Processing workflow

juggler

juggler

Suspended / Banned
Messages
5,059
Name
Simon
Edit My Images
No
@Livin The Dream asked in another thread about my workflow. This post has turned out rather epic so rather than derail his thread about consistency I've started this one. My output isn't especially consistent and what I do varies considerably depending on the audience for the result.

I use a light meter calibrated to my workflow - different camera profiles and the gamma curves applied when converting to different colour spaces can have an effect so it can't do any harm to calibrate, see http://www.frankdoorhof.com/site/2011/08/calibrating-the-light-meter-some-quick-notes/ and http://www.frankdoorhof.com/site/2012/06/light-meter-calibration-revisited/

My monitor is calibrated for colour but not brightness and contrast using a ColorMunki Smile. This is something I'm looking to improve. I find getting contrast levels right exceedingly difficult.

I then do some of the following. It's important to note that I very rarely do all of this and quite often I'll do virtually none. My processing varies from 2 minutes to 2 hours per image, perhaps more on a couple of occasions. I'm not saying it's right but it's a fairly complete list of what I do for the various various forms of portraiture I tend to shoot.
  1. Get exposure right using light meter. If I'm shooting tethered and can be sure I'm not blowing highlights I'll use ETTR and may over-expose considerably.
  2. Take a shot of a Datacolor SpyderChekr 24.
  3. Set WB using the 40% grey patch
  4. use Datacolor software to create an Lightroom develop preset and apply that to the whole set.
If the lighting changes then I'll go back to step 1 and repeat. This sounds a lot of work but I actually find it quicker than doing stuff by eye.
I use one of two different camera calibrations on import into Lightroom, either Camera Standard or Camera Portrait depending on what I'm doing. With my old Olympus I use a neutral profile I bought from Huelight instead - I may do the same for my new D750.

Disappointingly, the Datacolor software doesn't create it's own camera calibration - it works by tweaking the hue and saturation sliders.

Then
  1. I sometimes tweak the overall exposure, black and white points - especially if I've ETTR'd
  2. Noise removal (a) - either none or a little bit in Lightroom
  3. Raw pre-sharpening (a) - either none or a little bit in Lightroom
  4. Lens profile correction - straighten up and remove vignetting if required
  5. Crude crop if needed, leaving a fair bit of extra space. I don't want to process bits of the image I'm definitely going to crop.
  6. I may use the various tools in LR to locally tweak exposure, highlights or shadows.
  7. Open in photoshop as Adobe RGB.
  8. Noise removal (b) Nik Dfine - if I haven't done (a). Maybe in conjunction with a mask or control points
  9. Raw pre-sharpening (b) Nik Raw-presharpener - if I haven't done (b). Maybe in conjunction with a mask or control points
You could use smart objects for this stuff if you think you'll want to go back and alter the settings. I tend not to.
  1. Duplicate layer for basic background cleanup, extending the canvas, etc. Cleanup can be a mixture of spot healing, patching, content aware fill, cloning or just painting. On a plain studio background I may select the whole background, blur it and add some noise.
  2. More background cleanup - I might locally tweak the exposure, contrast or hue of some areas using a mask or Viveza
  3. Liquify, if I'm going to. I sometimes liquify the hairline downwards if it looks like someone has got a huge forehead. I'll tackle flashy bulges around straps, very occasionally reduce chins and sometimes tackle thighs - not to make them smaller, just to give them a cleaner line if they're a bit bulgy.
  4. Major spot healing, but only for those skin blemishes I know I'll have trouble dealing with any other way. I'll quite often use a camera raw filter on a duplicate layer. It's better than the PS spot healing tool at zapping spots automatically and undetectably.
  5. Hair - tidy up stray hairs. A mixture of cloning, straight painting and content-aware fill. I might slightly fill any bare patches exposed by partings.
  6. Skin smoothing - minor skin blemishes, lines, wrinkles. For beauty portraits I tend to dodge & burn on a small scale with a pressure sensitive brush on a grey soft light layer to remove pimples & deeper creases - this removes the highlights and shadows created by them but doesn't fundamentally alter the skin. This is time consuming but gives much better results than healing.
  7. Colour correction - sometimes dodging and burning can alter colours. One way to correct them is with a selective colour or hue/saturation adjustment layer clipped to the dodge & burn layer.
  8. If there are any spots remaining then I might just use the healing brush tool with an oval brush, or I might use frequency separation.
  9. Blotchiness. If someone has blotchy skin, bruises and the like - especially on legs and arms - then I'll do some frequency separation and paint on the colour layer or may use a blur on the colour layer. I sometimes use Nik Color Efex's dynamic skin softener but only on the low frequency layer.
At this point the image will just be a tidier version of the original - it shouldn't look majorly different. From here on in it's much easier to stuff up an image by over-doing things..
  1. Teeth. If I'm going to whiten teeth at all I'll do it here with masked hue / saturation & curves layers
  2. Lips. If I'm tidying up lipstick then I'll do it now by cloning using brushes in either darker colour or lighter colour mode.
  3. Eyebags. I use targeted curves adjustments rather than the patch tool or dodging and burning:
    View: https://www.youtube.com/watch?v=3lL8XUUWg_Q
    I sometimes use this technique to tackle blotchiness and bruises too.
  4. Eyes and eyebrows. I might add catchlights using a blurred spatter brush, then either dodge & burn eyelashes, eyebrows and the details of the iris using a softlight grey layer. Or more simply I might use Viveza with a control point to boost the structure.
  5. Large scale contrast work. Sometimes I'll tweak curves, perhaps targeted via luminosity masks. Sometimes I'll use large scale dodging and burning using a soft brush on a grey softlight layer to emphasise musculature or to do some 'beauty' contouring to bring out what makeup is there. Occasionally I'll use Topaz Clarity or one of a number of other techniques to emphasise local contrast. It's really, really easy to overdo this step.
Then we come on to the tricky bit - toning, giving the colours in the image a recognisable character. In the past I've been too influenced by what others do and have tried to slavishly copy. Now I try to enhance what is already there. That might mean lifting the black point to soften an image. Or boosting overall contrast. For toning I try to unify and simplify the colour palette so the predominant colours are all complementary or contrasting. I do that in a number of ways.. targeted curves (as above), 3d LUT adjustment layers, selective colour adjustments targeting the neutrals, Nik's Analogue Efex or Color Efex or by modifying a Florabella preset.

