Announcement

Collapse
No announcement yet.

G400 Capture Query

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • G400 Capture Query

    Hello Everybody.

    I am thinking about buying a G400 but I'm a little concerned about the "fuziness" that is constantly mentioned when compared to similar equipment such as Pinnacles' PC-TV Pro and Hauppages' Win TV.

    Here is my question:

    Does encoding with the HuffYUV codec or capturing in uncompressed eliminate this problem? (meaning that it is due to the MJPEG codec,) or doesn't it make any difference (which would mean the problem is due to the driver).

    I am aware that the HuffYUV codec will produce higher quality captures than the built in Zoran MJPEG codec, but all I really need to know is whether there is a noticable difference in the sharpness between the various capture routines.

    Any replies would be greatly appreciated.
    Thanks.

    Ian.

    [This message has been edited by White Thunder (edited 22 June 2000).]

    [This message has been edited by White Thunder (edited 22 June 2000).]

  • #2
    Fuzziness? Hmm.... Mine is actually quite good, especially when sent out through the S-Video connection.

    Dr. Mordrid

    Comment


    • #3
      If you line the antenna-cable from your VCR to the G400 and output it directly to the TV, and you put a scart cable from the VCR directly to the TV. Then you can easyly compare the quality of G400 (uncompressed, not MJPEG) to normal TV quality by simply shifting channels. Then you will see that the G400 is not even on par, the quality is much worse than plain TV. (I use component video and matrox PCVCR, maybe quality is better with SVHS output. If someone could do this test with G400 and SVHS output I would be interested in the result)

      I don't know how Pinnacles' PC-TV Pro and Hauppages' Win TV perform but according to rumors I've heard that they give equal or even better quality than plain TV in this test.., but, as I said, I'm not sure.

      What I'm sure of is that HuffYUV is much better than MJPEG and as long as I'm only capturing video for converting into DivX 352*288 the quality is good enough for me.

      But if I would have made the choise today, I think that I would have bought a PCTV pro + DC10+ combo. (plus a decent 3D card)

      Comment


      • #4
        I guess "fuziness" isn't the exact word I was after. I think "blurriness/ blurred look" would be more accurate.

        To get a general idea of what I mean, take a look at Intomisery's "Codecs & Marvel's less than pro quality" post.

        Thanks again.

        Comment


        • #5
          I think the fuzziness is inherent in the tuner of the Marvels. I have the G200 and I'm assuming the G400 has basically the same capture hardware. When I compare the picture quality on screen (not the quality of the capture) of the G200 to that of my ATI TV Wonder add-in TV Tuner card, there is no contest. The ATI TV Wonder provides superior color and clarity to the G200. The fuzziness and lack of correct color transfers to the captures no matter which codec you use. Using PICVideo I get much better results on my TV Wonder.

          Comment


          • #6
            The quality of the captured video is excelent but video-out (both composite and S-video) quality is rather bad. It is indeed blurred. What I find most annoying is that the video-out gets suddenly sharper when you start recording. It seems the VGA overlay and video-out are connected together (using the same circuit or buffer). The video-out is therefore deinterlaced (blurred) together with the overlay. Blurring video-out picture makes absolutely no sense. Matrox should kick out the deinterlacer completely as it only makes everything look worse.

            Comment


            • #7
              The more I hear this about blurry output the more I think there is another cause. It certainly does not sound ANYTHING like my Marvel G400-TV's output. Not even close. And I have two of them.

              In fact it's my recollection that most of the early users who switched from a Marvel G200 noted a marked increase in output quality.

              Just out of curiosity, how about a poll stating if you see "fuzziness" and listing;

              PAL, NTSC or SECAM

              Monitor type (composite, Y/C, TV w/vidin or ?)

              Monitor cable type

              Analog cable/Digital Cable/Antenna

              I'll start:

              No
              NTSC
              Y/C monitor
              S-Video to RCA Y/C adapter (circuit available)
              Digital cable

              Dr. Mordrid



              [This message has been edited by Dr Mordrid (edited 23 June 2000).]

