Pre-processing before distribution

Balance the light before the picture is encoded.

LivePictureAI rebalances harsh sunlight and deep shadow in real time, before the picture reaches viewers, and hands the next stage in the chain an already-corrected signal. Keep your existing cameras. Processing runs on GPU infrastructure.

Book a demoEvaluate your footage
The problem

Sun and shadow are hard on the encoder too

A sunlit scene spans far more brightness than a camera sensor, the broadcast signal and video compression can carry at once. In outdoor sport, that gap shows up as players lost in shadow and sunlit areas that blow out.

Compression makes it worse. It spends its bits where there's contrast, and a frame split hard between blown highlight and crushed shadow is the worst case: the encoder burns bandwidth on noisy, high-contrast edges, and the viewer still gets a picture where half the field of play is hard to read.

The solution

A real-time layer ahead of encoding and distribution

LivePictureAI processes the feed frame by frame, lifting shadow detail and taming highlight glare together, so the picture that goes on to encoding and distribution is already balanced. It works on what the camera captured: nothing is generated, added or invented. In smaller productions it sits on the program output, the last step before the picture leaves the production; in large productions it sits on each camera feed before the vision mixer.

Before the vision mixer

Large productions

Cameras / feeds
LivePictureAIOne per camera feed
Vision mixer / switcher
Broadcast / stream
The director cuts with the enhanced pictures. Needs more compute: a GPU per camera feed.
On the program output

Smaller productions

Cameras / feeds
Vision mixer / switcher
LivePictureAIOn the program output
Broadcast / stream
One stream to process, so a lighter compute footprint. The cut is made as today and enhanced on the way out.
The business case

Potential benefits before distribution

Potential bandwidth savings

Can reduce the bitrate required for comparable visual quality, depending on the footage and encoding configuration.

A corrected signal for every output

Hands the next stage in the chain an already-corrected, broadcast-grade signal, so every output downstream starts from the balanced picture.

Live or file

The same pipeline processes live streams and video files, so pre-recorded material can be prepared with the same lighting correction.

Results depend on your footage, encoding configuration and delivery settings, so the bitrate effect should be measured with your own encoder. Explore an illustrative scenario in the savings calculator, or discuss your setup with us.

Integration

Fits the setup you already run

  • No new cameras or lenses
  • On camera feeds or program output
  • SMPTE 2110, NDI, SDI, RTSP & file
  • GPU edge or production-hub deployment
  • Maximum 100 ms of added latency
  • Keeps the colors of the original production

LivePictureAI was introduced at IBC 2026 and has been tested on football footage. We are currently validating it with selected partners.

Before / After

See the difference

Drag to compare the original frame with the LivePictureAI-processed version.

Before: Sports footage with harsh sunlight, shadows and lighting issuesAfter: AI-enhanced sports footage with balanced lighting

LivePictureAI automatically reduces harsh sunlight and deep shadows to balance lighting, in real time.

FAQ

Pre-processing, answered

What does AI pre-processing before distribution mean?
LivePictureAI sits in the signal chain before the picture reaches viewers. It rebalances the sunlit and shaded areas of each frame in real time, so the encoder and everything after it receive an already-corrected, broadcast-grade signal.
Can it lower the bitrate we need for distribution?
A balanced picture can reduce the bitrate required for comparable visual quality, depending on the footage and encoding configuration. This should be assessed using your intended encoder and delivery settings, so we measure it rather than assume it.
Where in the chain does it go?
It depends on the production. In large productions it sits on each camera feed before the vision mixer, so the director cuts with enhanced pictures; this needs more compute, a GPU per feed. In smaller productions it sits after the vision mixer on the program output, so there is one stream to process before encoding.
Does it change the look of our production?
No. LivePictureAI balances the light rather than applying a look, and keeps the colors of the original production. Nothing is generated, added or invented: it only lifts what the camera captured, and the producer decides how much shadow stays. Any creative grade stays with your team.
Which formats does it accept, and how much latency does it add?
SMPTE 2110, NDI, SDI, RTSP and file-based streams, deployed on GPU-based edge or in the production hub. It adds at most 100 ms of latency, built to keep pace with live broadcast timing.
Does it work on recorded material as well as live feeds?
Yes. It was built for real-time live streams, and the same pipeline can also process video files, so pre-recorded material can be prepared with the same lighting correction before it is distributed.
Can we test it on our own footage?
Yes. Share a clip affected by sun and shadow and we'll return a sample enhancement, so you can run it through your own encoder and judge the result.

More on formats and placement on the broadcaster & OB page and in low-latency AI in the OB truck. For recorded matches, see archive football footage enhancement.

Get in Touch

Interested in LivePictureAI?

Book a call or send us a message, whichever suits you.

30 minutes Teams link Instant confirmation

Grab 30 minutes with the team and pick a time that suits you. You'll get instant confirmation and a Microsoft Teams link.

See available times

Prefer to write? instead.