Then I try to wait 24 hours and have a look at it again the next day, ideally on a few different screens including phones and iPad.

Lastly we come to the output step.
  1. Return to Lightroom. Perform final crop.
  2. Soft proofing to check for out-of-gamut colours - I haven't yet felt the need to do any correction
  3. Resize a copy in OnOne Perfect Resize (especially if upscaling)
  4. Open in photoshop
  5. Perform output sharpening for print using Nik's Output sharpener. Ususally masked using control points and backed off from the defaults
  6. Add borders for printing
  7. Upload to either Ilford, The Print Space or Metro Imaging for printing.
  8. Wait nervously...
  9. Frame using acid-free pure white matte board in a Nielsen Classic frame (it's got better glass than the others in the Nielsen range or any of Ikea's stuff).
  10. Hang / deliver
 
Last edited:
juggler said:
I use one of two different camera calibrations on import into Lightroom, either Camera Standard or Camera Portrait depending on what I'm doing. With my old Olympus I use a neutral profile I bought from Huelight instead - I may do the same for my new D750.

Disappointingly, the Datacolor software doesn't create it's own camera calibration - it works by tweaking the hue and saturation sliders.
Click to expand...

You could use Adobe's own DNG Profile Editor to generate a custom calibration profile using the shot taken of the target which might be better than using presets.
 
Many thanks for that Simon, there's much to chew on there for me. I shall probably keep dipping back into that and ask a couple of questions that spring up.

I haven't looked into the data colour software but assume it doesn't create camera profiles as above, just to set black, white and gamma points, presumably?

The first bit that interests me there is the selective targeted adjustments using the neutrals, care to expand on that please? 3D LUT adjustments are also new on me so I shall look that up.

I note you use more of Nik software than I do, only tend to use the silver Efex occasionally. But you do you define quite a bit which I haven't explored that much.

How easy to you find transferring your contrast and colour adjustments over to other shots for that consistent output in PS? Once I'm in PS I find that a little inconsistent myself.
 
Livin The Dream said:
Many thanks for that Simon, there's much to chew on there for me. I shall probably keep dipping back into that and ask a couple of questions that spring up.

I haven't looked into the data colour software but assume it doesn't create camera profiles as above, just to set black, white and gamma points, presumably?

The first bit that interests me there is the selective targeted adjustments using the neutrals, care to expand on that please? 3D LUT adjustments are also new on me so I shall look that up.

I note you use more of Nik software than I do, only tend to use the silver Efex occasionally. But you do you define quite a bit which I haven't explored that much.

How easy to you find transferring your contrast and colour adjustments over to other shots for that consistent output in PS? Once I'm in PS I find that a little inconsistent myself.
Click to expand...

