As promised, a discussion of my workflow and philosophy.
I shoot in Large Fine Jpeg, I have no experience with RAW, but I understand the quality is better if you go that route. My reasons are that the incremental improvement in quality is not worth the extra processing time, the content of the shot matters more than the image quality, and I can hardly tell the difference anyway.
First thing I do is shoot LOTS and LOTS of pictures. Everytime my wife asks me whether I got any good ones, when I've finished a 256MB card, I will deadpan "One or two". It's a standing joke between us, but usually quite true. Out of a couple hundred shots, only one or two will be keepers, and mostly caught by accident.
Next thing is to delete like mad when I've transferred the pictures to the computer. It might seem to defeat the purpose of taking lots of shots, but by deleting those obviously out-of-focus and awful moments shots (eg eyes half closed) which you'll never use anyway, you can find your good shots much quicker.
Then it comes down to choosing the best out of two or three of the same shot. Some shots are instant keepers with priceless expressions, some you have to agonize and choose between.
Now, let's say I've picked a nice picture to work on. First thing I do when I open it in PS is to crop it. I crop to remove distractions and recompose the shot, moving the subject to the side instead of dead centre, and rotate the crop by 1 or 2 degrees where necessary. If there's a diagonal line in the picture I try to make it end in a corner. I try to stick to the 3:2 ratio (usually 600 x 400), because that looks most natural and pleasing, and also subliminally implies that you did NOT crop, even though you did. If really cannot, then I will try a square format or panoramic. 4:3 ratio pictures just cry "Amateur!". Really.
After I crop, I will do an Auto-Levels and see if that improves the picture or destroys it (eg by changing the contrast or colour cast). If it looks horrid I simply use Ctrl-Z to undo the Auto-levels and adjust the levels manually. This usually entails moving the sliders on each end inwards to the foot of the histogram "mountain", then adjusting the centre slider until the details can be seen to your liking. If the slider on the right needs to be moved in too much, you usually end up with blown-out highlights. One way to avoid this is to then adjust the curves instead. The theory is a bit too complicated to go into, but suffice to say that using two points to make the curve "S-shaped" usually improves the contrast without blowing out the highlights or killing the shadows (which is desirable if you're doing B&W anyway).
Next step is to see if the picture works better in colour or B&W. Generally I prefer B&W because facial expressions are better appreciated in B&W, whereas colour distracts. But if the colours add something to the picture (eg similar coloured clothing or highly contrasting objects) then I will keep it in colour. The way to do this, of course, is simply to convert to grayscale and see if you like the result. If not, simply Undo the conversion. I tried the Lab Luminance route, but in the end I found that simple conversion to grayscale works best for me. The trick is in adjusting the levels again after converting to grayscale. Increasing the contrast (in the manner described above) works wonders. "Normal" grayscale usually lacks contrast and therefore "punch".
Once I am happy with that, I will decide if any further effects would help, like adding grain to the picture or a softness filter (the technique has already been described). But be cautioned that for this sort of thing, less is more. Don't get carried away, or your pictures will end up looking like Primary School Paint Shop Pro projects.
The final steps involve sharpening the picture using Edgesharpen (I posted the action somewhere - do a search) and reducing the noise if excessive. The principle behind these two actions are very simple - only the parts which need sharpening or noise-reducing will have the filter applied to them. This is done by applying the filter through a mask, which has been defined by the "Find edges" filter. These two actions are assigned to Function keys, so it's a matter of pressing these function keys.
Then comes the framing. I add a single-pixel black border, then use the "Drop Shadow (full image)" action that comes as a default with PS 4.0. Again, these are assigned to function keys, and two keystrokes complete the work.
And that's it. All my kung-fu secrets exposed. I know it's all text and no pictures, I'm too lazy to work on that right now. Maybe later. In the meantime (in the spirit of Sulhan), if you're not sure about anything I'm talking about, feel free to ask and I'll be glad to explain.
