For broadcasters & OB

Broadcast-grade lighting correction, in your live chain.

A real-time AI processing layer for live outdoor sports, on the camera feeds before your vision mixer or on the program output, designed for professional broadcast workflows. Keep your existing cameras. Processing runs on GPU infrastructure. Non-generative AI: nothing is added or invented, and no artifacts.

Book a demoEvaluate your footage
SDINDISMPTE ST 2110RTSPUp to 4K60Full specifications
Where it sits

One layer, two setups

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.
Built for live workflows

Designed for your signal chain

Standards-native

SMPTE 2110, NDI, SDI, RTSP and file-based streams, up to 4K60.

Flexible deployment

GPU-based edge or production-hub deployment.

Pipeline integration

On each camera feed before the vision mixer, or on the program output.

Real-time inference

Runs on live feeds, engineered for broadcast timing.

Keep your existing cameras

Processing runs on GPU infrastructure and adds a lighting-correction layer to your production chain.

Broadcast-grade out

Hands the next stage in the chain an already-corrected signal.

The business case

Potential benefits for your production

Potential bandwidth savings

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

Potential reduction in supplementary lighting

Potential to reduce reliance on supplementary daytime lighting, subject to venue and production requirements.

Less manual sun-and-shadow balancing

Can reduce manual intervention for sun-and-shadow balancing.

Operational and financial benefits depend on your footage, encoding configuration and production requirements. Explore an illustrative scenario in the savings calculator, or discuss your setup with us.

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

Integration, answered

Where does LivePictureAI sit in the chain?
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.
Which signal formats are supported?
SMPTE 2110, NDI, SDI, RTSP and file-based streams, up to 4K60 (2160p60) as well as 1080p60. Deployment is designed for GPU-based edge or the production hub, on-premises or in the cloud.
Can AI replace manual camera shading, or our vision engineers and graders?
Not entirely. LivePictureAI can reduce manual intervention for sun-and-shadow balancing. Vision engineers remain responsible for camera matching, exposure decisions and overall production quality.
Do we have to change our cameras or workflow?
Keep your existing cameras. LivePictureAI adds a GPU-based processing layer to the signal chain. Integration requirements depend on your existing infrastructure and are discussed as part of the evaluation.
What hardware does it need?
It runs on GPU infrastructure, either on-premises or in the cloud. Contact us and we will size the setup for your production.
Does the AI generate or invent anything in the picture?
No. LivePictureAI is non-generative AI: nothing is added or invented, and no artifacts. It only rebalances the light in what the camera captured, and the operator decides how much shadow to keep or lift.
What about latency?
Under 5 ms of processing per frame at 1080p60 and at most 10 ms at 4K60, with at most 100 ms end-to-end including I/O and buffering, built to keep pace with live broadcast timing.

Working in football specifically? See the live football page, or outdoor sports for rugby, cricket, tennis and more. All specs are on broadcast integration. For the integration detail, read NDI and SDI integration and low-latency AI in the OB truck. Balancing the picture ahead of encoding? See pre-processing before distribution.

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.