ADVANCED COMPOSITING & RETOUCHING - Chapter 10 (Capstone), Exercise 1 Why Prove the Recipe on One Photo Before Automating It ==================================================================================== QUESTION: Explain why this capstone builds and proves the entire recipe on one reference photo (Steps 1-4) before recording or scripting anything (Step 5), rather than starting directly with automation. SOLUTION / EXPLANATION: An Action or script only ever replays or executes exactly what it was given - it has no ability to notice on its own whether the underlying sequence of steps actually produces a correct result. If the masking, compositing, filter stack, and frequency-separation steps were recorded or scripted before being proven to actually work correctly, any mistake in that sequence - a mask that's slightly too aggressive, a filter setting that's too strong, a frequency-separation step applied in the wrong order - would get baked directly into the recorded Action or script, and then faithfully reproduced identically across every one of the 60 files. Automation doesn't fix or improve a flawed recipe; it multiplies whatever the recipe actually is, mistakes included. By contrast, working through Steps 1-4 manually on one reference photo first means every individual decision - which channel gives the best mask contrast, how much feathering the composite needs, what the brand's filter stack should actually look like, how the frequency-separation split should be tuned - gets checked and refined by eye, on a result that's easy to inspect and adjust, before any of it becomes a fixed, automated sequence. Only once that sequence has been confirmed to genuinely work correctly does it make sense to record or script it, since at that point automating it means reproducing a proven-good result rather than an unproven one. This is also consistent with Chapter 7's own tip about testing a batch operation on a small subset before running it on hundreds of files - this capstone applies the same underlying principle one level earlier, proving the recipe itself before even beginning to automate it at all. -------------------------------------------------------------------------- WHY THIS WORKS AS AN ANSWER: It explains specifically why automation reproduces whatever the underlying recipe actually is, mistakes included, rather than correcting for them, and connects the "prove it manually first" approach to the same caution about testing before scaling up that Chapter 7 already established for batch runs themselves.