MJPEG to PNG Converter

How to Convert MJPEG to PNG Online

1. Upload Your MJPEG File: Drag and drop or click "+ Add Files" to load one or more MJPEG (Motion JPEG) clips. Batch processing is supported, so you can queue several recordings together.
2. Pick Frame Selection: Under "Frame Selection," choose Specific Frame and enter a timestamp in seconds for a single still, or Multiple Screenshots to extract a sequence at a fixed interval (every 0.1 s up to every 10 s, or 1–50 fps).
3. Tune Compression and Resolution (Optional): In "Image Compression," pick a Quality Preset (Lowest through Very High), or set a Specific file size in KB/MB. Adjust Compression level (1–10, default 6) and Compression speed (1–10, default 4) to trade encode time against output size — PNG remains lossless either way. Under Resolution, keep original, set a Resolution Percentage (1–100%), choose a Preset Resolution (144p–4320p), or type exact Width × Height.
4. Convert and Download: Click Convert. Files are uploaded over an encrypted connection, processed on our servers, and deleted automatically after a few hours — no sign-up, no watermark, never shared.

Why Convert MJPEG to PNG?

Motion JPEG is an intraframe-only codec: every frame is encoded independently as a baseline JPEG, with no inter-frame prediction (per the Wikipedia M-JPEG entry). That makes random-access frame extraction trivial — any frame can be decoded without reading neighbors — but it also means each frame already carries irreversible JPEG quantization. Re-encoding to JPEG for stills compounds that loss; PNG (an ISO/IEC 15948:2004 lossless raster format using DEFLATE) preserves the decoded pixels exactly, so the output is the best possible representation of what the camera or capture card recorded.

• CCTV and IP-camera frame grabs — Axis, Hikvision, and Dahua cameras frequently record M-JPEG streams (often inside AVI containers). When pulling a still for an incident report, PNG avoids generation loss from the second JPEG round-trip.
• Forensic and evidentiary work — researchers detect duplicate-frame tampering and camera-model traces from per-frame JPEG quantization tables. Exporting frames as lossless PNG keeps those forensic artifacts intact for downstream analysis.
• Machine vision and quality inspection — industrial cameras stream M-JPEG over USB Video Class or HTTP because random frame access is cheap. PNGs feed directly into OpenCV, scikit-image, or ML pipelines without lossy-compression bias.
• Scientific and microscopy capture — labs that record at fixed framerates often archive a subset of frames at full fidelity. PNG's 8-bit and 16-bit depth options (supported in our advanced panel) cover both standard and high-dynamic-range outputs.
• Non-linear video editing thumbnails — M-JPEG is widely used in NLE workflows because every frame is a keyframe. Exporting reference stills as PNG keeps EXIF-free, alpha-capable assets for documentation or storyboarding.
• Compositing and overlay assets — PNG's alpha channel lets you mask or overlay extracted frames in Photoshop, GIMP, or DaVinci Resolve without re-encoding.

MJPEG vs PNG — Format Comparison

Frame-Extraction Settings Quick Guide

Frequently Asked Questions

Why does extracting frames as PNG produce larger files than the source MJPEG?

MJPEG is a lossy, per-frame JPEG stream — its small size comes from quantizing high-frequency detail. PNG stores the decoded pixels losslessly with DEFLATE, which is far less aggressive on photographic content. A 1920×1080 frame that occupies ~150–300 KB inside an MJPEG stream typically becomes a 1.5–4 MB PNG. The PNG is bigger because it is no longer throwing information away — that is precisely the point for forensic, archival, or re-edit use cases.

Should I extract as PNG or as JPEG from an MJPEG source?

Extract as PNG when you plan to crop, enhance, or composite the frame, or when the still is going into evidentiary or scientific records — re-saving as JPEG would add a second generation of DCT loss on top of the original encoding. Extract as JPEG (try our MJPEG to JPG tool) when storage cost matters more than fidelity, since each frame is already JPEG and re-encoding at high quality is essentially free in size terms.

What devices and software produce MJPEG files?

M-JPEG is widely used by IP-based security cameras (Axis, Hikvision, Dahua), USB Video Class webcams, machine-vision cameras, older Nikon/Canon DSLRs in movie mode, and some drone capture cards. Microsoft also documents an AVI fourCC for storing M-JPEG, so many.avi files from CCTV systems are M-JPEG inside an AVI wrapper rather than true .mjpeg files. Frame-by-frame editors such as Avid and earlier Final Cut also used M-JPEG as an intermediate codec because every frame is independently decodable.

Can I extract every frame from a long surveillance recording?

Yes. Choose Multiple Screenshots and either set seconds-per-frame (1, 0.5, 0.1) or use the framerate dropdown (1, 2, 5, 8, 10, 12, 15, 20, 24, 25, 30, 50 fps) to match the source. A one-hour clip recorded at 1 fps yields 3,600 PNGs; at 15 fps it yields 54,000. Extraction is sequential, so plan the capture rate around the disk space and review time you have — for routine review, 1 fps is usually enough to spot motion events.

What does the Compression level slider do, given that PNG is always lossless?

PNG compression is "tunable but not lossy." The slider controls how aggressively DEFLATE searches for matches: 1 is fastest with the largest files, 10 is slowest with the smallest files, and 6 is the libpng default that balances both. Pixel data is bit-identical at every level. Use a low setting if you are dumping thousands of frames and want speed; use 9 or 10 if you are archiving a small set and want minimum bytes on disk.

Will the extracted PNGs include a transparency channel?

By default, PNGs from an opaque MJPEG source are written as RGB (24-bit) without an alpha channel, since the source has no transparency to preserve. PNG itself supports an alpha channel and you can layer transparency in a downstream editor like GIMP or Photoshop — for example, masking out a privacy region before sharing a frame.

Can I trim or reframe the MJPEG before extracting stills?

Yes. Use Trim MJPEG to clip the segment containing the event you care about, then run that shorter file through this MJPEG-to-PNG extractor — the multiple-screenshots interval will then sample only your trimmed window. This is faster than extracting every frame and discarding the irrelevant ones, and it keeps your PNG output set small enough to review by hand.

My CCTV clip is an .avi file — will this tool still work?

If the AVI contains an M-JPEG video track, rename the extension to .mjpeg or first convert via AVI to MJPEG. If it contains H.264 or another inter-frame codec, you would not actually have an M-JPEG source — extract frames using the matching converter for that codec instead (for example our MP4 to MJPEG intermediate, then run this page).

What resolution will the PNG frames be?

By default, "Keep original" preserves the native resolution of the MJPEG track — typically 640×480 for legacy webcams, 1280×720 for mid-tier IP cameras, or 1920×1080 for HD CCTV. You can downscale by Resolution Percentage, pick a preset (4320p down to 144p), or type explicit width/height. Upscaling is allowed in the UI but cannot recover detail the source did not capture — the JPEG quantization is the ceiling on real resolution.