Maximum Quality Method!!!!????

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • I26
    Gold Member
    Gold Member
    • Jun 2002
    • 106

    Maximum Quality Method!!!!????

    I have been using Flask since day one and now i want to step it up a notch. What is the best method/program yielding the best result in overall quality? Thanx in advance!!!

    Homegrown Desktop:
    P4 2.4 @ 2.7
    ATI 9700pro @ 419.4/730.8
    3dMark01--17,189
    Air Cooled System Temps @ Benchtime: 2c--MB / 4c--CPU

    Dell 8600 Inspiron Laptop:
    Pentium M 1.4GHz
    NVidiaGF Go5650
    3dMark01--9,842
  • UncasMS
    Super Moderator
    • Nov 2001
    • 9047

    #2
    you used flask all these years and didnt find alternative tools reading this forum???

    loss-for-words

    Comment

    • setarip
      Retired
      • Dec 2001
      • 24955

      #3
      A simple suggestion, if i may (before you're deluged with suggestions from the fervent supporters of Gordion Knot) - before seeking other conversion programs, try using the recently released version of the cross-platform (PC and Mac) 3IVX codec in conjunction with good ol' reliable Flask. Quality settings in the range of 75% to 100% yield excellent single pass results...

      Comment

      • UncasMS
        Super Moderator
        • Nov 2001
        • 9047

        #4
        .... and unpredictable file size

        Comment

        • I26
          Gold Member
          Gold Member
          • Jun 2002
          • 106

          #5
          Originally posted by UncasMS
          you used flask all these years and didnt find alternative tools reading this forum???

          loss-for-words
          Wait....let me rephrase. I used many programs with mixed results. I seem to like Flask the best. What I am wondering is if anyone used Flask but found something way way better worth mentioning.

          Homegrown Desktop:
          P4 2.4 @ 2.7
          ATI 9700pro @ 419.4/730.8
          3dMark01--17,189
          Air Cooled System Temps @ Benchtime: 2c--MB / 4c--CPU

          Dell 8600 Inspiron Laptop:
          Pentium M 1.4GHz
          NVidiaGF Go5650
          3dMark01--9,842

          Comment

          • I26
            Gold Member
            Gold Member
            • Jun 2002
            • 106

            #6
            Originally posted by setarip
            A simple suggestion, if i may (before you're deluged with suggestions from the fervent supporters of Gordion Knot) - before seeking other conversion programs, try using the recently released version of the cross-platform (PC and Mac) 3IVX codec in conjunction with good ol' reliable Flask. Quality settings in the range of 75% to 100% yield excellent single pass results...
            Does that 3IVX setup like DivX 5.x codec? Don't worry I am not going to ask step by step but I am wondering and i will give it a whirl. So if single pass rocks how about 2 pass encode?

            Homegrown Desktop:
            P4 2.4 @ 2.7
            ATI 9700pro @ 419.4/730.8
            3dMark01--17,189
            Air Cooled System Temps @ Benchtime: 2c--MB / 4c--CPU

            Dell 8600 Inspiron Laptop:
            Pentium M 1.4GHz
            NVidiaGF Go5650
            3dMark01--9,842

            Comment

            • UncasMS
              Super Moderator
              • Nov 2001
              • 9047

              #7
              encoding video with virualdb using avisynth to frameserve is the best you could do.

              it doesnt realy matter if you use gordianknot as a gui for setting all needed switches, values, settings or do it manually in virtualdub or maybe use a tool like DVX

              vd is most reliable in combination with avisynth


              sound should be handled separately with besweet or any other way.

              the final video and audio stream(s) will be muxed (merged) together later on with virtualdubmod or nandub.

              Comment

              • setarip
                Retired
                • Dec 2001
                • 24955

                #8
                "Does that 3IVX setup like DivX 5.x codec?"

                The .3IVX codec is an ALTERNATIVE to the DivX (and Xvid) codecs...


                "So if single pass rocks how about 2 pass encode?"

                No need - you can choose from Constant Bit Rate quality or Variable Bit Rate quality - essentially the same concept as choosing between single pass and two pass processing...

                Comment

                • Enchanter
                  Old member
                  • Feb 2002
                  • 5417

                  #9
                  Quality settings in the range of 75% to 100% yield excellent single pass results...
                  Does not look any different from the standard quality settings available within the DivX 5 and XviD codecs. . .

                  In addition to that, I would not recommend the use of Flask for a number of reasons. One of them, which is the most crippling to Flask's quality, is the inferior resizing algorithm. VirtualDub's own resizer or Avisynth's built-in resizer (esp. for Lanczos) give more accurate results.

                  Comment

                  • I26
                    Gold Member
                    Gold Member
                    • Jun 2002
                    • 106

                    #10
                    Well i tried out 3ivX and I used constant quality at QP5 = 85%
                    The file was 983MB compared to the DivX 5 - 2 pass encode of 1360MB. I meant for it to be 2 disk. The question is this....what is better in 3ivX....Average Bitrate or Constant Quality? I see under avg. bitrate it has upto 500. Seems kinda low to me??!?!?!?!?

