Skip to content

Commit

Permalink
Update 01-l3d.inc.md
Browse files Browse the repository at this point in the history
  • Loading branch information
haudiobe authored Oct 2, 2024
1 parent 522291b commit 1cdef6a
Showing 1 changed file with 1 addition and 63 deletions.
64 changes: 1 addition & 63 deletions specs/l3d/01-l3d.inc.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,66 +4,4 @@

This feature description is an update to DASH-IF Low-Latency extensions [[!IOP5-PART4-LL]] adding additional features.

## Scenarios and Motivation ## {#l3d:scenarios}

In 2017, DVB (https://www.dvb.org) and DASH-IF (https://www.dashif.org) started to collaborate to address
Low-Latency DASH-based services. Initially, a joint group was formed that developed a report that was issued
by DVB/DASH-IF as _DASH-IF/DVB Report on Low-Latency Live Service with DASH_ [[!LL-REPORT]].

Based on this report, the DASH-IF Low-Latency extensions [[!IOP5-PART4-LL]] were developed and published, together with
test streams and dash.js implementation.

ISSUE: Add the references to dash.js

ISSUE: provide a summary of what could be achieved

* Reviewing the Report - what has changed
* Do have new requirements in terms of latency < 3 seconds?
* There are new requirements that sports events may needed less than 3 seconds, but not that hard
* We have the report here: [https://dashif.org/webRTC/report](https://dashif.org/webRTC/report)
* Betting use case may be of relevance.
* Please check if any of the use case matter as is or in variants for the work we are planning top
* Do we have new requirements to tune and switching?
* faster tune-in, time to first video is very important and should be low (1-2 seconds may be too high)
* Start playback faster – do not wait for up to 2 seconds
* Faster switching to avoid buffering
* Independency of joining latency, switching latency and segment duration
* Do we have new requirements for efficiency?
* _Compression efficiency_
* _Enables scene-boundary aligned segments_
* _Mini-GOP-sized partial segments for random access and events_
* Are there any new use cases that requires new technology?
* Ad insertion w/o stream conditioning
* Return from ad break to an arbitrary place – not necessarily an IDR
* LL-HLS compatibility
* Reuse LL-HLS partial segments
* Use DASH templates to keep same MPD update efficiency
* Trick modes
* Small partial segments can be used for smoother trick play
* Did the defined technology for LL-DASH not meet the requirements?
* The technology may meet the requirements in terms of latency, but at the expense of reduced quality, including more re-buffering and/or reduced compression efficiency
* Today's technology typically requires that segment durations are in the range of 2 seconds.
* Next Steps
* add the above logic to the introduction of the feature.
* Start documenting the technology by reference to 6th edition
* also document client requirements

## Content Offering Requirements and Recommendations ## {#l3d:content-offering}


## Client Implementation Requirements and Guidelines ## {#l3d:client}


## Examples ## {#l3d:examples}

NOTE: Add some MPD examples

## Reference Tools ## {#l3d:reference-tools}

NOTE: provide status for the following functionalities
* Dash.js
* Live Sim
* Test Vectors
* JCCP

## Additional Information ## {#l3d:additional-information}
The document is developed in [Google docs and can be accessed here](https://docs.google.com/document/d/1pj3Ex5AdmkfeYGd9KAXHEVercfee30NsjHg36L5EZos/edit?usp=sharing).

0 comments on commit 1cdef6a

Please sign in to comment.