[Experiment] Lip Color & Texture (Gloss/Matte) Test | 14 Conditions, 42 Images

Conclusions First

• Lip color specification confirmed color difference from control in 4 out of 5 colors (red, coral, dark berry, orange). Dark berry showed the most pronounced change; red and orange are also clearly distinguishable. Nude was nearly identical to control.
• Texture specification confirmed difference from control in 3 out of 4 types (glossy, matte, shimmer). Glossy and shimmer showed increased gloss; matte showed reduced gloss. Velvet showed minimal difference from control.
• Both lip color and texture are controllable via prompt, and higher-saturation color specifications tend to produce clearer effects.
• Combined color × texture specifications also work. In 3 conditions — glossy red, matte dark berry, and shimmer coral — both color and texture were reflected simultaneously. Glossy nude reflected only the gloss; color remained nearly identical to control.

Experiment Design

Base Prompt

extreme close up of lips, lip detail used to focus on lip changes. The variable section is swapped one at a time for comparison.

Evaluation Criteria

• Color change: Does lip color change according to the prompt?
• Texture change: Does gloss/matte quality change according to the prompt?
• Side effects: Are there unintended changes to elements other than the lips (skin tone, composition, etc.)?

Part A: Lip Color

Testing how well lip color specifications are reflected. Comparing control (no specification) against 5 color specifications.

A00: Control (No Color Specification)

seed 1	seed 2	seed 3

Observation: Natural pink-beige lips with light gloss and no signs of lipstick application. Serves as the comparison baseline for all subsequent conditions.

A01: red lipstick

seed 1	seed 2	seed 3

Observation: Compared to control, the lips shift overall to red/rose tones. Redness is clearly increased compared to control’s pink-beige, confirming the color specification effect. Composition differs from control with a diagonal angle (composition differences are not attributable to the prompt, as seeds differ).

A02: coral lipstick

seed 1	seed 2	seed 3

Observation: Compared to control, lips shift to an orange-tinged pink (coral tone). Warmer than control’s pink-beige, and a distinct hue from A01 (red). Color specification effect confirmed.

A03: nude lip color

seed 1	seed 2	seed 3

Observation: Nearly identical beige-pink to control. Since control already approximates a nude color, the specification effect cannot be visually distinguished. No color change was confirmed.

A04: dark berry lipstick

seed 1	seed 2	seed 3

Observation: The most pronounced color change of all conditions. The lips shift entirely to a dark, purple-tinged berry color — clearly distinct from control’s pink-beige. Lip line boundaries are sharp, giving the look of applied lipstick.

A05: orange lipstick

seed 1	seed 2	seed 3

Observation: Clearly shifts to orange tones compared to control. Color differs from A02 (coral) — more saturated orange than coral. Color specification effect confirmed.

Part A Summary

Color difference from control was confirmed in 4 out of 5 color specifications (red, coral, dark berry, orange). Dark berry shifts to dark purple-tones, red to rose tones, orange and coral each to warm tones — changes align with the specified color direction. Nude showed no difference from control, but this is likely because control already approximates a nude color, making distinction difficult. This model’s X lipstick lip color control works, with clearer effects for colors farther from control.

Part B: Texture

Testing how well lip texture specifications are reflected. Independently from Part A, only the texture is changed from control.

B00: Control (No Texture Specification)

Same condition as Part A’s A00.

seed 1	seed 2	seed 3

Observation: Same as A00. Natural pink-beige with light gloss. Serves as the baseline for texture comparison.

B01: glossy lips

seed 1	seed 2	seed 3

Observation: Compared to control, light reflection on the lower lip increases with a wet-looking gloss. Control already has light gloss, but the glossy specification further emphasizes it. Color remains nearly identical to control’s pink-beige.

B02: matte lipstick

seed 1	seed 2	seed 3

Observation: Compared to control, gloss on the lip surface is suppressed. Less light reflection with a dry surface texture. The wet highlight points from control and B01 (glossy) are barely visible. Color remains nearly identical to control.

B03: velvet lip texture

seed 1	seed 2	seed 3

Observation: Minimal difference from control. Appears to have slight gloss, looking intermediate between B01 (glossy) and B02 (matte), but confirming a clear difference from control is difficult. Color remains nearly identical to control.

