Initializing... drag & drop files here
Supports: MKV
12.450 for the frame 12 seconds and 450 ms in) to capture a single still. Switch to Multiple Screenshots to extract a sequence at a chosen capture rate (0.1s, 0.2s, 0.3s, 0.5s, 1s, 2s, 3s, 4s, 5s, 6s, 7s, 8s, 9s, or 10s per frame).MKV (Matroska) is the dominant container for movies, TV shows, anime fansubs, Blu-ray rips, and high-quality screen recordings — often holding an H.265 4K stream alongside multiple audio tracks and subtitle tracks in one file. Pulling PNG stills from that MKV (rather than JPG) gives you a lossless image with every pixel preserved exactly: no DCT ringing around hard edges, no smearing on subtitles, no degradation when the still is re-edited or re-saved. PNG is the right output when image fidelity matters more than file size.
If you'd rather have smaller files for a long sequence, see MKV to JPG. For an animated output, see MKV to GIF.
| Property | MKV (Matroska) | PNG (extracted frame) |
|---|---|---|
| Type | Multi-track video container | Single still image |
| Released | 2002 (Matroska project) | 1996 (PNG 1.0) |
| Typical codecs | H.264, H.265/HEVC, AV1, VP9, MPEG-2 | Lossless DEFLATE |
| Compression | Lossy stream compression | Lossless |
| Audio tracks | Multiple (English, dub, commentary) | None |
| Subtitle tracks | Multiple soft (SRT, ASS, PGS) | None |
| Transparency | No (opaque RGB) | Yes (8-bit alpha) |
| Plays in browsers | Limited — most browsers don't natively play MKV | Universal |
| File size, 1080p frame | ~2-8 GB for 1 hour | ~2-5 MB per frame |
| Best for | Storing full feature-length movies and shows | Tutorials, OCR, archival, print stills |
| Goal | Frame selection | Capture rate / settings |
|---|---|---|
| Single anime screenshot | Specific Frame | Exact timestamp, native resolution, 8-bit |
| Plex / Jellyfin episode poster | Specific Frame | Representative scene, 1080p or source, 8-bit |
| Software tutorial from OBS MKV | Multiple Screenshots | 1s or 2s per frame, native resolution |
| Storyboard contact sheet of full movie | Multiple Screenshots | 5s or 10s per frame, 720p |
| Editing image sequence | Multiple Screenshots | 0.1s (10 fps) or 0.2s (5 fps), source resolution |
| OCR / ML training input | Multiple Screenshots | 1s, native resolution, 8-bit |
| UI mockup with limited colors | Specific Frame | Indexed palette (16 / 64 / 256 colors) for smaller PNG |
| Print poster from 4K MKV | Specific Frame | 4320p preset, 300 DPI, 16-bit |
Use Specific Frame mode and enter the time in seconds with millisecond precision. For example, 12.450 means 12 seconds and 450 milliseconds into the MKV. Useful when you need the exact moment of an anime key frame, a movie's establishing shot, the perfect Plex poster scene, or a particular UI state in a long screen recording.
PNG is lossless DEFLATE compression — every pixel is preserved exactly. A 1080p photographic frame from an MKV is roughly 2-5 MB as PNG versus 200-500 KB as JPG. Across hundreds of frames the difference adds up. For graphic content with limited colors (animation, UI, screen recordings), an indexed-color PNG with a 16, 64, or 256-color palette can shrink dramatically while staying lossless within that palette.
Only burned-in (hardsubbed) subtitles will appear, because they're part of the actual video pixels. MKV soft subtitles — separate SRT, ASS, or PGS tracks stored alongside the video stream — are not rendered into frames during extraction. If you need the subtitle text in the still, hardsub the MKV first or composite the line in afterward.
Yes. H.265/HEVC is extremely common in 4K MKV rips (since around 2017), and AV1 is increasingly used for newer high-efficiency releases. Both decode for frame extraction here. 10-bit HEVC and HDR sources also extract — note that PNG output up to 16-bit per channel preserves more of the source range than 8-bit JPG, but HDR metadata itself is dropped.
Modern browsers (Chrome, Firefox, Safari, Edge) don't natively play the MKV container, even when the underlying codec (H.264, H.265) would work in MP4. Frame extraction runs on xconvert's servers, independently of the browser's playback layer — so codec quirks that block preview don't usually block extraction. If you also want native browser playback of the MKV, see MKV to MP4.
Almost no consumer video format carries an alpha channel — H.264, H.265, VP9, and AV1 inside MKV all encode opaque RGB. Exceptions are ProRes 4444 and VP9 with alpha. If your source has alpha, PNG preserves it; otherwise the extracted frame is fully opaque and you'd need to mask it manually in an image editor.
Multiply duration by capture rate. At 5 seconds per frame you'll get 1,440 stills — a manageable contact sheet of the whole film. At 1 second per frame you'll get 7,200. At 0.1s per frame (10 fps) you'll get 72,000 frames — fine for editing pipelines but a heavy ZIP at PNG file sizes. Pick the slowest interval that still captures the moments you need.
Frames extract on our servers. A single TV episode under ~2 GB extracts quickly. Full 4K Blu-ray rips (20-50 GB) take longer and are bound by upload size and connection speed; for those, consider trimming the relevant scene first or extracting at a sparser interval (5s, 10s) to keep the PNG count manageable.
PNG for sharp-line cel-shaded anime, screen recordings, UI screenshots, OCR / ML inputs, and print-quality stills where you want pixel-exact reproduction. JPG for live-action movies, TV, and photographic anime backgrounds where a 4K still under ~2 MB is acceptable. PNG is lossless but typically 5-10x larger than the equivalent JPG. See MKV to JPG for the lossy alternative.