VideoHelp Forum




+ Reply to Thread
Results 1 to 13 of 13
  1. hello there, i've recently posted questions reguarding scripts i was trying to use and even with suggestions nothing looks right. There always a bunch of grain/pixelation weither i use denoisers or not. I'm trying to encode the OC using 23 fps and 1200 bitrate. Could the all-wise experts suggest an ideal script, if they don't mind giving up their ancient chinese secrets. Thanks in advance.
    Quote Quote  
  2. Well, at least you've learned enough now to IVTC them back to 23.976fps. Did you bother to go back to Doom9 to thank Boulder for his advice, and me over here for my advice, in regards to IVTC versus ruining them with a deinterlace?

    Have you run a compression test through GKnot to get an idea of the final quality before encoding? Everything compresses differently, and all sources are different. No advice can be given without a source vob available. Do you have a place where you can upload a piece of the vob showing a section where the XviD has problems?
    Quote Quote  
  3. Yes i believe i thanked everyone for help with my ignorance but if i didn't i thank you, i definitly appreciate it. Not being able to figure something out is frustrating as hell and abeing able to come somewhere to ask questions really helps. I haven't tried the compression test yet but will. And i don't know of anywhere to upload.
    Quote Quote  
  4. 1200 kbps may not be enough bitrate for some scenes.

    If you don't need a file of a specific size, use Xvid in single pass Quantization mode. Pick a Q value between 2 and 4.

    If you need a file of a specific size use a variable bitrate 2-pass encoding.
    Quote Quote  
  5. Good point, junkmalle. I hadn't even thought of that. You aren't running a single pass at 1200 CBR are you, 12 gage? It's either 2-pass VBR or a single pass at a fixed quant (I wouldn't go over 3 myself) for best results.

    Another thing you might try is to run it through AutoGK, maybe a fast 1-pass Quality Based encode at the default 75%, and then study the log afterwards to see what you can learn. It has a very good analyze technique that determines if it can be Force FILM'd, has to be IVTC'd, or can only be deinterlaced. And it may do a better job by itself than you've been able to do so far with GKnot.
    Quote Quote  
  6. thanks for the response, i am using 2-pass and i'll try the AutoGK. I always see the program but it says spyware next to it and shyed away.
    Quote Quote  
  7. I still recommend you switch to Quantization (constant quality) mode. You can encode a short segment, see if you're happy with the quality. Then encode the whole video. After you've done this a few times you'll know what setting you want to use in the future. Why waste time with 2-pass encoding?

    At Q=2 it's hard to see the difference between any source and the Xvid file, even with enlarged still frames. At 3 you can see a little macroblocking if you enlarge still frames but they're not noticable at normal playback speed. At 4 you start seeing macroblocks on normal playback if you look very closely. I use 2 for high quality stuff, 3 for less important videos. You can also use decimal values like 2.5.

    The only time to use 2-pass encoding is when you need a file of a certain size -- say you want to fit a 90 minute movie on a 700 MB CD, or 2 hour long TV episodes on a CD.
    Quote Quote  
  8. I do need it a certain size, but if i use gordian knot to fix the bitrate will the Q=2 method still be effective? also is it the lower number better quality?
    Like 1 would be the best quality?
    Quote Quote  
  9. sorry i saw on your previous post you said Q=2-4. I dumm
    Quote Quote  
  10. The quantization value won't be used if you're using 2-pass VBR. There are

    Yes, 1 is the highest quality. But the resulting files will be very big and there won't much difference between 1 and 2.
    Quote Quote  
  11. can i use the i pass quantization method and still determine how big the file will be?
    Quote Quote  
  12. Originally Posted by 12 gage
    can i use the i pass quantization method and still determine how big the file will be?
    Not really. Quantization mode gives you constant quality -- every frame will have the picture quality you asked for, regardless of how much bitrate it takes. So you can't tell beforehand exactly how big the file will be.

    You can watch the estimated file size while encoding to get an idea. But it can be misleading. High action and noise will eat up lots of bitrate. Dark and static scenes compress better. So a film that starts out with dark static scenes and ends in bright action scenes will become much larger than the early estimates. And vice versa.
    Quote Quote  
  13. Thanks for the info, now i have a few things to play around with, much appreicated
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!