What do we do?
We analyse video codecs and find best encoding settings for given videos. Our settings increase encoding efficiency: faster encoding, lower bitrate for the same quality, or better quality for the same bitrate.
Video-codec tuning is a process of codec optimisation for particular tasks, such as compression of specific videos (movies, TV series, CCTV, interviews, CGI etc.) or working at given speed limits by variating encoding settings.
We offer a solution to the problem of optimal video codec configuration in the form of software for automatic selection of effective parameters and in the form of service, the result of which will be one or more of the most effective configurations of video codec parameters for your video sets.
Our clients are owners of video on demand (VOD) and streaming services, online cinemas, holders of large video archives, video codec developers, etc.
The challenge of video-codec tuning
Standard encoding presets are usually tuned by the developers to work efficiently on average for a variety of encoding cases. Adjusting your codec for content it works with allows to utilize unused reserves of video codec and boost its performance.
Is it simple to find good encoding settings?
Let's consider an example: tune x264 for one 20-second-long video
- x264 has ~50 encoding settings
- Different combinations of settings influence encoding performance in an unpredictable way
- Exhaustive search for optimal settings will last 22,000,000,000,000 centuries (~488,000 Earth ages!)
Our algorithms can tune any video codec for your videos
in just a few days
Do you want to optimize your encoding?
send you the report by email
Don’t believe it will work
for your videos?
Try it for free!
Send us your videos and we'll show you the results for your content
Our stack of technologies
Our stack of technologies and programming languages: Python, C++, NumPy, XGBoost, scikit-learn, CatBoost, OpenCV, FFmpeg, Docker, etc. We build, develop and keep up-to-date models linking video codec parameters, source video characteristics and compressed video characteristics through machine learning such as “learning with a teacher” and “reinforcement learning” using various machine learning methods.
Developed models are the basis of the software we offer for the automatic selection of effective encoding settings. The software recognizes the features of the video stream and selects the best codec configuration based on the encoder model. Our models also allow us to quickly find effective configurations of video codec parameters within the framework of consulting services. At the request of customers, we develop or customize software related to video encoding, decoding, video quality assessment, video processing to improve its quality
About company
Our team has 15 years of experience in video encoding and processing. We analyze video codecs and offer our software and consulting services to improve video encoding efficiency
Limited Liability Company “EVT”, TIN 7724494903, TRR 772401001, PSRN 1197746709984, address 22 Kaspiyskaya str., block 1, bld. 5, fl. 5, premises 9, room 17, office 16, 115304 Moscow, Russian Federation