Live delivery, recording and replay are often built as separate pipelines. Each pipeline composes speakers, slides, captions and branding again, which creates inconsistent outputs and repeated processing. This whitepaper defines Program Output as the single governed programme the audience is meant to see and hear, then follows that authority through live routes, recording, replay and destination delivery.
Executive summary
- Program Output is an authorised programme identity and timeline, not merely one encoded file.
- Compose combines approved elements according to versioned layout and programme policy.
- Live delivery, Record, Replay and destination fan-out can consume the same approved programme meaning.
- Recording readiness requires complete lineage and policy checks, not only the existence of a file.
- Reuse can reduce repeated work, but every performance or cost result must be measured against a matched baseline.
The problem with separate output pipelines
A live audience may receive one composition, the recording provider may create another and the replay workflow may produce a third. Speaker order, slide timing, captions, overlays and branding can diverge. When someone asks which version represents the approved programme, the answer depends on which pipeline they inspect.
Separate pipelines also repeat decisions and processing. They need their own layouts, source mappings, failure handling and readiness rules. A change made for live delivery may never reach replay. A recording file may exist even though the wrong source or incomplete presentation state was captured. Program Output creates one authority before those downstream paths begin.
What Program Output represents
Program Output is the governed programme the audience is authorised to receive. It connects the approved session, source elements, presentation state, composition policy, timeline and output identity. It can be rendered into more than one technical form, but those forms point back to the same programme meaning.
The identity is durable across transient provider resources. A live route can change. A recording job can retry. A replay asset can be packaged later. Those operations do not create a new customer-facing programme unless an authorised revision explicitly does so.
Composition is an authorised decision
Compose selects approved sources and places them into a versioned layout. A producer can change the layout or active source within granted authority. Captions, branding, overlays and presentation state join the output according to explicit policy. The customer product owns the experience and role decisions, while cRUD Morph owns the resulting media composition and evidence.
Server authority matters because a client preview cannot define what every destination received. The composition state must be durable and ordered. A late or reconnecting operator needs the current snapshot. Recording and destination workers need the same approved version rather than a client-local guess.
Live delivery can use several paths
Interactive speakers and a large audience do not need the same transport. Deliver can route approved elements and Program Output through paths suited to each role. A speaker may participate through an interactive path while most viewers receive a scalable audience output. The programme remains coherent because both paths reference the same session and output authority.
Destination fan-out follows the same rule. A social destination, customer CDN or other approved endpoint can receive an output with its own receipt and failure state. A destination failure does not rewrite the Program Output or silently mark the complete programme failed. It remains one observable branch of the authorised work.
Recording begins with intent
A reliable recording does not begin when a file appears. The session establishes recording intent, approved sources, output identity, policy and expected evidence before capture. Record then preserves the authorised Program Output or other explicitly approved source form. The resulting object carries lineage back to that decision.
A provider may produce segments, tracks or a composite. Those are implementation results. cRUD Morph normalises their state and relates them to the approved recording. Missing segments, incomplete composition or uncertain lineage remain visible. The product does not receive a ready state until the defined checks have passed.
Readiness is a product state
File existence proves only that storage contains bytes. Replay readiness may also require finalisation, integrity, expected duration, approved composition, captions, packaging, access policy, retention and lineage. The rule should be explicit and repeatable.
An operation can therefore be captured but not ready, processing but not ready or failed with partial assets retained for investigation. Unknown remains available when evidence is incomplete. This prevents a premature replay link from becoming the customer's first discovery that recording did not complete correctly.
Replay should reuse authorised work
Replay uses the approved programme, readiness state and entitlement policy. It should not rebuild the session from disconnected provider files unless a new authorised processing step requires it. A viewer receives the replay under time-limited authority, and the operating record can identify the source, output version and access decision.
The same lineage supports Simulive and bounded Evergreen use. A scheduled recorded experience can identify its source and programme version. If the content changes, the system preserves which viewers received each version. Reuse remains governed rather than becoming an endless public recording with unknown context.
Source, programme and rendition are different
A source is approved input, such as a speaker track, presentation asset or screen-content feed. Program Output is the authorised composition and timeline created from those inputs. A rendition is one technical form used to deliver or preserve that output. Keeping these identities separate prevents a provider file from being mistaken for the programme itself.
The distinction also makes later change safer. A new rendition can be created for a device or destination without changing programme meaning. A corrected source can create a new authorised programme version without erasing the earlier lineage. A replay package can expire while the retained recording and its evidence follow a different policy. Each transition is explicit rather than hidden behind a filename.
Product teams should test their data model against these three identities. If one provider asset identifier stands for source, composition, recording and replay at the same time, future migration and investigation will be difficult. The customer-facing contract should remain stable while provider resources and technical renditions can change beneath it.
Failure, retry and retention
Each downstream operation needs its own state while remaining connected to the output. A destination can retry without duplicating the programme. A recording job can resume or reconcile provider evidence. Replay packaging can fail without erasing the preserved recording. Idempotency keeps repeated commands from producing uncontrolled copies.
Retention and deletion follow explicit ownership and policy. Recording assets, replay packages and temporary processing objects may have different lifecycles. Optimize can remove unnecessary renditions or repeated work, but it cannot delete evidence or customer-authorised assets outside the approved rule.
How to evaluate a live-to-replay architecture
Run one real programme through live delivery, a recording interruption, late finalisation, replay packaging and access expiry. Change composition during the session and verify which version was captured. Fail one destination while another remains healthy. Repeat the same command. Replace a presentation source before a second run. The final evidence should identify every authorised version and outcome.
Compare the amount of repeated composition, processing and storage with a matched design. Inspect quality and timing, not only cost. Reuse is valuable when it preserves the intended programme and removes unnecessary work. It is not valuable if it hides differences or produces a lower-quality result.
Conclusion
One Program Output gives live delivery, recording and replay a shared authority without forcing every path to use the same transport or file. It connects composition, timeline, lineage, readiness and access so that reuse remains explainable. This is the foundation for producing the programme once in meaning, then delivering it through the forms each audience and destination require.
Basis and methodology
This whitepaper explains the public roles of Compose, Deliver, Record and Replay in cRUD Morph. It contains no external benchmark and makes no universal processing, storage or delivery saving claim. Supported outputs and measurable results are confirmed for each engagement.
For the public product boundary, review the cRUD Morph approach, capabilities and documentation.
Frequently asked questions
How can we verify that a replay comes from the intended Program Output?
Can the same output support Simulive and Evergreen delivery?
Does reuse always reduce media cost?
About this whitepaper. It explains the public product and operating principles of cRUD Morph. It does not disclose source code, private implementation, provider credentials or customer data.