Videos must be in mp4, mkv, or mov format.

Only videos encoded with h.265 (HEVC) and h.264 (AVC) are supported. These are among the most commonly used video codecs today.

Use https://mediainfo.js.org/ to check the codec used in your video. If the CodecID states HEVC or AVC format, then it is supported.

If your video uses a different codec, you can convert it using third-party tools such as the free ffmpeg command line tool, or video editing software like Premiere Pro or DaVinci Resolve, ensuring you encode the video with h.265 or h.264 codecs.

Here is a generic example of converting a video file from a non-supported codec to h.265 using ffmpeg:

ffmpeg -i <videofile> -c:v libx265 -c:a copy video_h265.mp4

Replace `<videofile>` with the name of the video you want to convert (e.g., video.mp4, video.avi, video.webm).

Video processing time varies, unlike images, as it involves larger files and more processing. However, our platform stands out for its swift processing compared to competitors. Many HD and 4K videos convert at near "real-time" speed - in other words, similar processing time as duration of video itself. Rest assured, we'll notify you via email with a download link once your video is ready, so you don't need to stick around whilst we do the work.

The output video will retain the same container as the input video, except for mov files. If your input is mp4, the output will also be mp4. If your input is mkv, the output will be mkv. For MV-HEVC video output, the format will always be mov.

The h.265 codec used for exported videos is becoming mainstream but may not yet be supported by all systems. If you have playback issues due to codec problems, try the following solutions, prioritized from most to least preferred:

- NVIDIA Hardware Decoding (preferred): Best results are expected with a codec provided by your NVIDIA GPU.

-Intel Processors: Recent Intel processors support h.265 and provide codecs. Install and use this codec if available.

- AMD Hardware Decoding: AMD processors generally support h.265. Install and use this codec if available.

- Official Software Codecs: Microsoft offers a codec for Windows systems, which incurs a small cost but is preferred over unofficial codecs.

- Unofficial Software Codecs: Third-party codecs, like those in the K-Lite Codec Pack, might work.  Download K-Lite Codec Pack.

Note: Using third-party codecs can cause compatibility issues and result in suboptimal viewing experiences.

No, support for upscaling and downscaling is not provided. You can use third-party tools to adjust the resolution before uploading your video to Immersity AI.

Audio will be passed through to the output video unchanged whenever possible.

Subtitles will be preserved in the output video whenever possible. If subtitles are embedded in the video itself, they cannot be separated and will be processed along with the video, potentially causing artifacts in the subtitle display.

Similar to image conversion, artifacts may be introduced at object edges and in areas with transparency or reflections. Temporal changes might also occur as each frame is processed independently. While some compensation is applied during processing, not all temporal jumps can be entirely eliminated.

The input video must not exceed 2GB in size.

This format is for viewing stereoscopic videos on the Apple Vision Pro headset.

The 360 format is designed for fully immersive VR viewing. The video will be mapped into a spherical view, ensuring both imagery and depth match at seam points. Selecting 360 field of view ensures this match at seam points and also corrects lens aberrations for fisheye lenses.

The free sample will have the same format and codec as the full video, providing an accurate preview of the final output. The only difference will be the length of the video.

To view the output video, you need a player capable of interpreting the exported format. Examples include:

- Apple Vision Pro: Use with MOV MV-HEVC.

- VR headset players: Such as Skybox, Bigscreen VR, and Virtual Desktop, using SBS formats.

- 3D TVs: Often support top/bottom formats.

- Lume-Pad 2, LeiaPlayer: Use SBS Full format.

Depth maps need to have the highest definition possible while maintaining a reasonable file size. PNG files support 16-bit depth, which is sufficient for good depth definition. JPG files on the other hand only support 8-bit depth, making them unsuitable for this purpose. Additionally, PNG is a lossless format, ensuring no compression artifacts are introduced.

You will receive a video file in the MP4 format, using the MPEG-4 container and encoded with the H.264 codec. The video will have a frame rate of 30 fps and an 8-bit color depth per channel. The aspect ratio and dimensions will match the original image, unless exported with the free tier, which downsizes the video to 720p.

Both formats are Side-by-Side (SBS) formats. Parallel SBS places the right eye image on the right and the left eye image on the left, suitable for most digital viewers. Cross View SBS swaps the images, with the right eye image on the left and the left eye image on the right, allowing for naked-eye viewing by crossing your eyes to merge the images.

LIF stands for Leia Image File format, suitable for viewing on Leia devices. It contains a hidden second eye image within the file. The image can be viewed as 2D on standard viewers and as 3D on Leia devices.

Images must have a width and height between 90px and 8000px on any edge and an aspect ratio between 32:5 and 5:32.

Supported formats include PNG, JPG, JPEG, and WEBP.

The exported image or video will retain the original dimensions, except for Side-by-Side (SBS) images, where the width will be doubled but the height will remain the same.

Depth maps must be in PNG format and should represent depth in terms of disparity, not absolute distance, to avoid distortions. Depth maps created by Immersity AI meet this requirement.

No, the depth map does not need to match the dimensions or aspect ratio of the image. It will be automatically adjusted to fit the dimensions of the original image.

Depth maps created by the latest AI technology perform well with most images. However, images with semi-transparent or reflective surfaces, or those with significant depth jumps, may produce artifacts or be less accurate.

Yes, some cropping may occur when using zoom settings.

Anaglyph exports combine two offset images in different colors (red-cyan). Viewing through red-cyan glasses merges the images to create a 3D effect, offering a cost-effective stereoscopic experience.

It is a format for viewing stereoscopic images on the Apple Vision Pro headset.nd Community Guidelines at any time.

A depth map provides detailed depth information using grayscale, where lighter areas are closer, and darker areas are further away. With 16-bit precision, it offers unparalleled definition for applications like 3D modeling and intelligent photo editing.

Abstract art will produce abstract results. However, since the AI may not interpret the shapes as you would, it can be an interesting experiment to see what it picks up on, that you may not.

Yes, the neural network has been trained to handle all types of images, including analog or computer-generated cartoons.

Some stretching of objects may occur on edges. Adding dilation (extending the object edges) can reduce these artifacts, but excessive dilation may bring background objects to the foreground.

No, currently this support is only available for videos, but it may be added in the future.