                    Homegrown Desktop:
                    P4 2.4 @ 2.7
                    ATI 9700pro @ 419.4/730.8
                    3dMark01--17,189
                    Air Cooled System Temps @ Benchtime: 2c--MB / 4c--CPU

                    Dell 8600 Inspiron Laptop:
                    Pentium M 1.4GHz
                    NVidiaGF Go5650
                    3dMark01--9,842

                    Comment

                    • Enchanter
                      Old member
                      • Feb 2002
                      • 5417

                      #11
                      Originally posted by I26
                      Well i tried out 3ivX and I used constant quality at QP5 = 85%
                      The file was 983MB compared to the DivX 5 - 2 pass encode of 1360MB. I meant for it to be 2 disk. The question is this....what is better in 3ivX....Average Bitrate or Constant Quality? I see under avg. bitrate it has upto 500. Seems kinda low to me??!?!?!?!?
                      It is obvious that quality settings are never meant for precise filesize predictions (this is not an opinion of mine, rather a fact to be aware of). Quality settings at, say, 75% means that the encoder will compress every frame in a given video stream at 75% quality. Given that each frame compresses differently, the bitrate distribution will be variable (if it is constant, start getting worried). IMO (and others, I'm sure), quality settings should ONLY be used for archiving purposes when filesize does not matter. Using it for DVD-2-CD conversion isn't a good idea.

                      On the other hand, 2-pass encoding will create a bitrate graph that will be used for determining how the bitrate SHOULD be distributed in the second pass. This results in more accurate usage of bits and allows for precise filesize prediction (by either setting the bitrate or desired AVI filesize, rather than using an ambiguous value of x% quality).
                      Last edited by Enchanter; 25 Jan 2003, 06:57 AM.

                      Comment

                      • setarip
                        Retired
                        • Dec 2001
                        • 24955

                        #12
                        "The question is this....what is better in 3ivX....Average Bitrate or Constant Quality?"

                        I personally always opt for "Constant Quality" (similar to two-pass) because I'm more than happy to use two inexpensive CDs for consistently excellent output throughout the video, rather than sacrificing this for the sake of fitting the video onto one CD.

                        As a consequence, you sacrifice the ability to be CERTAIN, in advance, of the final filesize. In your instance, I would ask if you are totally satified with the quality of your 983Meg video. If so, split it and burn to two CDs - just as you intended to do. If you think it could be still better quality, I'd suggest that you simply re-do the conversion afresh, using 95% "Constant Quality (since you know you've got approximately 400Megs of additional CD space with which to "play").

                        Despite the apparent shortcoming of the lack of predetermination of filesize (using "Constant Quality), after several conversions you'll develop a "feel', based on the amount of action, etc., as to the highest percentage level you can use.


                        I'm curious to know your opinion of the quality of the generated output...

                        Comment

                        • I26
                          Gold Member
                          Gold Member
                          • Jun 2002
                          • 106

                          #13
                          Well the first movie I did with the outcome at 983MB for a 2.5hour movie was as good as if not a tad bit better than flask 2 pass with divX5. The outcome divx5 file was 1.36GB. Big difference in size but not quality. I say here 3ivX takes the prize.
                          I just did another movie with QP5=85% and the movie is not that good of quality. Getting some of that blurryness around the people and such. Its a 2hour movie and the outcome file size was 626MB which isn't bad. I may either try to up the quality and use the 3ivX again or do a 2 pass flask/divX5 on it. What do you think would be a good idea? Anything over 2 hours I am ok with using 2 discs I suppose....that is alot to cram on one cd.

                          Homegrown Desktop:
                          P4 2.4 @ 2.7
                          ATI 9700pro @ 419.4/730.8
                          3dMark01--17,189
                          Air Cooled System Temps @ Benchtime: 2c--MB / 4c--CPU

                          Dell 8600 Inspiron Laptop:
                          Pentium M 1.4GHz
                          NVidiaGF Go5650
                          3dMark01--9,842

                          Comment

                          • khp
                            The Other
                            • Nov 2001
                            • 2161

                            #14
                            I'd like to point out that one of the differences between quality based encoding and bitrate controlled encoding (both 1 pass and 2 pass), is that quality based encoding encodes all frames at the same quality, while bitrate controlled encoding will encode low motion scenes at a much higher quality than high motion scenes. This means that, how you compare the encodings may make a huge difference.

                            The reason bitrate controlled encoding, uses higher quality for low motion scenes than high motion scenes, is that any quality problems are much more noticable durring low motion scenes than high motion.

                            With this in mind I can't recommend using quality based encoding, for anything but very short clips testing, and for 100% quality rips (if you absolutly don't care about filesize).
                            Last edited by khp; 25 Jan 2003, 12:31 PM.
                            Donate your idle CPU time for something usefull.
                            http://folding.stanford.edu/

                            Comment

                            • I26
                              Gold Member
                              Gold Member
                              • Jun 2002
                              • 106

                              #15
                              Originally posted by khp
                              I'd like to point out that one of the differences between quality based encoding and bitrate controlled encoding (both 1 pass and 2 pass), is that quality based encoding encodes all frames at the same quality, while bitrate controlled encoding will encode low motion scenes at a much higher quality than high motion scenes. This means that, how you compare the encodings may make a huge difference.

                              The reason bitrate controlled encoding, uses higher quality for low motion scenes than high motion scenes, is that any quality problems are much more noticable durring low motion scenes than high motion.

                              With this in mind I can't recommend using quality based encoding, for anything but very short clips and testing.
                              ok.....point made, but what do you recommend? So far Flask 2 pass is still my choice. I still am tinkering with 3ivX.

                              Homegrown Desktop:
                              P4 2.4 @ 2.7
                              ATI 9700pro @ 419.4/730.8
                              3dMark01--17,189
                              Air Cooled System Temps @ Benchtime: 2c--MB / 4c--CPU

                              Dell 8600 Inspiron Laptop:
                              Pentium M 1.4GHz
                              NVidiaGF Go5650
                              3dMark01--9,842

                              Comment

                              Working...