Choose a Photo for Hit the Road Dude

Start with one unambiguous subject

Choose the person or animal that should replace the main character. A group photo asks the tool to resolve an extra question before it can interpret appearance. Cropping to one subject is useful when the remaining face and upper body are still clear enough to use.

Avoid a collage containing several views, outfits or pets. The current interface accepts one image, and a collage is still one image with competing information inside it. It does not function as several separately labeled references.

For a person, use an image they have agreed to have edited. If the photo belongs to someone else, get permission before selecting it. For a pet, choose a normal photograph that makes its features readable; it does not need a human face or a costume.

The chosen subject should remain recognizable at the size of the upload preview. If you have to zoom far in to decide which face matters, a tighter crop or another source image will be more useful.

Leave room for hair, ears and clothing

A face-only crop can leave out the very details you want to replace. Include the full hairstyle, ears where visible, neck and the upper part of the outfit. The reference performance shows more than a face, so your photo should supply information about the rest of the visible character.

Keep a little breathing room above the head. Cutting through the hairline or an animal's ear makes that boundary ambiguous. You do not need a large empty border, but the important silhouette should be complete.

Look at the lower edge of the crop as well. A collar, jacket opening or bandana can help establish the subject's appearance. If the crop ends immediately under the chin, the tool has little visual evidence about what should appear below it.

A full-body image can work if it is sharp and the face remains readable, but distance often hides useful details. A head-and-upper-body view is a practical starting point for this particular scene. You do not need to imitate the truck pose in your source photo.

Check the eyes and edges, not just the file size

A large file can still contain a blurred subject. Zoom into the eyes, hairline and collar. If those edges are smeared by movement or missed focus, exporting the image at a higher resolution will not recreate the missing detail.

Even light makes the photo easier to assess. One side of the face can be brighter than the other, but avoid a picture where an eye disappears into a deep shadow. Strong backlight may preserve a silhouette while losing the features needed to recognize the person.

For pale clothing or white fur, check that bright areas retain texture. For dark hair or a dark coat, make sure the outline is not lost against a matching background. The issue is visible separation, not a requirement for a particular background color.

Heavy beauty filters and aggressive smoothing can change a face's identifying details. If you want the output to resemble the person as they normally appear, use a photo that already does. Simple exposure adjustments are different from reshaping the face or inventing texture.

Remove avoidable ambiguity

A hand across the mouth, a phone covering half the face or an object in front of the chest hides useful information. Choose another frame when possible. The replacement must already interpret movement and partial occlusion in the reference, so a clear starting image gives you one less uncertainty.

Sunglasses can be part of a person's usual appearance, but they conceal the eyes. Decide whether keeping that accessory or showing the eyes matters more for the version you want. Do not assume the tool will reconstruct hidden features exactly.

A readable front or three-quarter view is generally easier to compare with the final character than an extreme angle. A photo taken directly from below can exaggerate the jaw or muzzle. A sharply turned profile can hide the far side of the face.

For animals, check ears and paws separately. A good facial photo may still crop off a defining ear or hide the chest. The pet guide goes further into species-specific inspection; this page focuses on making the input clear before submission.

Use a supported file without unnecessary conversion

The input accepts JPG, PNG and WebP files. Large originals are resized and compressed locally before they are uploaded. Both image dimensions must be at least 256 pixels. These are acceptance limits, not a promise that every image meeting them will produce an equally good result.

If your phone saves a HEIC file, export a real JPG copy using its photo or sharing tools. Changing the filename extension does not convert the image. Likewise, a screenshot of a photo can be easier to upload but may lose detail or include interface elements; use the original image when practical.

The browser prepares the selected image before upload. This reduces transfer size while keeping a local preview. You do not need to compress an otherwise suitable photo repeatedly just to make the original file tiny.

If an image has transparent areas, review how the subject appears against a plain background. Transparency is not a separate instruction to remove the background in the generated video. The fixed reference still supplies the scene.

• The file is actually encoded as JPG, PNG or WebP.
• The browser can open the image; large files are prepared automatically.
• Width and height are each at least 256 pixels.
• The preview opens upright and shows the intended subject.
• No app controls, screenshot borders or unrelated text cover the photo.

Use the local preview as your last input check

Selecting a photo for Hit the Road Dude does not immediately upload it. The preview stays on your device until you start generation. Use that pause to check the crop and choose a better image if something important is missing.

If the preview shows the wrong photo, remove it and select the intended one. If the image looks sideways, export an upright copy and open it again. Check the visible result instead of relying only on how the file looks in another application.

When a file cannot be opened, try a fresh export in a supported format. When it is too small, return to the original photo rather than enlarging a low-resolution thumbnail. When it is too large, export a reasonable-size copy while keeping the subject's detail.

Do not send sensitive documents or unrelated personal information as part of the image. The tool needs the subject's appearance. A visible address, screen notification or document in the background adds no useful information for the replacement.

If two photos look equally good

Compare the details that matter for your specific edit. One photo may have a sharper face but hide the outfit; another may show the jacket clearly while the eyes are blurred. Prefer the image that makes the face readable and preserves enough of the visible appearance to identify the subject.

Choose one variable to improve if you make another attempt. For example, replace a tightly cropped picture with a wider upper-body view while keeping the same subject and outfit. That makes the change easier to evaluate than switching to a different pose, hairstyle and costume all at once.

A well-chosen photo for Hit the Road Dude improves the information available to the tool, but it does not guarantee exact likeness or flawless movement. Watch the completed video through every gesture, then download a local copy whenever you want. The saved video stays private in your account until you delete it.

Once the photo is ready, follow the production guide for sign-in, the 10-credit submission and My videos. The reference is already fixed; the useful decision at this stage is which clear version of your subject should appear in it.