B04: lip gloss with shimmer

seed 1	seed 2	seed 3

Observation: The strongest gloss of all texture conditions. Multiple light points scattered on the lower lip, giving a sparkling finish. Even stronger light reflection than B01 (glossy), with more fine highlight points. Color remains nearly identical to control.

Part B Summary

Difference from control was confirmed in 3 out of 4 texture specifications (glossy, matte, shimmer). Glossy shows increased gloss, matte shows reduced gloss, shimmer adds fine sparkle — all changes align with their specifications. Velvet showed minimal difference from control, making effect determination difficult. Gloss strength ranks as shimmer > glossy > control > velvet > matte. Lip texture is also controllable via prompt with this model.

Deep Dive: Color × Texture Combinations

Parts A and B tested color and texture separately. Here we test whether both are reflected when color and texture are specified simultaneously. 4 combinations were selected from the 4 colors confirmed effective in Phase 1 (red, dark berry, coral, nude) and 3 texture types (glossy, matte, shimmer).

D00: glossy red lipstick (gloss + red)

seed 1	seed 2	seed 3

Observation: All 3 show reddish lips with a wet-looking gloss. Color is the same red/rose tones as A01 (red lipstick alone); gloss is equal to or stronger than B01 (glossy lips alone). Both color and texture are reflected.

D01: matte dark berry lipstick (matte + dark berry)

seed 1	seed 2	seed 3

Observation: All 3 show the dark purple-tinged berry color. Color matches the tones of A04 (dark berry lipstick alone). For texture, 1 out of 3 (seed 789) shows almost no light reflection, making the matte quality prominent. The remaining 2 have masked areas making texture judgment difficult, but no strong gloss highlights are visible. Both color and texture are reflected.

D02: shimmer coral lipstick (shimmer + coral)

seed 1	seed 2	seed 3

Observation: All 3 show orange-tinged pink (coral tone) with fine light points. Color matches A02 (coral lipstick alone); texture shows the same sparkle as B04 (lip gloss with shimmer alone). Both color and texture are reflected.

D03: glossy nude lip color (gloss + nude)

seed 1	seed 2	seed 3

Observation: All 3 show beige-pink lips with gloss. Color is nearly identical to control (A00) — no color difference was confirmed, consistent with Part A’s A03 (nude lip color alone). However, gloss is clearly increased compared to control, showing wet-looking reflection equivalent to B01 (glossy lips alone). Only texture was reflected; no color change occurred.

Deep Dive Summary

Of 4 combinations, 3 conditions (glossy red, matte dark berry, shimmer coral) reflected both color and texture. Glossy nude reflected only gloss (texture), with no color difference from control — consistent with the weakness of nude in Part A, which persists even in combination. No mutual interference (one weakening the other) was observed in combination specifications.

Summary

• Lip color control: Color difference from control confirmed in red, coral, dark berry, and orange. Dark berry showed the most clear change — consistently shifting to purple-tinged berry across all seeds. Nude was indistinguishable from control.
• Lip texture control: Difference from control confirmed in glossy, matte, and shimmer. Glossy and shimmer showed increased gloss; matte showed reduced gloss. Velvet showed minimal difference.
• Color × texture combinations: glossy red lipstick, matte dark berry lipstick, and shimmer coral lipstick all reflected both color and texture simultaneously. No mutual interference was observed in combinations. glossy nude lip color reflected only texture.
• Practical recommendations: For changing lip color, specifications with large color difference from control like dark berry lipstick or red lipstick are effective. Texture is controllable with glossy lips or matte lipstick. Simultaneous color and texture specification also works, allowing single-phrase descriptions like glossy red lipstick. Colors close to control (nude) or intermediate textures (velvet) tend to show weak effects.

Lab Director Comment: Dark berry actually looks like it has lipstick on — that’s solid. The difference between glossy and matte is obvious, and the shimmer sparkle is great too. The fact that simultaneous color and texture specification doesn’t cause mutual interference is quietly impressive. matte dark berry lipstick as a single phrase hitting the target is very practical.

Related Articles

⚠ 関連記事が見つかりません: /en/tips/blush-contour-highlight-test

⚠ 関連記事が見つかりません: /en/tips/eye-color-shape-control-test