When you make an exposure with a digital camera and save it as a raw > file, you are capturing an [x,y] array of linear gamma luminance > intensities organized in an RGB mosaic. The process of 'raw > conversion' into something intelligible to our eyes involves an > interpolation of those RGB values into chrominance values per pixel > and gamma correction of the luminance intensities to suit the way our > eyes and brain work. A more precise word than 'conversion' is > 'rendering'. > > Rendering an image to monochrome values rather than RGB values, well, > since what we appreciate as B&W photography is a translation of color > values into luminances without chroma, and B&W films and colored > filters help us to control that translation by separating or smashing > together color/intensity values into luminance values, what you're > doing with the sliders in Photoshop is directly analogous to putting > filters on the lens when exposing B&W film. It's not a "filter > effect": it's filtering, period. ;-) > >> The biggest problem I have when shooting digital is blown highlights that I >> cannot burn and basically have to replace using cut a paste from an >> adjoining area. I've gotten to the point that I often have to use exposure >> compensation of -1 to avoid blown highlights. > > That's a matter of proper exposure for the digital capture medium, > Jeffery, which requires a different approach than metering for film > negatives. You should only very rarely have to use negative EV > Compensation UNLESS your subject matter is mostly dark and the > significant area where you need detail is mostly bright highlights. (I > would say if I looked at exposure compensation values for all of my > past couple years exposures that my properly exposed photos showed a > 100:1 preponderance of +EV valued EV Compensation, not -EV values...) > > The old adage used for negative film was "expose for the shadows, > develop for the highlights", the notion being to get enough light > energy onto the medium to activate the chemicals and record detail > where you wanted it in the dark areas, and then control the gamma (or > contrast curve) to keep from blocking up the highlights through > development techniques. > > Digital capture sensors, as said above, always capture in a linear > gamma. Their behavior at the limits of exposure are different from > film media: the highlight limit is a hard stop when the photosite > cannot record any additional light energy, the minimum exposure limit > is a soft threshold where detail can no longer be distinguished from > noise. Another factor: since the capture gamma is linear and has to be > stretched and squeezed into a more curvaceous shape for our eyes and > brain to interpret it correctly, it turns out that we need to stuff as > many bits towards the high end of the range as we can (without hitting > the saturation limit) so that we can stretch the values down into the > low end without losing too much data along the way. > > So the goal in exposing properly for a digital sensor is to consider > them as more similar to transparency film ... "Avoid over-exposure on > the highlights like the dickens and let the rest fall where it might" > ... but with a lot more control since we can push the rendering curve > around with great freedom in the raw processing phase. I usually look > at a scene with the idea of evaluating a) what's the overall > reflectivity of the scene? and b) where are my Zone IX highlight > values? A scene which has a lot of bright in it and a small contrast > ratio to handle usually means adding exposure from an averaging > meter's normal recommendation (most scenes, as it turns out). A scene > which has big contrast and small areas of significant Zone IX detail > that has to be preserved is one where I will pull down the EV > compensation to keep from overexposing the details I want and let the > rest fall into blackness (if it's outside the DR of the sensor). These > latter are where the Spot metering pattern is helpful in evaluation, > but don't let it do the whole job for you... it's pretty dumb in AE > mode. > > "Expose for the highlights, and render for the shadows" is a > simplification of the technique often called "Expose to the Right", > but it works. It presupposed raw capture, btw, because JPEGs don't > have the range of adjustability required for processing high contrast > scenes most of the time. > > But I'm beginning to ramble ... ;-) > > > -- > Godfrey > godfreydigiorgi.posterous.com > > -- > PDML Pentax-Discuss Mail List > PDML@pdml.net > http://pdml.net/mailman/listinfo/pdml_pdml.net > to UNSUBSCRIBE from the PDML, please visit the link directly above and follow > the directions. -- PDML Pentax-Discuss Mail List PDML@pdml.net http://pdml.net/mailman/listinfo/pdml_pdml.net to UNSUBSCRIBE from the PDML, please visit the link directly above and follow the directions.
