First of all, I would like to congratulate you with the fine work on
Dirac, the specification and both implementations. I have followed
Diracs development for years. Several years ago I worked on an
implementation of Dirac for FFmpeg (which I still have to finish when
I have some spare time).
Recently I have started to work as a PhD student. For my research I
am looking for an adaptive streaming application. This means that
environmental conditions can, for example, influence the required
processing power. Usually such changes do not occur very often and
after the application adapted to these changing circumstances, it
remains static for a while in this sense.
My question is: Does Dirac (in some of its applications) require
adaptivity? In that case, I might use Dirac as a driver application
for my research. The are some things which at least gave me the idea
that Dirac might have the kind of adaptivity I am looking for:
- Multiple sequences in a stream (changing frame rate / picture size, etc)
Is this used in practice?
- A varying quality, by using different IDWT algorithms, changing
interpolation or by using more reference frames (for example,
because not enough bandwidth is possible). Although in this kind of
application my "static" requirement might not hold anymore if the
stream adapts continuously, requiring different processing power
continuously.
Although these are just two ideas that might not make sense (I am not
a video coding expert at all), I am mainly looking for (future)
applications of Dirac that require such (static) adaptivity, any
suggestion is more than welcome.
Regards,
Marco
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello,
First of all, I would like to congratulate you with the fine work on
Dirac, the specification and both implementations. I have followed
Diracs development for years. Several years ago I worked on an
implementation of Dirac for FFmpeg (which I still have to finish when
I have some spare time).
Recently I have started to work as a PhD student. For my research I
am looking for an adaptive streaming application. This means that
environmental conditions can, for example, influence the required
processing power. Usually such changes do not occur very often and
after the application adapted to these changing circumstances, it
remains static for a while in this sense.
My question is: Does Dirac (in some of its applications) require
adaptivity? In that case, I might use Dirac as a driver application
for my research. The are some things which at least gave me the idea
that Dirac might have the kind of adaptivity I am looking for:
- Multiple sequences in a stream (changing frame rate / picture size, etc)
Is this used in practice?
- A varying quality, by using different IDWT algorithms, changing
interpolation or by using more reference frames (for example,
because not enough bandwidth is possible). Although in this kind of
application my "static" requirement might not hold anymore if the
stream adapts continuously, requiring different processing power
continuously.
Although these are just two ideas that might not make sense (I am not
a video coding expert at all), I am mainly looking for (future)
applications of Dirac that require such (static) adaptivity, any
suggestion is more than welcome.
Regards,
Marco