sk-video

library1.1.10pypypi✓ verified 83d ago

scikit-video is a Python library for video processing built on top of numpy, scipy, and ffmpeg. Current version 1.1.10 provides functions for reading, writing, and manipulating video files as NumPy arrays. The library is in maintenance mode with very infrequent releases; no new features are expected.

devops

pip install sk-video

homepage ↗

INSTALL

IMPORT

SIG · SK-VIDEO

sk-video

devopspythonv1.1.10

harness data pending

Install & Compatibility

Where this runs

No compatibility data collected yet for this library.

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

vread

✓ from skvideo.io import vread

Standard import for reading video files

FFmpegReader

✓ from skvideo.io import FFmpegReader

✗ from skvideo.io import VideoReader

VideoReader class was renamed to FFmpegReader in early versions; VideoReader is a legacy alias that may not exist

skvideo.measure

✓ import skvideo.measure

Module for video quality metrics (e.g., SSIM, PSNR)

Basic reading and writing of video files. Ensure ffmpeg is installed and in PATH.

import numpy as np from skvideo.io import vread, vwrite # Read a video file video = vread('input.mp4') print(f'Video shape: {video.shape}') # (T, H, W, C) # Write a video file vwrite('output.mp4', video)

Debug

Known issues

breakingInstallation requires ffmpeg executables on system PATH. sk-video does not bundle ffmpeg. If ffmpeg is missing, import skvideo will raise ImportError.

fix

Install ffmpeg via your system package manager (e.g., apt install ffmpeg, brew install ffmpeg) or download from ffmpeg.org.

affects: all

gotchavwrite returns None on success, does not raise exceptions on failure. Always check that the output file exists after writing.

fix

Manually verify file existence with os.path.exists after vwrite.

affects: all

deprecatedThe skvideo.datasets module is no longer maintained and may not work with newer versions of the library.

fix

Use your own video files for testing or download from external sources.

affects: >=1.1.0

breakingskvideo.io.FFmpegWriter has inconsistent behavior across platforms; may produce corrupted files on Windows.

fix

Consider using vwrite or switch to another library (e.g., OpenCV, imageio) for writing video on Windows.

affects: all

Errors

Common errors & fixes

ImportError: libavformat.so.58: cannot open shared object file: No such file or directory

Missing ffmpeg dynamic libraries at runtime (Linux only, when ffmpeg is not installed system-wide).

fix

Install ffmpeg via apt or compile from source. Ensure LD_LIBRARY_PATH includes ffmpeg's lib directory.

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

Passing a NumPy array as a boolean condition inside skvideo's code (often due to unexpected dtype or shape).

fix

Ensure video arrays are contiguous and have dtype uint8. Call np.ascontiguousarray(video) before passing to skvideo functions.

RuntimeError: Could not determine frame count

FFmpeg reader could not parse the video header; often due to corrupted file or unsupported codec.

fix

Verify the video is playable with ffplay. Re-encode with standard codec (H.264, AAC).

Upgrade

Version history

1.1.10latest on PyPI · released Dec 12, 2017

Audit

Dependencies

numpyrequiredcore dependency for array operations

scipyrequiredused for image processing filters

ffmpegrequiredsystem-level dependency for video codec support; must be installed externally

Agent activity

24 hits · last 30 days

node

Amazon

Resources

homepagehttp://scikit-video.org/ ↗