Trim a Video Without Losing Quality (No Re-encoding)

Most video editing is lossy. You trim a clip, the software decodes every frame, applies the cut, re-encodes the whole thing, and the result is slightly worse than what you started with. For a one-time trim this does not matter much. For a workflow where you trim, then compress, then trim again, the quality drop compounds fast.

There is a way around this. Trimming a video can be almost entirely lossless. The frames in the middle of the video are copied byte-for-byte, and only the frames at the very start and end are re-encoded (a few frames at most). The result is a smaller file with identical quality to the source, produced in seconds instead of minutes.

Why re-encoding hurts

When a video is encoded, it uses a complex technique called inter-frame compression. Instead of storing every frame as a complete image, it stores one "keyframe" (a full image) and then describes the next several frames as "small changes from the previous frame." This is why videos are so much smaller than an equivalent sequence of images.

Re-encoding breaks this. To re-encode, the software has to decode every frame back to a full image, apply the edit, and recompress. Each re-compression loses a little quality, and if the encoder uses different settings than the original, the loss can be visible: soft edges, blocky shadows, colour banding.

Re-encoding also takes time. A 5-minute video on a modern laptop might take 30 seconds to re-encode, but a 30-minute video can take 5 minutes. Multiply by "I am just trimming" and it becomes clear why lossless trimming matters.

How lossless trimming works

A lossless trim takes advantage of the way MP4 files are structured. Every MP4 file has a set of keyframes (called I-frames) placed at regular intervals, usually every 1 to 10 seconds. Between two keyframes is a group of intermediate frames (P-frames and B-frames).

When you trim from a point exactly on a keyframe to another keyframe, the software can literally copy the byte range from one keyframe to the other, without decoding anything. The result is a new video with the identical frames and identical compression.

Modern lossless trimmers find the nearest keyframe before your chosen start point and the nearest keyframe after your chosen end point, copy that whole range, and optionally re-encode only the first and last few frames to hit your exact timestamps. The result is identical quality in 90 percent of the video.

How to trim a video in your browser

The simplest method, no software to install, works on Windows, Mac, Linux, Chromebook, iPhone and Android.

1. Open the Video Trimmer tool.
2. Drop in your MP4, MOV or MKV video.
3. Drag the start and end handles to set your trim. Most tools snap to keyframes automatically for maximum quality.
4. Preview the trimmed section to verify.
5. Download the trimmed video.

Everything runs in your browser using FFmpeg compiled to WebAssembly. No upload, no sign-up, no watermark. A 5-minute video trims in a few seconds.

What a browser trimmer cannot do

• Cut out a chunk from the middle and join the two halves. That requires re-encoding, because the two halves have to be re-stitched, which changes compression.
• Apply effects, transitions or filters. Those require decoding every frame.
• Change the resolution or frame rate. Both require full re-encoding.
• Trim with frame-perfect precision without re-encoding the whole video. You trade exactness for speed and quality.

For those tasks you need a real video editor. But for the most common use case ("cut the first 10 seconds off" or "cut the last minute"), a browser trimmer is dramatically faster and produces a better file.

The command line method (for power users)

If you are comfortable with a terminal, FFmpeg is the tool every lossless trimmer is built on. Here is the exact command:

# Lossless trim from 00:00:30 to 00:02:15
# -ss before -i = fast seek (jumps to keyframe before the start point)
# -c copy = copy streams without re-encoding
ffmpeg -ss 00:00:30 -to 00:02:15 -i input.mp4 -c copy output.mp4

The -c copy flag is the whole trick. It tells FFmpeg to copy the video and audio streams directly instead of re-encoding them. The result is byte-identical video in the middle of the clip.

Two more useful variants:

# Frame-accurate but with re-encoding only for the start/end
# Slower than -c copy, but exact timestamps
ffmpeg -i input.mp4 -ss 00:00:30 -to 00:02:15 -c:v libx264 -crf 18 -preset veryfast -c:a copy output.mp4

# If -c copy gives you a broken start (audio plays but video doesn't)
# Re-encode just the first 2 seconds:
ffmpeg -ss 00:00:30 -to 00:02:15 -i input.mp4 -c copy -avoid_negative_ts make_zero output.mp4

The keyframe-snap tradeoff

With -ss placed before -i, FFmpeg seeks to the nearest keyframe *before* your requested timestamp. This means the output will start slightly earlier than you asked, usually by 1 to 10 seconds depending on the video's keyframe interval.

If that is a problem, you have two options:

• Accept the offset and manually remove the extra seconds on the other end.
• Re-encode just the first GOP (group of pictures). Use -ss before -i for the seek, then encode the whole thing with a fast preset. The result is slightly larger and slightly slower but the timestamps are exact.

When you have to re-encode (and why it is sometimes fine)

There are cases where re-encoding is unavoidable:

When you do have to re-encode, use the highest quality setting you can afford. A CRF of 18 in H.264 produces visually lossless output. The file is 2 to 3 times larger than a typical CRF 23 encode, but the quality is indistinguishable from the original.

Four common mistakes

• Re-encoding when you only need a trim. This is the biggest one. If you are just cutting the start or end, never re-encode.
• Using -ss after -i with -c copy. This forces FFmpeg to decode every frame up to the start point, then copy. It is slower and can produce a corrupted start.
• Trimming a WebM file expecting lossless. VP9 does not support the same lossless range copy as H.264/H.265. Expect some quality loss.
• Not checking the output. Always play the first and last 5 seconds of the trimmed file. Lossless trims occasionally produce a broken start if the source has variable frame rate or a corrupt keyframe.

The bottom line

If your goal is to trim a video (cut the start, cut the end, keep the middle) a browser-based trimmer will do it in seconds at identical quality. There is no faster or better way on any device.

For real editing (cuts, joins, effects) you need a proper editor and you accept the re-encoding cost. That is fine. It is the only way to do it. But for the common case of "I just want to shorten this clip," the lossless path is dramatically better.