Two defined phi comparisons
Compare visible face length/cheek width and mouth width/nose width with φ ≈ 1.618.
Compare two facial proportions with the golden ratio, φ ≈ 1.618. See the actual spans and differences side by side.
Drag and drop or choose a file
Compare two named proportions with phi, with the endpoints shown.
See the report in actionLoading the measured example…
Phi is 1.618… . Face length/width and mouth/nose width use different endpoints, so read each comparison separately.
Compare named photo spans with phi and see the signed difference. The result describes those spans, not a universal beauty percentage.
Check whether you are reading visible face length/width or mouth/nose width.
Follow the ruler endpoints on the original portrait.
A smaller deviation means closer to phi for that ratio; it is not a grade for the whole face.
The free result opens three source readings and two derived comparisons with phi. Each comparison keeps the numerator, denominator and signed difference visible, so the calculation can be checked.
Compare visible face length/cheek width and mouth width/nose width with φ ≈ 1.618.
See each ratio, its difference from phi and its signed percentage deviation from that reference.
Open the original length, cheek, mouth and nose spans, plus the head-roll reading, for free.
These two comparisons use different selected relationships from the same portrait. Their definitions matter more than whether a result happens to be close to the constant.
The first comparison uses the upper visible forehead landmark to the chin, divided by the cheek-to-cheek span. It uses the same aspect reading shown elsewhere in your report. Because the model does not locate the hairline, this should not be compared with a hairline-based face-length formula as though they were identical.
The second comparison divides mouth-corner distance by the span between the detected nose-wing points. The stored report expresses nose width/mouth width, so this view explicitly takes its reciprocal. A smile can change the mouth span and therefore the result. Open the underlying reading to verify the endpoints.
They can be traced to existing measured spans without inventing a hairline or estimating hidden anatomy. Their inclusion is an exploratory product choice, not evidence that either ratio ought to equal phi. For a wider set of descriptive relationships, use the proportions table.
Explore eight measured facial proportions ↗Choose the question you want to explore. The definitions, examples and research are here when you need them.
A constant does not decide which spans should be compared. The meaning changes when the numerator, denominator or landmark definition changes.
| Quantity | Definition in this tool | How to read it |
|---|---|---|
| Phi, φ | (1 + √5) / 2, approximately 1.618034. | The mathematical comparison constant. |
| Visible face aspect | Upper forehead landmark to chin, divided by cheek width. | Not a hairline-based face length. |
| Mouth / nose width | Mouth-corner width divided by visible nose-wing width. | The reciprocal of the stored nose/mouth reading. |
| Signed relative difference | 100 × (ratio / φ − 1). | Below zero means smaller than φ; not less attractive. |
The golden ratio is a mathematical constant. Applying it to a portrait first requires a clear choice of numerator and denominator, which this calculator makes explicit.
The golden ratio is φ = (1 + √5) / 2, approximately 1.618034. If a longer span A and a shorter span B form that ratio, A/B is about 1.618. It is a relationship between chosen lengths, not an intrinsic property that a whole face either passes or fails.
Changing endpoints changes the calculation. Forehead-to-chin length divided by cheek width is different from eye-to-mouth distance divided by total length. A golden ratio face test needs to name its spans before its result can be checked. This tool makes two specific comparisons and does not average them into beauty.
Facial-proportion research does not make every selected span an ideal of 1.618. Pallett and colleagues studied feature arrangements under controlled image conditions; their paper is not a validation of this calculator or a universal phi-based grading rule. We keep the arithmetic separate from that stronger claim.
Sources: Pallett, Link & Lee (2010) · New ‘Golden’ Ratios for Facial Beauty
Research using “golden” in its title does not automatically test the constant 1.618 for every feature. The chosen stimuli, landmarks and rating task matter.
Pallett, Link and Lee manipulated eye-to-mouth and interocular distances in female face images. Preferred arrangements in their experiments were around 36% vertically and 46% horizontally, rather than a universal rule that every ratio should equal phi.
The study used particular face-height and width definitions and a restricted set of stimuli and observers. FaceRater’s forehead/cheek and mouth/nose comparisons are different quantities. The paper therefore does not validate our endpoints or a beauty prediction.
The useful question is “How does this named proportion compare with this constant?” Our two comparisons answer it directly. The report leaves aesthetic preference to you instead of disguising the calculation as scientific consensus.
