Browse categories
Explore
Fiverr Pro
English
$
USD
I will build multiview pip and video compositing pipelines
Sasikiran G
About this gig
Looking to implement Multiview, Picture-in-Picture (PIP), or multi-channel video compositing for sports streaming, OTT broadcast, or surveillance?
I am a Linux Multimedia Systems Engineer with 4+ years of production experience building high-density GStreamer and FFmpeg compositing pipelines in C, C++, and Python.
What I Can Build For You:
Server-Side Multiview: Synchronize and composite multiple live RTSP, SRT, or RTMP streams into 2x2, 3x3, or custom grid layouts (compositor / glvideomixer).
Dynamic Layouts: Implement seamless Picture-in-Picture (PIP), L-Band graphics, and Pushup transitions without stream drops.
Audio/Video Synchronization: Fix PTS/DTS drift and eliminate clock desync across disparate multi-camera feeds.
Hardware-Accelerated Mixing: Offload video decoding, scaling, and compositing to GPU/SoC hardware (NVENC, VAAPI, Jetson, QuickSync) to minimize CPU load.
Multi-Protocol Delivery: Ingest from SRT/RTMP/UDP and transcode directly into low-latency HLS, DASH, or RTSP.
Tech Stack:
GStreamer (compositor, videomixer, interpipes), FFmpeg, C, C++, Python, Linux, SRT, RTMP, HLS, WebRTC, H.264/H.265.
Please message me with your input streams, resolution.
Respect third-party rights
Please be aware that it is against Fiverr's policies for sellers to include themes, templates, or any other elements that infringe third-party rights or applicable laws in the delivered work. Read more about in our Guide to Responsible Digital Creation.
Get to know Sasikiran G
Sasikiran G
Media Developer Research Engineer
- FromIndia
- Member sinceMay 2026
- Avg. response time1 hour
Languages
English
I am an innovative Software Engineer with 5 years of experience in multimedia systems, backend development, I specialize in GStreamer, C/C++, and Python, with a proven track record of designing scalable multimedia applications and optimizing video/audio streaming protocols.
FAQ
What input stream protocols can you composite into a multiview grid?
I can ingest and mix almost any standard live protocol, including RTSP, SRT, RTMP, UDP, RTP, HLS feeds, and local V4L2 capture cards or video files.
How do you prevent audio and video from drifting out of sync across multiple feeds?
I implement shared pipeline reference clocks, configure accurate jitter buffer queues, and lock presentation timestamps (PTS/DTS) across all branches using GStreamer compositor elements so all feeds stay synchronized.
Can you dynamically change layouts (e.g., switch PIP to full-screen or toggle an L-band)?
Yes. Using GStreamer pad properties, dynamic element linking, or controller APIs in C++ or Python, feeds can expand to full-screen, switch positions, or display overlay graphics without dropping the active stream.
Do you support hardware acceleration for multi-stream mixing?
Yes. Multi-stream compositing is GPU-heavy, so I configure hardware acceleration using NVIDIA (NVDEC, NVENC, CUDA/NVMM), Intel VAAPI/QuickSync, or embedded ARM SoCs (like Jetson) to keep CPU usage low.
What will I receive as the final deliverable?
Depending on your package: fully documented, ready-to-run GStreamer pipeline commands, a native C/C++ or Python media application, and step-by-step instructions to deploy and run it on your target Linux machine or server.
