Initializing... drag & drop files here
Supports: MPG, MPEG
.mpg or .mpeg video — both extensions are accepted. Batch is supported, so you can queue several clips and extract frames from each.MPEG-1 (ISO/IEC 11172) and MPEG-2 (ISO/IEC 13818) are the codecs that powered Video CDs, DVD-Video, and the first wave of digital broadcast TV — DVB, ATSC, and ISDB all adopted MPEG-2 as the over-the-air standard. Even today, plenty of archived .mpg and .mpeg files survive on home-burned discs, camcorder backups, and old broadcast captures. Pulling individual frames out as JPEG is the natural way to harvest stills from that footage when you don't need the moving picture.
| Property | MPEG (.mpg/.mpeg) |
JPEG (.jpg/.jpeg) |
|---|---|---|
| Type | Video container + codec | Still image |
| Standard | ISO/IEC 11172 (MPEG-1, 1993), ISO/IEC 13818 (MPEG-2, 1995) | ISO/IEC 10918 (1992) |
| Compression | Inter-frame (I/P/B) DCT, motion-compensated | Intra-frame DCT, lossy |
| Typical bitrate / size | ~1.5 Mbit/s (MPEG-1), up to 9.8 Mbit/s (DVD MPEG-2) | ~50–500 KB per frame |
| Color | YUV 4:2:0 | YCbCr 4:2:0 or 4:4:4 |
| Audio | MPEG audio (Layer I/II/III) muxed in | None |
| Primary uses | DVD-Video, VCD, DVB / ATSC broadcast | Web photos, thumbnails, prints |
| Browser support | Native playback uneven; often needs transcode to MP4 | Universal — every browser since 1996 |
| Method | What it does | Best for |
|---|---|---|
| Quality Preset | Highest → Lowest fixed presets | Quick one-off extraction with no tuning |
| Image Quality % | Manual 1–100% JPEG quality | Fine control; 85–95% is the sweet spot |
| Target file size (%) | Aims for a percentage of an internal reference | Predictable, roughly uniform output |
| Specific file size | Exact KB/MB target with auto-scale | Hard upload caps (email, forum, CMS) |
A JPEG at 85–95% quality is visually near-indistinguishable from the source frame; below ~70% you'll start to see blocking around high-contrast edges and ringing around text. For DVD-resolution material (720×480 or 720×576), 90% quality typically lands in the 50–150 KB range per frame; for a 1080p MPEG-2 broadcast capture, expect roughly 100–300 KB at the same quality.
Specific Frame extracts a single image at the exact timestamp you enter in the Seconds field — useful when you already know the moment you want. Multiple Screenshots samples the clip at a regular interval (every N seconds or every N frames) and gives you a sequence of JPEGs in one pass — useful for contact sheets, picking the best frame from a candidate set, or rotoscoping reference. You'll choose one or the other in the Frame Selection group; you can't mix both in a single conversion run.
.jpeg or .jpg?Both extensions produce byte-identical files. JPEG is the original three-letter+ name from the ISO/IEC 10918 standard; .jpg was the DOS-era three-letter shortened variant and is now more common on the web. Pick .jpg for maximum compatibility with older software and shorter filenames, .jpeg if your platform or pipeline standardizes on that suffix.
It depends on the source. DVD MPEG-2 was interlaced 720×480 (NTSC) or 720×576 (PAL); each "frame" is actually two interleaved fields, so a raw grab can show comb artifacts on moving subjects. Our converter deinterlaces by default. The output JPEG can't be sharper than the source — MPEG's inter-frame compression already discarded high-frequency detail — so a 480i grab will always look like 480i, regardless of which quality preset you pick.
MPEG-1 and MPEG-2 use I-frames (full pictures), P-frames (predicted from a previous frame), and B-frames (predicted from frames on either side). If you ask for a timestamp that lands on a P or B frame in a heavily-compressed clip, the reconstructed image inherits that prediction noise. Try nudging the timestamp by a fraction of a second to land on a different position in the GOP, or raise the source bitrate before extracting if you control the encode.
Yes, when the file is genuinely MPEG-1 or MPEG-2. Many DVR exports use a .mpg extension; some cable boxes wrap the stream as a .ts (MPEG-TS) instead — those have a different extension but the same underlying codec. If your file ends in .mpg/.mpeg, upload it directly. If it's .ts or .m2ts, you'll usually want to rewrap to .mpg first (or use a video-to-image route built for transport streams).
Choose Multiple Screenshots and set the interval to 1 second. The tool walks the clip and saves a JPEG at each one-second mark. For longer reels you can also drop the rate — 1 frame every 5 or 10 seconds gives a manageable contact sheet for a feature-length recording without producing thousands of files.
If you just want a faithful frame grab, leave Image Resolution on Keep original — for DVD MPEG-2 that's 720×480 or 720×576, for HD broadcast captures it's typically 1280×720 or 1920×1080. Pick a preset (1080p, 720p, 480p) or a percentage if you specifically need to downscale for web or upload. Upscaling a low-resolution MPEG to 4K won't recover detail; it just enlarges the existing pixels.
JPEG is a still-image format, so audio is dropped — there's nowhere to store it. If you need both the audio and the frame, run a separate audio extraction with MPEG to MP3 or MPEG to WAV in parallel.
Yes. Files upload over HTTPS, are processed on isolated workers, and clear from the server after your session ends. Nothing is shared, indexed, or kept past the session window. Related conversions you might want next: MP4 to JPEG, MPG to JPEG, or Compress JPEG to shrink the extracted stills before upload.