Project 52 - Post Processing - Week 1

[Lon Howard]

I have resisted a Project 52 so far because my picture taking is extremely intermittent but I'm thinking this concept might work for me since I can use older images, and hopefully others can learn as much as I hope I can. This project is for post-processing devotees since the emphasis is directed there, but I should add that C&C on any aspect of the image is valuable and desired. Each week I will post three versions of the same image.

1. A saved version of the unprocessed raw file, along with a link to the actual raw image.
2. The file after ACR processing only.
3. A 'final' version after processing in CS5 or CS6.

A normal critique is desired but if possible, compare the three images and offer C&C on the degree and nature of the processing done in ACR and PS; and whether or not it could/should be different. Re-dos are encouraged and even desired; seeing an image being taken in another direction is of immense value, as it broadens the sphere of thinking and sparks the imagination even if it's something you yourself wouldn't do with a particular image. Please advise which version you worked on.

For re-dos using the link to the actual raw image, Lightroom users will use that as opposed to ACR. The link works for ACR 6.6/LR 3.6 and later, according to the host, Mediafire.

Questions about what was done in either ACR or PS are also encouraged - I will try to make sure I keep notes on the ACR work. Also don't be shy about being forthrightful; a "what were you thinking?" epiphany is just as valuable if it comes from without as opposed to within. If I'm over-cooking or under-cooking, I'd like to know as soon as possible, and I think most would feel the same way.

Well, lets get on and see how this goes. Here is the first image. I'll call it Week 1 even though it's time for Week 3 ... and I'll catch up very soon.

EDIT: I've just noticed that all three of these images look somewhat fuzzy ... first time that's happened to me and I have no answer for it now. Please ignore that and consider all other aspects.


Unprocessed raw: Link to actual raw image: http://www.mediafire.com/?h4mlbbcxv9dpo3m




After ACR processing:




Final after CS6 processing:

[Geoff F]

Looks fine to me, Lon, a perfect example of correct processing.

If you wanted to spend quite a bit more time, I suppose you could have created two Raw conversions, one to add a little more highlight brightness to the buildings and another with a darker sky, then merged them with a mask.

Which would prove to be a good example of working with Smart Objects.

Quite a lot of work for little benefit though.

Any straightening of the buildings would have to be left until after the merging.

[Brownbear]

Beautiful!

[Lon Howard]

Looks fine to me, Lon, a perfect example of correct processing.

If you wanted to spend quite a bit more time, I suppose you could have created two Raw conversions, one to add a little more highlight brightness to the buildings and another with a darker sky, then merged them with a mask.

Thanks Geoff; that's an interesting thought.

I did brighten up the buildings but did so in PS after the raw conversion (perhaps I should have gone a little further with it?). I thought briefly about darkening the sky as well but thought that the combination of doing both might look a tad artificial.

I've never used smart objects ... guess it's time I looked into that .

[Geoff F]

Using Smart Objects as Raw conversions is something which I regularly use, Lon. Very useful for a HDR type of image when you can only take one shot.

Not true HDR but two different Raw conversions from the same image can be useful.

Create the first conversion but hold the Shift key whilst transferring to the main edit window. This produces a Smart Object.

Right click the image thumbnail layer and choose New Smart Object from Copy. Double click the thumbnail picture of the new layer.

This returns you to the ACR Editor where you can edit this layer as required. Click OK when finished as you go to the main edit window but now have two copies but with different conversions.

Merge the two layers as required with masks, selections, etc.

A little while ago we did discuss using two different images from their Raw files and combining them as separate layers in the same window. Double clicking the thumbnail of either image returns you to ACR for further adjustments.

ps Smart Objects can do other things - but I haven't gone there yet.

[Letrow]

Lon, good to see a PP related thread. As a JPEG shooter I am interested in seeing what you can get out of the RAW files. So far I haven't really felt compelled to change to RAW, but you never know, I have an open mind.
I like the colours you came up with in the last photo, that is how I would process the photo too in Gimp. It is a bit difficult to see the end product next to the first (bleaker) shot, but that is because you notice the difference so much now.

One thing that in my opinion worsened in PP is the building in the right corner (more exactly the second building coming out of the right lower corner). It seems to have more moire now than it had at the start, making it slightly unpleasant to look at.

I like how the sky worked out. Nice detail in there.

