I now want to implement bit rate switching so that the video switches from high to low bit rate whenever there is a low bandwidth seamlessly. I understand I can use the Wowza trasnscoder addon to achieve the transrating so that the incoming stream is split into low, medium and high bitrate streams which are key-frame aligned. After this, I should be creating an SMIL file to specify the various bit rates.
Now, how will this work with nDVR recording API ?
Say I have the incoming stream split into stream_high_bitrate, stream_medium_rate and stream_low_bitrate which are of different bitrates.
When I require to start the recording with a name ‘recording_1’, should I be starting up all the 3 bitrates like this ?
In that case, would SMIL file contain the recording names instead of the stream names ? The recording names keep changing, so how can this be updated in SMIL file dynamically ? And what if there is a latency between starting the three recordings ?
Starting the recording of the streams simultaneously would be very helpful. Could you provide the source code to do this? I would imagine it is a common requirement?
you look at the recording ndvr article,
There is http provider source code. Which you can modify to do
Typically HTTP streaming requires 3 chunks of data to begin streaming, so there is some expected latency. I don’t think it would be different latency for adaptive bitrate. What are you seeing?
Starting the recording of the streams simultaneously would be very helpful. Could you provide the source code to do this? I would imagine it is a common requirement?
There were questions from db_2011 but he asked them many time ago.
I ask again similiar questions. Maybe there was any progress in this case:
How can I achieve adaptive bitrate with dvr recording API? I mean I have got 3 bitrates (200 kbps, 300 kbps, 500 kbps) of channel “first_ch” which I can record. And by recording API I’ll record them into 3 different folders which I’ll name like “first_ch_200” “first_ch_300” “first_ch_500” by requests like these:
I can do this by default code of your Recording API. And I can request them or parts of them selectively by Playlist Request Delegate API by URLs like that:
And my question is how can I get them simultaneously with adaptive feature? As I can do with plain streams by grouping them into smil file and then I can request them with adaptive feature:
When I starting up all the 3 bitrates like that requests that I wrote above there will be a little latency between start times of that three recordings. So what if it is? And does Playlist Request Delegate API process it right? And in general does Playlist Request Delegate API can be implemented in such requests?
Typically HTTP streaming requires 3 chunks of data to begin streaming, so there is some expected latency. I don’t think it would be different latency for adaptive bitrate. What are you seeing?
-Lisa
Hi!
Yes I know about this but I’ve asked about case when I use Recording API and questions regard only that case.
I’ve run three streams at once with requests that I wrote above and than generate smil file which refers to folder names of my recordings and all works!