<html>
<head>
<style>
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Verdana
}
</style>
</head>
<body class='hmmessage'>
Throw away unreferenced b-frames, they're not worth correcting. <br>Otherwise, you should take this discussion to the ffmpeg mailing list, <br>this list isn't meant for decoding questions.<br><br>JB<br><br><br>武南龚 wrote:<br>> yes, there is frame-freeze technology<br>> I mean<br>> if B frame is lost<br>> I substitute it with the previous P frame, that is I P P b<br>> then how is the b frame estimated?<br>><br>><br>> On Fri, Aug 21, 2009 at 2:34 PM, Loren Merritt<lorenm@u.washington.edu> wrote:<br>> <br>>> On Fri, 21 Aug 2009, 武å~M~W�¾~Z wrote:<br>>><br>>> <br>>>> I have a question about error concealment. I used x264 to encode a<br>>>> video. And in the encoded video, assume there are some consecutive<br>>>> frames: I P B b .....<br>>>> If the P or B frame is lost, how to reconstruct(or estimate) the next<br>>>> frames?<br>>>> <br>>> If frames are lost, you're screwed. There is no way to reconstruct anything<br>>> worthwhile. If you want an algorithm that may or may not look slightly<br>>> better than discarding everything until the next keyframe, you could assume<br>>> that the lost frames are identical the preceding valid frame.<br>>><br>>> --Loren Merritt<br>>> _______________________________________________<br>>> x264-devel mailing list<br>>> x264-devel@videolan.org<br>>> http://mailman.videolan.org/listinfo/x264-devel<br>>><br>>><br>>> <br>> _______________________________________________<br>> x264-devel mailing list<br>> x264-devel@videolan.org<br>> http://mailman.videolan.org/listinfo/x264-devel<br>> <br><br><br><br /><hr />Hotmail® is up to 70% faster. Now good news travels really fast. <a href='http://windowslive.com/online/hotmail?ocid=PID23391::T:WLMTAGL:ON:WL:en-US:WM_HYGN_faster:082009' target='_new'>Try it now.</a></body>
</html>