The datacolor software creates a develop preset which alters the hue, luminosity and saturation of each of the colours in the HSL panel to ensure colour accuracy. From memory, it doesn't alter the black / white / contrast section at all. You're meant to get that right yourself using their instructions before sending the shot of the card to the Datacolor software. Altering contrast alters perceptual colour saturation but their software needs consistent input to do the colour correction.

Using selective colour for toning.. see https://fstoppers.com/post-production/selective-color-possibly-best-tool-photographers-7954
The advantage of targeting neutrals is that it tones the image but leaves the blacks black and whites white.

I hardly use Silver Efex any more, it's very good at making images look as though they've been through Silver Efex. When I do use it I'll tend to blend the results with a more vanilla black and white conversion.

3d LUTs - It's now relatively easy to create your own and it's a very quick way getting consistent, reliable toning between images. I spent ages a while ago searching for free ones to download and have a heap but I can't find the source any more. I've only recently really started exploring what you can do with them.

I often use blend_if to target toning, contrast or other adjustments. It's quicker and easier to use than luminosity masks and seems to me to give better results unless you're going to put a lot of effort in.

Transferring colour adjustments between shots is easy - create a 3d LUT. Or duplicate your toning group from one image to another. You could even use Image | Match colours which sometimes works really well - but often fails. I rarely do local modifications to toning except to mask off skin.

I find getting contrast to match between images much harder, especially if I've done any local adjustments. Hence my desire to improve my monitor environment and calibration. I'm not sure how my wife will react to calibrated lighting in our living space though :)
 
juggler said:
The datacolor software creates a develop preset which alters the hue, luminosity and saturation of each of the colours in the HSL panel to ensure colour accuracy. From memory, it doesn't alter the black / white / contrast section at all. You're meant to get that right yourself using their instructions before sending the shot of the card to the Datacolor software. Altering contrast alters perceptual colour saturation but their software needs consistent input to do the colour correction.

Using selective colour for toning.. see https://fstoppers.com/post-production/selective-color-possibly-best-tool-photographers-7954
The advantage of targeting neutrals is that it tones the image but leaves the blacks black and whites white.

I hardly use Silver Efex any more, it's very good at making images look as though they've been through Silver Efex. When I do use it I'll tend to blend the results with a more vanilla black and white conversion.

3d LUTs - It's now relatively easy to create your own and it's a very quick way getting consistent, reliable toning between images. I spent ages a while ago searching for free ones to download and have a heap but I can't find the source any more. I've only recently really started exploring what you can do with them.

I often use blend_if to target toning, contrast or other adjustments. It's quicker and easier to use than luminosity masks and seems to me to give better results unless you're going to put a lot of effort in.

Transferring colour adjustments between shots is easy - create a 3d LUT. Or duplicate your toning group from one image to another. You could even use Image | Match colours which sometimes works really well - but often fails. I rarely do local modifications to toning except to mask off skin.

I find getting contrast to match between images much harder, especially if I've done any local adjustments. Hence my desire to improve my monitor environment and calibration. I'm not sure how my wife will react to calibrated lighting in our living space though :)
Click to expand...

Thanks Simon, looong day at work today, so I'll digest this next few days.

I have come across 3D LUT briefly as I use an x-rite Display Pro and have seen LUT files. Haven't really looked into calibration in this much detail, just calibrate to RGB which is quite straight forward.

Interesting what you say about the match colour command in PS, I too have found that it doesn't do a great job so will look into what you say.

I don't use blend if, but keep meaning to do so, so another worthwhile prompt.
 
What software do you use to create the LUT files Simon?

I tend to use selective color for tweaking skin tones but I think it's the LUT element that is missing in my process.

This is all a little more relevant to the high-end work that you produce I guess, but I would like to take my PP to the next level. This is a good starting point though so thanks.
 
Livin The Dream said:
What software do you use to create the LUT files Simon?

I tend to use selective color for tweaking skin tones but I think it's the LUT element that is missing in my process.

This is all a little more relevant to the high-end work that you produce I guess, but I would like to take my PP to the next level. This is a good starting point though so thanks.
Click to expand...

First, you need to do your own toning manually via curves, selective colour, colour balance layers or whatever - then export those changes as a LUT for using on another image:
http://blogs.adobe.com/richardcurti...shop-for-video-grading-and-image-enhancement/
http://blogs.adobe.com/jkost/2014/09/5555-exporting-color-lookup-tables-in-photoshop-cc.html
There is also http://3dlutcreator.com/
 
You must log in or register to reply here.
Back
Top