2pass encode, the bitrate set in first pass will affect the output ??

[solution]

Hello everyone~

I'm now using Mpeg Mediator with divx 5.05 to encode some video. Now i use 2pass as encoding method, but i wanna know, in the first-pass step( i set it in divx, which is "2-pass, the first pass" one, not original one), will the bitrate affect the output log file generated so that have different analysis of the screen??

and also, later when i do the 2nd-pass encode, how do the "bitrate" and "max bitrate" affect the output avi file??

and, the last question, what's the different between "original 2-pass" and the new 2-pass encoding method?

thx~

[zx50]

In multi pass first pass What you call first pass) you specify a bitrate to analyse the video with, say if you have 950kbps then it will look through the video to see how it can make the video file look good with that amount of bit's. But if you have a higher bitrate then it makes it easier for the codec to improve the quality of the video because it is not so restricted with the bitrate, The lower you go with the bitrate the harder the codec has to try to figure out where to put the bitrate. I thought i would first explain it yo you but an answer to your first question is yes, it will improve the analysis of the video.<br><br>secondly, what's the difference between multi pass and original one pass. Well the difference is original one pass does'nt analyse the video like multi pass does, so therefore can't do as good of a job as multi pass can.

[solution]

Thx for your help zx50~

but sorry that, maybe my description not so well. The second question that i asked, in fact, i mean the "Original two-pass" mode in divx5.05. In divx5.05, there are "Multi pass, 1st/nth pass" mode and "original 2-pass, 1st/2nd pass" mode for user to choose, so what are their difference??

and also, my next another question asked before, at the "multi pass, nth-pass" mode, there is a "max bitrate" appear just above the main bitrate slider, what do this max-bitrate relate with the main-bitrate setting?

thx!!

[zx50]

Well yeah that's what i meant (should have put it right myself) when i said original one pass i did'nt actually mean just the one pass i actually meant the original one pass first pass. Secondly in the nth pass, the bitrate what you specify (next to the slider) is the average bitrate, of course the codec can go under this bitrate if needed but this is what the codec will normally use, and as for the max bitrate in the box, when the codec comes across scenes that need a lot of bitrate the codec can jump up to what it hinks the scene needs, but it can't go over what you have in the box (to tell you the truth it might even get nowhere near) but anyway that's what max bitrate means.

P.S just to let you know again the original one pass first pass does'nt analyse the film whereas multi pass first pass does. I done a film in original one pass first pass and then done a film in multi pass and i can tell you that the quality of the multi pass was much better.

[solution]

OHHH~~ really thanks zx50!!!
I'm clearly understand what these setting mean now!!

but i still got some little questions. I wanna know that, how to judge, what bitrate should i use to analysis the video in the first-pass then??

and, i found that, in the first-pass, i have checked the box "write MV file > c:\mvinfo.bin" but i found that there is no such file write out during or even finish the process, is it a divx5.05 bug? Do this BIN file important for the nth-pass encoding? Since i have try MPEGMediator and FlaskMPEG, both of them no such BIN file output during first-pass process...

and the last question, in the nth-pass, there appear a checkbox "update divx.log", is it that if the checkbox is checked, there will do analysis again, but, write out the movie without check the box??

thanks again anyway!!

[zx50]

Firstly, yes i think it's a bug of some sort in the codec because it did'nt write an mvinfo bin file when i have enabled it (when i used the 5.0.5 that is). And update log file means that when checked it sort of updates the quality if you like. If you don't have update log file enabled then you might as well be doing a 1 pass because the quality will be just the same as the first pass, whereas if you have update log file checked then the log file gets updated therefore meaning that the quality will be better by shifting the bitrate around.

[solution]

oh.... that's the bug?? so poor design 5.05...

so zx50, what do that BIN file exactly used to? and now there is no BIN file generated in my divx5.05, will it affect the nth-pass process results??

and, in the last question, how to judge the bitrate in first-pass step??

the last question, i still not clear about the "update log file" stage...
do u mean that, if i check the "write log" in the first-pass, and then no-check the "update log" in 2nd pass, the results movie will be just the same as 1-pass movie?? but ... in my idea, there already do 1 analysis, and then use the analysis result to encode the movie, don't this already a complete 2-pass encoding process? so if i do the "update log" in the 2nd-pass, the movie come out at the 3rd-pass stage, don't this called 3-pass encoding then???

[zx50]

The mvinfo bin file is to track the motion in the film on the first pass and then use that info on the second pass meaning that the second pass will be quicker (because it does'nt have to track motion, because it has all the motion in the mvinfo bin file) and about calculating the bitrate, well i'm afraid that's the only question that i can't answer for you because i have'nt a clue how to calculate that (i let gordian knot do that for me) and lastly why do people just do 2 passes when it comes to the nth pass method ?? i don't know beats me, i always do three as well, because that's the way i see it. If the log file gets updated in the second pass then it's going to show the results in the third pass.

[zx50]

Btw no it won't affect the results if there's no mvinfo bin file produced, like i said just to track the motion that's all. Basically just to save time on the nth passes that's all.