[Lon Howard]

Very good eye on the moire, Peter. The culprit was a poster edges layer I added in CS6. Please don't ask what possessed me to do that ... I hadn't used that filter in a long time. This is a perfect example of trying to do too much. Here's a revision, in which I just deleted that layer, also toned down the vibrance. Here I'm trying a high pass filter with soft light blending mode instead of unsharp mask to reduce artifacts, though it still looks somewhat soft.

Thanks very much again!

[Letrow]

The colour looks better now Lon, a bit toned down from the first attempt. Still, I like strong colours, but always have to be careful to keep it normal looking. Especially skin tones can be difficult to get right if you use vibrance.

[FrankMi]

Hi Lon, like you I don’t get out to shoot as much as I like and I too am trying to learn as much as I can about post processing.

Thank you for providing the link to the DNG file. It makes learning from the image so much easier.

When I opened the image in ACR I noticed that Lens Corrections image settings for your Canon 18-55mm lens had been applied. Switching between Lens Correction (on) and Camera Raw Defaults (off) demonstrated the difference the lens correction made. Although I usually make the ACR Lens Correction with my images, I need to try switching back and forth to see how much of a difference it makes on images taken with my lenses!

This brings up an interesting point. Assuming that the ACR correction is valid, unless we shoot RAW and use the ACR Lens Correction for the specific lens in use, the image can contain unexpected distortion. In this case, the shape of the buildings changes fairly dramatically between the original and corrected images.

[Lon Howard]

This brings up an interesting point. Assuming that the ACR correction is valid, unless we shoot RAW and use the ACR Lens Correction for the specific lens in use, the image can contain unexpected distortion. In this case, the shape of the buildings changes fairly dramatically between the original and corrected images.

Thanks for your thoughts, Frank. Also - at least as far as I know - if there are obtrusive vignettes and other sky distortions caused by a polarizer, enabling lens profile corrections in a raw converter is the only way to eliminate them, unless you've got a load of time on your hands and like to blink a lot. You can open a jpeg in raw and enable lens profile correction that way, but I just noticed it doesn't find the correct lens automatically so you have to find it from the drop down list. I wonder if it would be as effective as it is with a raw image though??

[FrankMi]

I wonder if it would be as effective as it is with a raw image though??

I guess we'd need to take a view like this with a lot of lines that should be straight and test it. I really don't know as I hadn't thought to compare before now. I used a grid display on the before and after versions of your image and you can really see a significant difference. There may be effective Lens Correction available for SOOC JPG images but I had never previously thought about needing that kind of PP for JPGs!

I had a try at playing with your image and managed to recover the edges lost in the ACR processing. When I was done, I compared it to your final result and realized that, likely due to personal preferences, my version was a bit more vibrant but not as sharp. I then realized that yours has a higher contrast than my version. Increasing the contrast and brightness a bit made my copy look sharper.

I wasn't sure if you wanted modified images posted to your thread but I can send it to you in a PM if you prefer and if you are interested, the steps I used to do the PP. I suspect that if you give an image to 50 people, you'll get 50 different versions of the results and as many combinations of PP techniques to get there!

[splashy]

You can import it straight in LR 4.3 and with a couple of sliders you're finished, I took away CA also (small but it's there)
I think you sky is a bit overdone puple-ish.

[Lon Howard]

I wasn't sure if you wanted modified images posted to your thread ...

Thanks Frank, and yes as mentioned in the first post, re-dos are highly desired. My thought when starting the thread was not to just improve my skills but to allow all an opportunity to improve their PP skills. It really doesn't matter to me if someone else's version is better than mine ... in the end the more people who learn, the better. Some probably don't have the time for re-dos and that's fine, but when possible I do like to see them, and also to know whether they were done in RAW/LR or PS. It's one thing to go through online portfolios of professionals and try to figure out what and how they did, but I sense that seeing someone else do a little magic with your own image heightens your focus.

Your thoughts about 50 people wanting 50 different things is a good point ... maybe it could get out of control we'll just have to see .

@ Mike: Thanks for the insight, Mike ... I'll look at it some more!

[FrankMi]

re-dos are highly desired.

OK, here is what I did with this marvelous image. As mentioned above, I opened it initially in ACR and reset to defaults. When I did I realized that it was worthwhile to see how much the Lens Correction changed the image so I applied a View/Show/Grid and found that corrections for the Canon 18-55mm lens really do help. I also noticed that when the Lens Correction is applied, we lose some of the image around the edges of the scene.

