Modificateur de vitesse vidéo

The chipmunk effect at 2× is expected behavior and stems from the mechanism the tool uses to change speed. The pipeline sets the HTML video element's playbackRate property to your chosen speed multiplier. When playbackRate is 2, the video element advances through both the video and audio tracks at twice the normal rate. The audio decoder outputs samples twice as fast, which effectively raises the pitch by one octave — the same octave shift a tape recorder produces when you run the tape faster. This is consistent with how native video players, YouTube's speed controls, and VLC behave without pitch correction enabled. The chipmunk effect becomes progressively more pronounced at 3× and 4×. At 0.5× the voice deepens noticeably. At 0.25× speech becomes a slow, bass-heavy rumble. True pitch-corrected speed change — where duration changes but pitch stays constant — requires a time-stretching algorithm such as WSOLA, phase vocoder, or the Web Audio API's AudioWorklet with a time-stretch processor. These algorithms are computationally intensive and not yet available in the MediaRecorder pipeline without a custom WebAssembly implementation. The Mute Audio option in the tool strips audio from the output entirely, which is useful for speed-changed video that will receive a separate voiceover or background music track. Practical tip: for tutorial and lecture recordings where you want faster playback without pitch change, use YouTube's built-in speed control or VLC's Synchronization panel, both of which include pitch correction by default.

Slow motion at 0.25× using the browser's playbackRate mechanism produces smooth output only if the source video was shot at a high enough frame rate to begin with. When you slow a 30 fps video to 0.25×, the output plays at an effective 7.5 frames per second. At 7.5 fps, each frame is displayed for approximately 133 milliseconds, which is long enough for the human visual system to detect frame-to-frame jumps in high-motion scenes. The result looks jittery on fast-moving subjects — sports, hand gestures, water, vehicle movement. True smooth slow motion requires either a high-frame-rate source or an optical flow interpolation step that synthesizes intermediate frames between captured ones. A source recorded at 120 fps slowed to 0.25× plays at 30 fps, which is perfectly smooth. Most modern smartphones record at 120 or 240 fps in dedicated slow-motion modes specifically for this reason. For a source shot at 30 fps, 0.5× gives 15 fps output, which still looks slightly choppy but is often acceptable for talking-head footage and controlled movement. For 0.25× from a 30 fps source with smooth results, you would need to run the video through an interpolation tool like DAIN or RIFE before slowing it. These tools are available as local applications and some web services. Practical tip: check the frame rate of your source in the tool's metadata display before selecting a slow speed — if it shows 60 fps or higher, 0.25× will produce acceptable smoothness without interpolation.

The WikiPlus tools are independent single-purpose utilities, so chaining effects requires running each tool in sequence and downloading the intermediate result between steps. A typical trim-then-speed workflow is: open the Video Trimmer, set your in and out points, download the trimmed clip, then upload that clip to the Video Speed Changer and set your desired rate, download the speed-changed result, then optionally upload it to the Video Compressor to reduce file size. Each step re-encodes the video through MediaRecorder, which introduces a generation of quality loss per pass. For most content — social media clips, tutorial excerpts, marketing loops — three passes through MediaRecorder at the tool's default bitrate produce output that is visually indistinguishable from the source. For professional production work, multiple re-encodes compound visible artifacting on dark gradients and high-frequency textures. In those cases, the correct approach is to do all operations in a single FFmpeg command that encodes only once: ffmpeg -i input.mp4 -ss 00:00:10 -to 00:01:30 -filter:v setpts=0.5*PTS -filter:a atempo=2.0 -b:v 2M output.mp4 trims, doubles speed, and compresses in one encode pass. The WikiPlus browser tools are optimized for quick, no-install use where near-professional quality is sufficient. Practical tip: plan your edit sequence before starting — do trim first since it produces the shortest source for subsequent passes, which makes speed change and compression faster.

The audio in the output file reflects the pitch-shifted version of the source audio at the chosen playback rate. The pipeline captures both the video stream and the audio stream from the video element's captureStream() and feeds both into MediaRecorder simultaneously. Because the video element's playbackRate setting applies uniformly to both the video decoder and the audio decoder, the audio in the captured stream is the rate-shifted, pitch-shifted version. There is no option to keep the audio at original pitch while changing video speed within the current pipeline. The output codec for audio depends on the container: WebM outputs Opus audio encoded at approximately 64 Kbps, and MP4 outputs AAC at a similar rate. Both codecs introduce mild lossy compression on top of the pitch shift. The Mute Audio toggle removes audio entirely from the capture stream before it reaches MediaRecorder. Use this option when the speed-changed audio would be unusable — for example, 4× speed where voices become incomprehensible — and you plan to add a separate audio track in a video editor afterward. Some use cases benefit from the pitch shift intentionally: timelapse-style montages at 3× or 4× with no audio, or comedy clips at 2× where the voice effect is part of the joke. Practical tip: for instructional videos where you need slightly faster pacing without pitch change, export at 1.25× from this tool and then apply pitch correction in Audacity or iMovie on the downloaded file rather than correcting in the browser.