Previously I posted about the visual development of an image for a small project titled Return to Sender. That post can be found here. Essentially I walked through the various steps I used to make the image (planning, model-making, photography and compositing) and ended with a note that I was somewhat satisfied with the image but considered it not entirely finished. I have since gone back and revised the image, so this post is meant to document those updates.
The image of concern was meant to represent a "top-level" of reality within the short story Return to Sender and the different layers of reality in the story were to be visually distinguished from each other by the style of rendering used for each one. The "top-level" became critical to solve, since it would represent the ceiling beyond which the rendering complexity of lower levels should not encroach. Having felt that I had completed the image, I moved on to doing the next level down, which I determined would be entirely a digital painting.
Once I had completed the next level I revisited the top-level image to see how the images compared next to each other. I was satisfied with the level of visual differentiation between the two but the distance I gained from the top-level image enabled me to recognize a number of glaring issues that I had previously been somewhat blind to.
For reference, here is the image which up to that point I had considered "mostly" final, albeit in need of some tweaks.
Doing a mini-crit1, I saw the following main problems:
- The image is all dark: A dark background combined with dark foreground figures killed any sense of depth, making it difficult to find focus and even read the image. In showing a large environment to give a sense of space one wants to have a relatively light, low-contrast background with dark, high-contrast foreground elements. Oddly enough this is something that I knew and had planned out roughly to varying degrees in my early tone and colour studies but which I didn't hold to as I yielded to the photographic material that I was able to obtain. I would need to look into ways of amending this to regain any sense of depth to the image.
- The mechanical tendrils in the background are bad: They are largely lost due to being light over a dark background which makes their position in space uncertain. They are also overly "scruffy" and seem to twist and bend without an appropriate sense of tension. They are the wrong colour, being shifted magenta while the far background is more red.
- The mid-ground tube is bad: It is overly stretched and too blurry and does not feel consistent with the tubes in the foreground.
- The image feels airy but not immersed in fluid: Key elements associated with images that are under fluids are not present (low contrast, light shafts, bits of floating debris).
In the previous post I noted that I used Affinity Photo to composite the image and do digital painting and I complained of the sluggishness experienced throughout the process. Since then I started doing digital painting using Clip Studio Paint - I had used Affinity Photo for its photo editing rather than painting capability - and the difference was palpable. Going back to the image in Affinity Photo I was reminded of just how impossibly slow making even the slightest change was, an impediment that would make getting the image up to scratch a Herculean task2. I exported out the elements I needed from Affinity Photo to re-compose into a new image using Clip Studio.
I actually started adjusting the image trying to use exclusively Affinity Photo, tweaking the background and foreground elements while using a black and white filter to get a more readable image. However, once I had the image in Clip Studio and began making further changes to the background it became apparent to me that the entire composition was very flawed. I made numerous quick sketches to arrive at a tube layout that would help the image read more easily.
Once I had the rough shapes in place, I traced them out very roughly to create a tone study. This was basically a more disciplined approach to what I had done earlier for the image before I had even begun model-making.
Going from black and white to colour proved to be a little challenging. I was able to apply a preset gradient map to get a not-terrible very rough starting point for an underwater scene, but applying the same results for an entirely red scene proved more challenging. It took me a little while to figure out that the value ranges for the reds needed to be much more clipped than would occur in an all-blue or black and white scene; pure red lights tend to go from dark-values to a high saturation mid-value rather than a low saturation light-value. Because I was trying to keep the background low saturation I wasn't quick to realize that this was the direction I needed to head down.
After I finished the painting I set it aside to complete the rest of the images for the project. When I returned to it I found myself looking at it from different angles and realized that it read better when flipped horizontally, so I adjusted it accordingly.
Conclusion
The lesson I took away from this experience was the value of using the right tools. Using Affinity Photo I had simply grit my teeth and endured, but this severely impeded my ability to iterate and get to the outcomes I had envisioned, leading to lots of compromise. With Clip Studio Paint all of the sluggishness was gone and I found myself much more focused on and attentive to the image itself rather than getting sucked into small details, waiting for the program to respond and then being beset by crashes. The difficulty of using Affinity Photo had caused me to suppress issues with the image in my mind so as to avoid having to try to deal with them.
Comparing the revised image to what I had previously I feel it comes much closer to communicating what I want out of the image. There is much more a sense of depth and motion to the image and the glowing red gives the impression of an internal and organic space rather than some potentially cosmic expanse. The image doesn't do a great job of communicating vastness of scale and would probably benefit from more work being done to bring the detail level of the tubes more in line with the attendant caretakers. However, I'm willing to call it complete for now.
Return to Sender is a visual short story, a kind of prologue or introduction, and can be viewed up on my website.
Footnotes
1 I have found self-critiques done quickly with some distance from the image to be very valuable. Keeping them brief helps ensure that they are focused and that the feedback is actionable. I always write them out; if what is wrong cannot be articulated then the ability to adequately address it becomes very difficult.↩
2 I use a Microsoft Surface Pro 5 running Windows 10. Affinity Photo is generally not the snappiest program on my machine, but its performance slows to a crawl when using adjustment layers. My image, being a composite of many elements, used quite a lot of adjustment layers, which then made trying to further tweak the image excruciating. With just about everything I would want to do in Affinity Photo being through adjustment or live filter layers, this handicap makes it almost useless to me.↩
