<div>Hi, Shaaban </div>
<div> I haven't caught what you said.</div>
<div>First, you want to transcode mpeg2 bitstram to h.264 bitstream.<br>Second, the paper introduced an method which can reduce bit-rate of mpeg2 bitstream in frequency-domain,<br> In the paper, the author just changed QP to do bit-rate reduction and didn't use rate control.
<br>Third, you hope to see the effect of requantization error.<br> Bitrate reduction will cause quality degrade which can be expressed with PSNR.<br> You mean you want to do bit-reduction with rate control?<br>Best Regards.
<br>Liwei Song <br><br> </div>
<div><span class="gmail_quote">On 11/17/06, <b class="gmail_sendername">Lists administration</b> <<a href="mailto:videolan@lists.videolan.org">videolan@lists.videolan.org</a>> wrote:</span>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">----- Forwarded message from Mohsen Shaaban <<a href="mailto:mmm5554@cacs.louisiana.edu">mmm5554@cacs.louisiana.edu
</a>> -----<br><br>From: Mohsen Shaaban <<a href="mailto:mmm5554@cacs.louisiana.edu">mmm5554@cacs.louisiana.edu</a>><br>To: <a href="mailto:x264-devel@videolan.org">x264-devel@videolan.org</a><br>Subject: RE: [x264-devel] Re: Using I-frame instead of IDR
<br>Date: Thu, 16 Nov 2006 12:51:31 -0600<br>X-Mailer: Microsoft Office Outlook 11<br><br><br>The deleted attachment is at:<br> <<a href="http://lists.videolan.org/attachs/20061117/frequency-domain.pdf">http://lists.videolan.org/attachs/20061117/frequency-domain.pdf
</a>><br><br><br>Hello Loren,<br><br>First, I'm a member of x264-devel but I do not know how to post questions<br>from within the x264-devel mailing list. Can you please tell me how to do<br>that?<br><br>Second, I have a question about
H.264 transcoding from MPEG-2 bitstream, I<br>hope you can help me.<br><br>I want to do some experiments on the MPEG-2 bitstream before Transcoding, so<br>I want to see the effect of requantization error on the stream and I read in
<br>one paper that you can do that by having a fixed triplet quantization<br>scaling (fixed scaling value Qi, Qp, Qb) and my question is there any<br>feature in MPEG-2 TM5 reference software that you can change to do that or
<br>one have to change the rate control algorithm in there?<br><br>I attaching the paper for your reference (experiment shows in fig. 3 in the<br>paper)<br><br>I hope you have an answer or a guide for me to follow.<br>I appreciate it and thank you.
<br><br>Mohsen Shaaban<br>PhD Student in Computer Engineering<br>The Center for Advanced Computer Studies<br>University of Louisiana at Lafayette<br><br>-----Original Message-----<br>From: <a href="mailto:x264-devel-bounce@videolan.org">
x264-devel-bounce@videolan.org</a> [mailto:<a href="mailto:x264-devel-bounce@videolan.org">x264-devel-bounce@videolan.org</a>]<br>On Behalf Of Loren Merritt<br>Sent: Wednesday, November 15, 2006 10:42 PM<br>To: <a href="mailto:x264-devel@videolan.org">
x264-devel@videolan.org</a><br>Subject: [x264-devel] Re: Using I-frame instead of IDR<br><br>On Wed, 15 Nov 2006, Axel Gunter wrote:<br>> Loren Merritt wrote:<br>>> On Wed, 15 Nov 2006, Axel Gunter wrote:<br>>>
<br>>>> I noticed all the I-frames introduced by the x264 code were IDR Access<br>>>> Units. Is there a way I can have non-IDR I-frames in x264 encoding?<br>>><br>>> No. Why do you want non-IDR I-frames? The only reason I can think of is
<br>>> open-GOP B-frames, which x264 also doesn't do.<br>><br>> I guess the compression would be better if the frames following the IDR or<br><br>> I-picture could use frames temporally before the IDR/I-picture since they
<br>> have more reference frames to chose from. Since IDR forces a decoder<br>> reference buffer flush, this isn't possible. Correct me if I am wrong.<br><br>In that case, look at --min-keyint. If a scenecut is too close to the
<br>previous IDR (default: 25 frames) it gets encoded as a non-IDR I-frame.<br>Note that this is not optimal, but the metric that would really optimize<br>for compression ratio isn't possible to measure in 1 pass, and I haven't
<br>written a 2 pass scenecut/IDR algorithm yet.<br><br>--Loren Merritt<br><br>--<br>This is the x264-devel mailing-list<br>To unsubscribe, go to: <a href="http://developers.videolan.org/lists.html">http://developers.videolan.org/lists.html
</a><br><br><br>----- End forwarded message -----<br><br>--<br>Lists administration <<a href="mailto:videolan@lists.videolan.org">videolan@lists.videolan.org</a>><br><br>--<br>This is the x264-devel mailing-list<br>
To unsubscribe, go to: <a href="http://developers.videolan.org/lists.html">http://developers.videolan.org/lists.html</a><br><br></blockquote></div><br>