Expanding Transcoding Architecture with AWS Elemental MediaConvert

Customer Challenge

TrackIt was recently engaged by a video technology provider that relied on a Kubernetes-based transcoding pipeline built around FFmpeg to process and prepare video assets. As part of its platform evolution, the customer initiated the implementation of a MediaConvert-based workflow alongside its existing FFmpeg pipeline.

The objective was to recreate the existing pipeline’s behavior on AWS while maintaining parity across key elements such as job metadata, transcoding profiles, output formats, and overall processing logic. This included accounting for existing API and messaging integrations.

In parallel, the customer required a clearly defined target architecture covering orchestration, transcoding, event handling, observability, and security controls, along with a structured approach to validation and cutover.

Implementation

TrackIt designed and implemented a target architecture centered on AWS Elemental MediaConvert.

The engagement began with a detailed assessment of the existing Kubernetes and FFmpeg-based workflow, including job definitions, profile configurations and throughput patterns. This informed the design of a target AWS architecture aligned with customer requirements.

The implementation covered both infrastructure and workflow integration:

• Designing the AWS architecture for job orchestration, transcoding, event handling, observability, and security controls aligned with customer requirements.
• Provisioning the messaging and routing infrastructure using Terraform, and defining the target AWS resources required to support the workflow.
• Implementing integration with existing API and messaging systems.
• Integrating MediaConvert into the existing platform to support routing between the legacy and AWS-based transcoding workflows.
• Integrating job status and completion notifications with existing systems to maintain compatibility with downstream consumers.
• Recreating and validating transcoding profiles based on the existing FFmpeg configuration, including MP4 and WebM renditions, rate-control behavior, and audio track mappings.
• Configuring MediaConvert job settings and output profiles to match existing transcoding behavior. 

Validation and cutover planning were also integral to the implementation:

• Validating output for in-scope profiles and transcoding behavior.
• Benchmarking transcoding quality against the existing FFmpeg pipeline using objective video quality metrics.
• Defining a side-by-side validation approach, enabling the execution of the existing pipeline and the AWS workflow in parallel.
• Establishing quality benchmarks, performance metrics, and criteria for determining rollout readiness.

This resulted in a MediaConvert-based workflow integrated with existing systems and ready for controlled rollout.

Outcome

The engagement resulted in the implementation of an AWS-backed transcoding workflow built on AWS Elemental MediaConvert, integrating orchestration, API-driven job handling, event propagation, and profile-based processing aligned with the existing pipeline.

The system reproduces key aspects of the previous FFmpeg-based workflow, including transcoding profiles and integration patterns, while introducing a structured AWS architecture with defined security, observability, and infrastructure provisioning practices.