Integration of Dataflow Components Within a Legacy Video Transcoding Framework

Tewodros Deneke 1 Lionel Morel 2 Sébastien Lafond 1 Johan Lilius 1
2 SOCRATE - Software and Cognitive radio for telecommunications
Inria Grenoble - Rhône-Alpes, CITI - CITI Centre of Innovation in Telecommunications and Integration of services
Abstract : Recently the RVC-CAL dataflow language has enabled video codecs to be specified in a more natural way than imperative languages by allowing implicit expression of parallelism and side effect freeness. The tools developed for RVC-CAL have also enabled the automatic generation of parallel C code, among others, from dataflow specifications. This paper introduces a new approach allowing the integration of dataflow components within legacy code. The approach makes use of a generic interface definition that allows seamless interaction between I/O components, which are mostly state operations and are best implemented in imperative languages with data processing components which are mostly stateless dataflow operations and are best implemented in dataflow languages. The advantage of the approach is the ease of development by allowing each language to be used on those parts of the application that it is most appropriate for. The functionality of the approach is demonstrated by using the generic interface to add a new dataflow based MPEG and HEVC decoder into the legacy video transcoding library FFmpeg.
Document type :
Conference papers
Complete list of metadatas

Cited literature [14 references]  Display  Hide  Download

https://hal.inria.fr/hal-01289532
Contributor : Lionel Morel <>
Submitted on : Wednesday, March 16, 2016 - 8:48:22 PM
Last modification on : Tuesday, November 19, 2019 - 12:30:23 PM
Long-term archiving on: Sunday, November 13, 2016 - 7:53:41 PM

File

dataflow_transcoder.pdf
Files produced by the author(s)

Identifiers

Collections

Citation

Tewodros Deneke, Lionel Morel, Sébastien Lafond, Johan Lilius. Integration of Dataflow Components Within a Legacy Video Transcoding Framework. 2015 IEEE Workshop on Signal Processing Systems, SiPS 2015, Oct 2015, Hangzhou, China. ⟨10.1109/SiPS.2015.7345013⟩. ⟨hal-01289532⟩

Share

Metrics

Record views

139

Files downloads

209