For a value study from a photo, start with three values—light, middle, and dark. Move to five only when a boundary you need to paint disappears. Use two values to test the composition and nine to inspect subtle transitions, not to plan the first block-in.
That answer came from reducing one controlled still-life source four ways. Five values gave us the most useful working map: enough separation to distinguish the teal pitcher, orange pear, cream linen, navy book, and cast shadows, but not enough room to copy every change in tabletop grain or ceramic texture.
What the test held still
The source in the cover is a fictional editorial simulation, not a photograph of a physical studio setup. We generated one still life containing exactly four objects under one soft side light, then locked the 1536 × 1024 source, crop, object positions, and exposure.
For every variation we:
- converted that source to grayscale;
- divided the full black-to-white range into 2, 3, 5, or 9 evenly spaced levels;
- mapped each pixel to its nearest available level;
- displayed every result at the same size.
We did not add local masks, repair object edges, blur texture, change contrast, or hand-paint a preferred result. Adobe describes Posterize in similar operational terms: choose a number of tonal levels and map pixels to the closest level. Adobe also notes that a grayscale conversion is necessary when you want a specific count of values rather than a multiplied set of RGB-channel colors.
This mechanical reduction is the control, not the finished painting plan. A painter still has to decide which broken pieces belong together.
Two values: test the arrangement, not the form
The two-value panel answers one severe question: does the dark-light arrangement survive when every middle value disappears?
The pitcher, book, window, and main cast-shadow direction still organize the frame. The pear remains a compact dark accent. That means the composition has a usable large-scale pattern.
But look at the damage. The pitcher’s body becomes nearly flat. Linen folds break into black-white flecks. Wood grain becomes louder than the objects. An automatic threshold does not understand that one table is a single supporting plane; it faithfully reports every pixel crossing the cutoff.
Use two values as a gate:
- Can you identify the focal mass at thumbnail size?
- Is the negative space around that mass distinct?
- Does one dark shape connect too much of the frame?
- Does a small but important subject vanish into its background?
Do not use the threshold result as paint-by-number instructions. Redraw the large shape by hand and remove islands that do not affect the composition.
Three values: establish the light map
Three values restore the middle family. Now the wall and tabletop separate, the pear sits against its background, and the linen reads as one light object with a few dark turns. This is the smallest version that explains the lighting setup rather than only the silhouette.
The Metropolitan Museum of Art’s value-scale exercise uses the same practical trio—lightest, medium, and darkest—then asks the viewer to locate those areas in a thumbnail sketch. The Getty’s formal-analysis guide defines value as how light or dark a color appears and connects shading with the illusion of three-dimensional space.
For a three-value map, assign jobs rather than objects:
- Light: the window, lit linen, and selected highlights.
- Middle: most of the wall, tabletop, and lit pear.
- Dark: the book, pitcher shadow family, contact shadows, and deepest accents.
This version is ideal for deciding whether the light family stays connected. It is less reliable for distinguishing two neighboring objects that occupy the same broad family. In our panel, part of the pear and parts of the cast shadow compete for the same dark note.
Five values: make a block-in map
Five values solved the useful separations without returning to photographic accounting. The pitcher gains a lit plane, middle body, shadow side, and dark opening. The pear separates from its cast shadow. The cream linen remains light while its major fold direction becomes paintable. The book stays one dark mass but no longer collapses into the tabletop shadow.
That is the point at which the reduction becomes a working map. It answers four different questions at once:
- Composition: large wall, table, object, and shadow masses.
- Light direction: lit side, turning plane, and shadow side.
- Object separation: pear from wall; book from cast shadow.
- Focal hierarchy: small bright and dark accents without full texture.
Five is not a universal ideal. It won this source because the scene contains four materials, readable side light, and several adjacent middle-dark shapes. A foggy landscape might need only three. A portrait with a subtle far cheek, dark hair, pale skin, and white shirt may need five before the likeness anchors separate.
Nine values: observe first, paint later
The nine-value panel looks smoother and more photographic. It reveals gradual turning on the pitcher, smaller folds in the linen, and distinctions within the wall and table. That makes it useful for checking the finish.
It is a poor first block-in for the same reason. Nine bands let surface variation compete with structure. A painter can spend the opening pass matching a minor wall transition while the pitcher silhouette or main shadow remains wrong.
Use nine values as a reference layer you consult after the three- or five-value structure works. Restore only transitions that explain form, likeness, material, or focal emphasis. Leave the rest merged.
A repeatable photo-to-study workflow
Start by making the image neutral and small. Convert a copy to grayscale, shrink it until detail stops being persuasive, and create a three-value reduction. Ask what disappeared.
If a missing boundary changes the subject, light direction, or object overlap, move to five values. If it only restores pores, fabric weave, leaf chatter, or background texture, stay at three. Keep the original color source beside the map so grayscale does not become permission to forget hue and temperature later.
When you translate the map into a painting brief, name the groups:
preserve one connected light shape across the linen and pitcher rim; keep the book and contact shadows in the dark family; separate the pear from the wall with one middle-light step; omit tabletop grain until the final pass.
Then run a neutral first repaint in the FreePainter photo-to-painting workflow. Compare the result to the value map before asking for more brush texture. If the large families drifted, fix those first.
What this experiment cannot prove
This is one synthetic, side-lit still life with a broad tonal range. It does not prove that five values suit every subject, medium, or painter. Equal digital bands are not the same as perceptual grouping, and a different exposure can move many pixels across a threshold without changing the scene. Transparent watercolor, compressed fog, dark skin under soft light, or a white subject on a white background deserves its own controlled test.
The portable rule is narrower and more useful: begin with three value families, add a fourth or fifth only to recover a necessary boundary, and postpone subtle transitions until the structure already reads. A value study succeeds by removing decisions—not by reproducing the photo with fewer colors.