Sources: Pallett, Link & Lee (2010) · New ‘Golden’ Ratios for Facial Beauty
Choose one clear portrait, check the spans and inspect both comparisons with phi. There is no account requirement or paid step to see these calculations.
Keep both cheek edges, the upper forehead and chin visible. Use a relaxed mouth and even light. Confirm adult photo permission, then choose JPG, PNG or WebP up to 10 MB.
The result names both the numerator and denominator for each comparison. Read the ratio beside phi rather than treating a different measurement or an inverse ratio as interchangeable.
Select a comparison to inspect the underlying photo spans. A signed deviation describes arithmetic distance from phi. It does not indicate how beautiful, healthy or typical your face is.
This is a calculator for named proportions. It is useful when you want to inspect the geometry behind a familiar claim, rather than apply a decorative mask to a face.
Identify the two spans before comparing the number. If the video measures hairline to chin and this tool uses a forehead mesh point, the results are not interchangeable even when both are called “face length”.
Read the sign and units. A negative relative difference simply means the selected ratio is smaller than the constant. It is neither a negative score nor evidence that the feature needs changing.
Use the facial ratios view when you want eye spacing, jaw taper or visible vertical bands. Those quantities answer different geometry questions and are not combined into a made-up phi compliance score.
The displayed difference retains its sign and units. It states how a measured ratio compares with phi; it does not estimate a percentage of beauty.
For a measured ratio r, the signed difference is r − φ. A negative value means below phi; a positive value means above phi. For example, a ratio of 1.50 is about 0.118 below phi. The sign describes the calculation, not whether a feature is good or bad.
The displayed percentage is 100 × (r/φ − 1). A ratio of 1.50 gives about −7.3%. This is a relative difference from a selected constant. It is not a confidence interval, percentile, beauty percentage or estimate that your face matches an ideal person.
These comparisons are derived from the report’s existing rounded ratios and shown with appropriate display precision. They add no terms to geometry-v1.1. The overall score retains its declared six dimensions and reference bands; the proportion band is not silently replaced with 1.618.
See how the separate face rating is calculated ↗Visible spans change with photographic conditions. Check the endpoints and use comparable captures before reading a small numerical difference as meaningful.
A ratio removes changes from uniform image resizing, but not changes in perspective. Keep the camera at eye level and avoid an extreme close selfie. If you compare two photographs, use similar distance, framing and expression instead of interpreting a small deviation change as a structural improvement.
A hidden cheek contour or an open smile can change these particular spans. If a point falls off the visible feature, retake the photo. More decimal places cannot fix an incorrect endpoint. The overlay is a way to audit detection, not a guarantee of anatomical measurement accuracy.
If you want a styling outline, use face shape. If you want left/right alignment, use symmetry. Phi comparisons answer a narrower arithmetic question and should not substitute for either result. The same portrait can support several distinct views without giving those numbers the same meaning.
Find your visible face-shape estimate ↗Sources: Bryan, Perona & Adolphs (2012) · Perspective distortion and social judgments
These answers cover the golden ratio face questions that affect how you use and interpret the result. Open a question for its measurement definition or a practical next step.
Yes. Both phi comparisons, their signed differences, the underlying span readings and head-roll context are free. No customer account or card is needed for the scan.
This calculator does not issue one. A ratio equal to phi has zero arithmetic deviation for that selected pair of spans. It does not establish a perfect face or predict everyone’s attractiveness judgment.
No. This tool does not calculate a golden-ratio match percentage. Its signed percentage is deviation from phi, with the formula shown. That quantity should not be read as beauty or a population ranking.
No. Face length starts at an upper visible forehead landmark. A test using the actual hairline would need a separately verified endpoint and should be labeled differently.
Compare the endpoints, ratio direction, reference constant and photograph first. Some apps use different spans or transform several differences into an undisclosed score. Those outputs are not directly interchangeable.
No. It is already open in your free result. The optional $17.49 full report adds the remaining geometry readings, regional explanations, styling plan and PDF.
Written and reviewed by FaceRater · . Research provides context; it does not validate our scoring formula. Product definitions and limitations are explained in the measurement method.
Experimental feature-arrangement research; not evidence that FaceRater’s selected spans should equal phi or that deviation predicts attractiveness.
Controlled research on camera distance and judgments of photographed faces; does not validate FaceRater scores.
The phi calculations remain free. The complete report explains the wider set of facial measurements and connects them with a preference-based styling plan. Its value is context and organization, without a hidden golden-ratio beauty score.
Compare two named proportions with phi, with the endpoints shown.
Explore the sample report