This part was an experiment - I opened copies of both the Default and Lens Corrected images in CS5, placing the larger Default image above the Lens Corrected one in Layers. I then used Edit/Transform/Warp on the Default layer to match the Lens Corrected image thereby retaining the Lens Corrections for the bulk of the image but recovering the missing pieces around the edge of the image. The result then looked like this:



Next, I used the View/Show/Grid to help me align the vertical and horizontal lines of the buildings and adjusted the buildings using Edit/Transform/Skew. At this point I knew I might lose much of what I recovered on the edges if I simply cropped to include just the parts that were visible after the alignment as you can see in this image:



I noticed that the top left would be easy to simply clone the sky and the top right was doable, but with a little more cloning finesse so that left the bottom to level out. Fortunately, the amount of adjustment needed at the bottom was slight so a little bit of Edit/Transform/Warp did the trick. Now the image is aligned and fills the frame with just a little cropping.

From here on out it is all up to personal taste on how one sees the colours, saturation, contrast, brightness, etc. in their mind so it may not match what you actually saw when you took the shot. In the end, I compared my result to your final posting and realized that yours had a higher level of both brightness and contrast. Realizing that you saw the scene and I did not, I boosted the brightness and contrast by 20 points each using an Adjustment Layer. Here is what I ended up with as my view of your image:



By the way, I love the way the main street leads the eye all the way back to the Space Needle and the glint on the green dome!

Hope this helps!

[Lon Howard]

@ Frank: Frank, I confess my skills with the various Transform functions are far short of yours ... I can use skew, etc. for basic straightening but I couldn't match the default layer with the corrected one without some serious bending so I'm putting that on hold for now; I'll find a Kelby or youtube tutorial later. The primary improvement in using lens correction in ACR appears to be that it eliminates middle of the frame from protruding toward you as you view it, at least in this case. Thanks for your redo, I think I will try to add some vibrance back into the image but as you say, to each his own.

I'm catching up so here is week #2's image. The main problem here is the drab sky ... I tried to add some color there both in ACR and PS but the foliage defeated everything I tried: adjustment brush, selections of all kinds including using threshhold layers, masking; everything ending up looking like mush. Two images with varying exposures was a no-go because less exposure just darkened the gray without adding color. So here at the three versions. Thanks again for the input!


Unprocessed raw: Link to actual raw image: http://www.mediafire.com/?v7qs3n51hgk3qn5




After ACR processing:




Final after CS6 processing:

[Lon Howard]

This is what I was able to do using a blue photo filter and then masking all but the sky. Couldn't go much deeper with the color without the imperfections around the foliage leaping out more than they do here. Anyone who's done this knows it's more work than they care to do anyway. Obviously with no texture or variation in the sky it's not a great image anyway but still, I wish there was a better way to get even this far.

[Lon Howard]

Catching up, so here's week 3. This one is definitely out of my comfort zone ... I really don't do flowers or animals much but my wife would take a picture of every flower she saw so when we went out together, well I was kind of compelled to join in. This is hand held from three years ago at the Antelope Valley Poppy Preserve near Lancaster, CA, with a Canon 400D, EF28-135 f/3.5-5.6 IS USM at f/5.6 and 105 mm. With more suitable equipment I'm sure I could have done better, but here's what resulted.

I didn't enable lens correction in the uncorrected raw image this time.


Unprocessed Raw file. Link to actual raw image: http://www.mediafire.com/?lz3e3kw4jwuz6w8




After Raw Processing:




After Processing in CS5:

[Brownbear]

The images of the roses are gorgeous, great detail, clarity and colour.

I also think that the whitish streaks/highlights distract from an otherwise lovely bokeh and would clone them out or soften them somehow. I don't have enough editing experience to advise on the how tos, but your final image appeals to me the most.

[Lon Howard]

Thanks, Christina. In reviewing, I saw that the brightness of the whitish blades in the background was enhanced by a curves layer. I did a 100% mask over those white areas and it seemed to have helped near the top of the frame ... not so much further down because the curves layer didn't add much - if any - exposure in that area in the first place, hence nothing to mask. But, some improvement and that's what we're after. Thank you again!

[Geoff F]

I think, Lon, I would try to clone out that purple area at the top edge.

Where it is situated, on the top edge and close to the palest area of the flower, I find it distracting.

The other little background flower is fine. In a darker area without any important areas close by. I may darken it just a little more, but I'm not really worried either way.

Otherwise, everything looks good to me.