Announcement

Collapse
No announcement yet.

NSV + Flash = yes

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Originally posted by Jkey
    I take back anything bad i ever said about you smelter
    Me too

    although, if anyone asks, I'd say jkey made me say it!

    Of course ... I still say bad things about red5 - just for the giggles
    "If you don't like DNAS, write your own damn system"

    So I did

    Comment


    • #17
      If anyone is serious about getting some action and want me to set you up with this, hit me on a P.M.

      You can wait for the the functions and sample code to be added to the red5 trunk and D.I.Y. or you can get started with me now. My rates are very reasonable.

      Comment


      • #18
        Originally posted by Smelter
        My rates are very reasonable.
        about 128kbit?

        boom tish

        I've spent a couple of hours trying to achieve the same result using haxe

        stream from nsv to haxe (nekovm) - check
        stream from haxe (nekovm) to flash - well ... there's data flowing, but no pic or sound, no errors either

        I think I need to solve the pickle matrix
        "If you don't like DNAS, write your own damn system"

        So I did

        Comment


        • #19


          YOu have to pull frame data out of nsv container and put it in flv container, and you need to parse the aac data to get time-span values for the frames. Then you need to serialize the flv data to rtmp packets. After that then you can create a rtmp port in haxe for flash player to connect to and after the flash player requests the stream, begin sending the data.

          Comment


          • #20
            Here is the rtmp spec you need to 'talk' to flash player.



            You have a long row to hoe, my friend.

            Comment


            • #21
              yeah, the aac is proving to be difficult

              not sure about the video either ... doesn't the handshake need to be "correct" for flash player to stream vp6 (and aac for that matter) anyway - I don't think I have the handshake quite right

              oh, the rtmp side of things is all fine

              the AAC sequence header thing has me totally going round the bend at the moment - my linux box died a horrible death, much of my aac "knowledge" was on it
              "If you don't like DNAS, write your own damn system"

              So I did

              Comment


              • #22
                Hmm really, you are parsing nsv in to flv?

                Wow, good job, although I figure you still have a couple months to go.... There are quite a few project out there to copy from, so maybe sooner, but I expect later if at all. Prove me wrong

                Comment


                • #23
                  Originally posted by Smelter
                  Hmm really, you are parsing nsv in to flv?
                  Parsing is such a relative word

                  Currently, I'm successfully processing the incoming NSV, identifying in each frame the aux data (ignored for now), video data and audio data - that was easy - I could at this stage very easily output raw VP6 data and raw aac data (actually, it's ADTS, but that's another story)

                  Originally posted by Smelter
                  Wow, good job, although I figure you still have a couple months to go.... There are quite a few project out there to copy from, so maybe sooner, but I expect later if at all. Prove me wrong
                  I haven't found anything that specifically handles NSV to FLV - knowing the specs of both is a start though

                  My main concern at the moment is trying to ascertain if I have the handshaking right
                  "If you don't like DNAS, write your own damn system"

                  So I did

                  Comment


                  • #24
                    Im talking about handshaking. As far as nsv->flv, unless anyone else is holding back, Im the only one with it and offering to set it up. Hell, we are the only ones who know what it is.

                    Comment


                    • #25
                      doesn't someone already offer it ... streamsolutions, http://www.streamsolutions.co.uk/audio/flash
                      "If you don't like DNAS, write your own damn system"

                      So I did

                      Comment


                      • #26
                        audio only.

                        I am the only one supporting video

                        Comment


                        • #27
                          I think they use wowza, and red5 is the only one relaying video.

                          Comment


                          • #28
                            ugh ... this is just way too much like real work!! I'm sure my issue is the handshake ... but I'm sure I've duplicated red5's handshake "authentication" - barely any different from what I had - there's just no indication if the handshake is "working"
                            "If you don't like DNAS, write your own damn system"

                            So I did

                            Comment


                            • #29
                              It is very clear what the rtmp spec says to do. You should be using that instead of copying red5.

                              What is even harder than the handshake and connection is the serialization to rtmp.

                              Comment


                              • #30
                                [QUOTE]Originally posted by Smelter
                                It is very clear what the rtmp spec says to do. You should be using that instead of copying red5.
                                [/B]
                                so, forget all the fancy hmac sha256 rubbish?? Interesting

                                Originally posted by Smelter
                                What is even harder than the handshake and connection is the serialization to rtmp.
                                Really? I didn't find that hard at all.

                                In flash, I'm getting keyframes only for the video now - so I guess I'm doing something right - boo-yah, video works - audio still has me beat though ...
                                "If you don't like DNAS, write your own damn system"

                                So I did

                                Comment

                                Working...
                                X