Joe,
Thanks, I had envisioned a potential NiFi processor that could take the audio
stream and save things in some sort of sound file form like WAVs. Once in
something like a WAV form, then NiFi already helps us run things like sox to
convert in whatever format we need and then we can try to run tools on the wav
and grab more metadata in json format if we can milk anything out. However, we
seem to also be having a requirement that is in flux as to the details that
would require us to take a stream and forward it elsewhere. I also realize
that this second thing (stream forwarding) is the not necessarily best
situation to pick NiFi for. But, to extract sound files from a stream from one
of these protocol sets and dump things into a file, that could be a useful
thing for a NiFi custom processor down the road to be able to do. However, I’m
currently leaning on not attempting to write such a custom processor for myself
for either stream forwarding or chunking the files as I found SIP/RTP and
RTSP/RTP protocols on the surface to be a little bit more than I might be able
to chew in the painful time constraints provided.
Ideally, I’d love the extraction ability down the road of a custom processor
from a stream to something like a WAV as that would potentially enable me to
take one NiFi instance and forward the WAV or a zip of WAVs onto another remote
NiFi instance we have for onto further analysis. The eventual ability to
extract a stream and chunk it to a file would be nice as that would allow us to
forward the file to another NiFi instance for futher analysis.
Regards,
Jeremy
On 9/8/17, 3:52 PM, "Joe Witt" <joe.w...@gmail.com> wrote:
Jeremy
There are no current plans or efforts to ingest SIP/RTP/RTSP streams
at this time in NiFi that I am aware of. The processor to do so would
be fairly straight forward but you would have to think about the
chunking of the audio into flow files but this is pretty reasonable.
NiFi wouldn't be great most likely if the idea was to ingest these
audio streams then restream them back out as they were received. But
if it is to ingest them into chunks, do some sort of
extraction/analysis and then produce those results it could be a great
fit.
Thanks
On Fri, Sep 8, 2017 at 10:21 AM, Jeremy Taylor
<jeremy.tay...@acesinc.net> wrote:
> Greetings,
>
> We have received requirements for ingesting streaming audio that are in
flux
> and subject to change. We will either need to ingest audio being sent via
> SIP/RTP means or RTSP/RTP. I’ve looked for whether any existing NiFi
> processors would be able to ingest this and I’ve looked for whether there
> are any custom processors already written that might be able to read in
> these streams as well. Yes, I’ve noted that we have a ListenTCP and
> ListenUDP set of processors, but those won’t be enough in this situation.
> Is anyone aware of any custom processors in existence that I might not be
> aware of yet to read in streams via SIP/RTP or RTSP/RTP? Also, out of
> curiousity, is anyone aware of any plans for NiFi to have an eventual
> processor or processors to read streams in either via SIP/RTP or RTSP/RTP?
> My expectations are currently low and I realize that we may not be able to
> use NiFi for this particular ingest requirement, but I wanted to just ask
> just in case.
>
>
>
> Regards,
>
> Jeremy
>
>
>
>