I shoot in Large Fine Jpeg, I have no experience with RAW, but I understand the quality is better if you go that route. My reasons are that the incremental improvement in quality is not worth the extra processing time, the content of the shot matters more than the image quality, and I can hardly tell the difference anyway.
First thing I do is shoot LOTS and LOTS of pictures. Everytime my wife asks me whether I got any good ones, when I've finished a 256MB card, I will deadpan "One or two". It's a standing joke between us, but usually quite true. Out of a couple hundred shots, only one or two will be keepers, and mostly caught by accident.
Next thing is to delete like mad when I've transferred the pictures to the computer. It might seem to defeat the purpose of taking lots of shots, but by deleting those obviously out-of-focus and awful moments shots (eg eyes half closed) which you'll never use anyway, you can find your good shots much quicker.
Then it comes down to choosing the best out of two or three of the same shot. Some shots are instant keepers with priceless expressions, some you have to agonize and choose between.
Now, let's say I've picked a nice picture to work on. First thing I do when I open it in PS is to crop it. I crop to remove distractions and recompose the shot, moving the subject to the side instead of dead centre, and rotate the crop by 1 or 2 degrees where necessary. If there's a diagonal line in the picture I try to make it end in a corner. I try to stick to the 3:2 ratio (usually 600 x 400), because that looks most natural and pleasing, and also subliminally implies that you did NOT crop, even though you did. If really cannot, then I will try a square format or panoramic. 4:3 ratio pictures just cry "Amateur!". Really.
After I crop, I will do an Auto-Levels and see if that improves the picture or destroys it (eg by changing the contrast or colour cast). If it looks horrid I simply use Ctrl-Z to undo the Auto-levels and adjust the levels manually. This usually entails moving the sliders on each end inwards to the foot of the histogram "mountain", then adjusting the centre slider until the details can be seen to your liking. If the slider on the right needs to be moved in too much, you usually end up with blown-out highlights. One way to avoid this is to then adjust the curves instead. The theory is a bit too complicated to go into, but suffice to say that using two points to make the curve "S-shaped" usually improves the contrast without blowing out the highlights or killing the shadows (which is desirable if you're doing B&W anyway).
Next step is to see if the picture works better in colour or B&W. Generally I prefer B&W because facial expressions are better appreciated in B&W, whereas colour distracts. But if the colours add something to the picture (eg similar coloured clothing or highly contrasting objects) then I will keep it in colour. The way to do this, of course, is simply to convert to grayscale and see if you like the result. If not, simply Undo the conversion. I tried the Lab Luminance route, but in the end I found that simple conversion to grayscale works best for me. The trick is in adjusting the levels again after converting to grayscale. Increasing the contrast (in the manner described above) works wonders. "Normal" grayscale usually lacks contrast and therefore "punch".
Once I am happy with that, I will decide if any further effects would help, like adding grain to the picture or a softness filter (the technique has already been described). But be cautioned that for this sort of thing, less is more. Don't get carried away, or your pictures will end up looking like Primary School Paint Shop Pro projects.
The final steps involve sharpening the picture using Edgesharpen (I posted the action somewhere - do a search) and reducing the noise if excessive. The principle behind these two actions are very simple - only the parts which need sharpening or noise-reducing will have the filter applied to them. This is done by applying the filter through a mask, which has been defined by the "Find edges" filter. These two actions are assigned to Function keys, so it's a matter of pressing these function keys.
Then comes the framing. I add a single-pixel black border, then use the "Drop Shadow (full image)" action that comes as a default with PS 4.0. Again, these are assigned to function keys, and two keystrokes complete the work.
And that's it. All my kung-fu secrets exposed. I know it's all text and no pictures, I'm too lazy to work on that right now. Maybe later. In the meantime (in the spirit of Sulhan), if you're not sure about anything I'm talking about, feel free to ask and I'll be glad to explain.