Back to guides

Choose a photo for your office-concert lead

Use light, framing and a single clear subject to make your photo a useful appearance reference.

Tiny Desk AI
Subject; Light; Framing
Evaluate the subject, light and framing before choosing the photo.

The short answer

Choose an image in which the intended subject is easy to identify without zooming. Even light and visible facial or species features matter more than a studio background. A clear input helps you judge the result; it does not guarantee identity or motion accuracy.

In this guide

  1. Keep the intended subject unambiguous
  2. Prefer readable light over dramatic lighting
  3. Leave room for the head and upper body
  4. Match the subject mode to the photograph
  5. Keep the original for a useful comparison

What to check

StepCheck
Keep the intended subject unambiguousOne clear lead is easier to inspect than competing faces.
Prefer readable light over dramatic lightingKeep the eyes and important contours visible.
Leave room for the head and upper bodyAvoid cutting through essential identifying features.
Match the subject mode to the photographPeople and pets require different anatomy.
Keep the original for a useful comparisonCompare features throughout the clip.

Keep the intended subject unambiguous

Use a photograph where the intended person or pet occupies a useful part of the frame. Several equally prominent faces, a collage or a mirror that introduces another subject can make the identity request ambiguous.

A photograph of several people is not the same as several independent identity inputs. Follow the current mode’s slot labels and keep supporting band members separate from the identity you are supplying.

Prefer readable light over dramatic lighting

A naturally lit phone photo can be useful when you can see the face without deep shadow or blown highlights. Strong coloured light and heavy beauty filters can hide the features you later want to recognise.

Do not try to match the office set in your input photograph. The source image is for appearance; the effect supplies the performance setting. A simple background often makes the intended subject easier to read.

An everyday portrait in a naturally lit cafe.
A source photo supplies appearance.

Leave room for the head and upper body

For people, keep the whole head and some shoulders visible when possible. For pets, retain ears, muzzle, markings and enough body to understand proportions. Choose a second photo rather than inventing missing features by enlarging a tiny thumbnail.

A close portrait does not establish unseen hands, clothing or instruments. Expect generated areas outside the photo to require particular care during playback review.

A clear face or animal profile is an appearance reference; a crowd is an ambiguous identity request.
A clear face or animal profile is an appearance reference; a crowd is an ambiguous identity request.

Match the subject mode to the photograph

If the current studio offers subject modes, choose the one matching the uploaded image. An animal’s muzzle, paws and body shape should not be judged with the same motion expectations as a human keyboard player.

Pet results need their own review for species, markings, limb continuity and interaction with props. A successful person video does not establish that the corresponding pet version will work.

Keep the original for a useful comparison

Save the selected photograph with the work record. During playback, compare recognisable features at the opening, after camera changes and near the end. A flattering opening frame is only one point in that check.

If identity drifts, record when it happens and what changes. Changing input or settings creates a new request with its own quote; do not repeatedly submit an unresolved task.

Questions about this task

  • No. An ordinary authorised photo can be useful if the subject is clear and the important features remain visible.

Related guides