Transformations

Transcode between formats

You can specify the output audio and video formats you want to produce when building Transformer. For example, the following code shows how to configure Transformer to output H.264/AVC video and AAC audio:

If the input media format already matches the configurations for audio or video, Transformer automatically switches to _transmuxing_, that is, copying the compressed samples from the input container to the output container without modification. This avoids the computational cost and potential quality loss of decoding and re-encoding in the same format. ## Remove audio or video {: #remove-audio-video} Remove audio or video using `EditedMediaItem.Builder`, for example: ## Trim a clip {: #trim } You can remove any media outside specified start and end timestamps by setting the clipping configuration on the input media item. For example, to produce a clip containing only the media between 10 seconds and 20 seconds: ## MP4 edit lists {: #mp4-edit-lists } For faster trimming, Transformer supports MP4 edit lists, allowing for more efficient "trim-only" edits without full video re-transcoding. This method utilizes existing encoded samples and a "pre-roll" within the edit list, which instructs the player to begin playback at a specific point, effectively skipping the undesired initial segment. To make trim-only edits significantly faster, call `experimentalSetMp4EditListTrimEnabled(true)`. It is important to note that not all media players support a "pre-roll" position. This means that when such a player is used, the file will start playback from the absolute beginning of the encoded sample, regardless of any edit list information that might specify a different starting point. Caution: Trimming video often implies permanent removal of unwanted sections. However, using edit lists creates a privacy risk: users might unknowingly share sensitive, "deleted" information "hidden" by the "pre-roll" position. ## Optimizing trims {: #trim-optimization } To reduce the latency of trimming the beginning of a video, enable trim optimization. This speeds up the export by decoding and re-encoding as little of the video as possible, then stitching the re-encoded data with the rest of the original video. The optimization relies on being able to stitch part of the input file with newly-encoded output, which means that the encoder's output format and the input format must be compatible. So, for example, if the file was originally produced on a device with a different encoder implementation then it's likely that it won't be possible to apply the optimization. For the optimization to succeed, the encoder provided to Transformer via the `EncoderFactory` must have a level and profile compatible with the input format. This optimization only works with single-asset MP4 input with no effects except no op video effects and rotations divisible by 90 degrees. If the optimization fails, Transformer automatically falls back to normal export, and reports the outcome of the optimization in `ExportResult.OptimizationResult`. We are validating this functionality and expect it to become non-experimental in a later release. ## Video edits {: #video-edits } `EditedMediaItems` have lists of audio processors and video effects to apply in order. The library includes video effect implementations for common use cases, or you can write custom effects and pass them in when building edited media items. You can rescale media, which can be useful to save on processing resources or bandwidth when dealing with very high resolution input, such as 4k or 8k video. For example, to scale proportionally to 480 pixels high: Alternatively, you can scale by a given factor, for example, to halve the size: You can configure rotation in the same way: ### Custom video effects {: #custom-video } The `Effects` constructor accepts a list of audio and video effects to apply. Internally, Transformer's effects framework converts the list of video effects into a sequence of GL shader programs that are applied in order. In some cases, the effects framework is able to apply multiple effects with one shader program. For example, one shader program can apply multiple consecutive matrix transformations, which improves efficiency and quality. Video effects are also supported for preview in ExoPlayer, using `ExoPlayer.setVideoEffects`. For an example of how to use this API, check out the [effect demo app][effect-demo-app]. The [demo app](demo-application) includes examples of custom video effects. ## Image input {: #image-input } Transformer supports image inputs by treating them as static video clips. To configure an image as an input source, follow these steps: * Create a `MediaItem` using `MediaItem.Builder`. Specify the display duration of the image in the output video by calling `setImageDurationMs`. * Construct an `EditedMediaItem` wrapping the `MediaItem`. Specify the target frame rate for the generated video stream using `EditedMediaItem.Builder#setFrameRate`. The following example demonstrates how to configure an image input to generate a 5-second video at 30 frames per second:

Audio edits

Audio effects are implemented by applying a sequence of AudioProcessor instances to raw (PCM) audio. ExoPlayer supports passing audio processors to the DefaultAudioSink.Builder, which allows previewing audio edits.