I don't have time to do the diff and I think there may be enough differences that a refactor into a base and derived class is probably warranted. The philosophy is a bit strange. I need to continue for a jan release but wanted to send you this before the holidays, (including the warts.)
1) It is written only for H264, and my [7][8][5] keyframes are aggregated. 2) Both the [7][8][5] keyframe and the [1] diff frames are prefixed with the access_unit_delimeter nal [9] 3) It is more frame by frame that by low water mark, As small resolution video was coming in under the 1000 setting and causing me some trouble. Keeping it frame aligned for me worked because it made it easy to detect the keyframe. If we had a pointer to a struct that contained the buffer, then we could have metadata like the sourcerec does, a little tweaking and that may work using the sourceRec. Speaking of which, I temporarily made that array a LOT bigger because I needed to implement the whole standard and have 13 bit PID's. Would like to try a map as a sparse array. 4) The segmentDurationSecs_ is a static int at the moment, this obviously needs to be part of the CTOR 5) So what I am doing is detecting the keyframe and when I have enough keyframes to fill a segment, inserting a PAT then a PMT. I use mod so it always starts with a PAT and PMT. I have in my code an assumption of a keyframe per second, so this is really a keyframe count not seconds and it will be changed and will be enforced that segment size is an even multiple of the keyframes. Then my segmenter is trivail, all I have to do is detect the PAT and start a new segment file and add it to my index. Obviously this is a lot higher freq than the original class design, but apple likes it that way, every segment start with it. http://pastebin.com/PZvXpE7R http://pastebin.com/VHdMP8PQ http://pastebin.com/C7ufnTd8 http://pastebin.com/jE1KiYsa I didn't change any headers but I should of but a modified for iOS by... If you do, my name is spelled Jeff Shanab :) From: live-devel-boun...@ns.live555.com [mailto:live-devel-boun...@ns.live555.com] On Behalf Of Ross Finlayson Sent: Friday, December 23, 2011 9:24 AM To: LIVE555 Streaming Media - development & use Subject: Re: [Live-devel] invalid ts stream No, Imade a lot of changes back and forth trying to figure it out so I am not sure what is actually required and what is not. Could you please try to figure out the smallest set of changes that you needed to make to our code in order to get your TS file to work - and then send us a "diff" file, so I can make equivalent changes in our released code. Ross Finlayson Live Networks, Inc. http://www.live555.com/ ________________________________ No virus found in this message. Checked by AVG - www.avg.com<http://www.avg.com> Version: 2012.0.1901 / Virus Database: 2109/4698 - Release Date: 12/23/11
_______________________________________________ live-devel mailing list live-devel@lists.live555.com http://lists.live555.com/mailman/listinfo/live-devel