              Comment


              • #8
                Thanks for the reference, WhiteThunder..
                Mordrid: it's not exactly fuzziness it's just..
                let's put it this way. if I take a piece of video (I've actually tried this with a VHS copy of The Matrix) and I use either S-video or RCA cables (2 seperate VCRS, one a newish Pansonic S-VHS deck, the other a cheap RCA 2 head VCR) in, record a clip, then output that clip and compare the two, the quality is not identical to the original. Something is lost- be it colour, clarity, whatever. It looks like it's been run through a computer. For myself, going from editing on a $35,000 Media100 system to this is a huge downgrade (of course). All I want is exact quality reproduction of the original. Is that too much to ask from a $400 card? perhaps. Will I try my hardest to achieve this lossless reproduction while complaining at the same time? you betcha!


                I have no idea what Y/C, etc are, but I'll run down my specs as best I can.

                YES (big yes)
                NTSC
                Monitor: TV using coax cable connection from VCR
                VCR out into Marvel B.O.B. in: S-video/RCA
                Marvel B.O.B. out to VCR in: S-video/RCA
                (when I use S-video to input, I use S-video to output and vice versa)

                P.S. Did notice a big difference when I used S-Video and a good VTR as opposed to my regular crappy VCR which has only RCA in (I borrowed the S-VHS VTR from my school to test out the quality), yet still not "lossless"

                [This message has been edited by intomisery (edited 24 June 2000).]

                Comment


                • #9
                  ok, I've got it!
                  The "fuzziness" sort of looks like movement. it looks like there are 2 pixels that keep jumping back and forth or switching colours or something. this happens all over the screen.. .like if I look at the edge of letterboxing, the edge isn't clearly defined by 1 pixel width all the way horizontally across that portion of the screen, it's definied by 2 or 3 that keep jumping between different colour blacks- it's exactly like anti-aliasing in photoshop or whatever.. From far back it looks ok, your eye blends it together and it creates a smooth line. but closer up you can see the movement, the different shades of black..

                  [This message has been edited by intomisery (edited 24 June 2000).]

                  Comment


                  • #10
                    I don't know about your system, but on my G200 the output is representative of the captured video.

                    It is the overlayed video that is poor in quality. Video through the S-Video produces far better results than with composite or the TV tuner, however the color is still a little lacking even with adjustments and calibration with MSP6.

                    The overall picture is just not as sharp and lacks the color vibrancy when compared to my ATI TV Wonder. No matter how much adjustments are made, the quality does not approach the default settings of the ATI TV Wonder.

                    This is being viewed on my 17" Sony Trinitron Multiscan 220GS which is a very good monitor, though not as good as Sony's current completely flatscreen offerings.

                    I think most people don't realize the lack of quality because they don't compare the Matrox's picture to anything else. When you immediately switch back and forth between the two, the difference is plainly obvious. For me, I was never satisfied with the G200's video input quality from day one.

                    I don't think it's anything wrong with my Marvel G200 because I've had both an AGP and a PCI version and they both were identical in video quality.

                    Also, numerous other postings on the Usenet and here confirm the "soft" picture and lack of color vibrancy. Others I know who purchased Matrox Marvels G200 or G400 gave up on video editing on the computer altogether because they were disappointed in the quality of the Marvel products.

                    Also, the Matrox Marvel G200 and G400 products seem to have almost a cult following where people downplay, ignore, deny or blame the user for the inherent problems these cards have.

                    Comment


                    • #11
                      Viewing video on the computer, full screen, is a bit weird. there's the same sort of problem I talked about earlier, the unclearness of the picture and the anti-aliasing/movement... but more of it. This is using the MJPEG Codec. I have however tried only 320x240 YUV and the quality looks a lot better.. especially when using s-video cables.

                      I've heard that the ATI All-In-Wonder was shite compared to the marvel, that's why I bought it instead...

                      Comment


                      • #12
                        Question: are these moving pixels after editing or in the raw capture?

                        Dr. Mordrid

                        Comment


                        • #13
                          They're moving in the raw capture.
                          I've set something up that might help.. a bit. They're stills from both YUY2 and MJPEG captures.
                          unless I send you a tape of captured footage, this is as best it gets to showing you..
                          http://www.geocities.com/deluminite/...on/compare.htm

                          [This message has been edited by intomisery (edited 24 June 2000).]

                          [This message has been edited by intomisery (edited 24 June 2000).]

                          Comment


                          • #14
                            actually, I havent noticed too much of a diff. between MJPEG and YUY
                            when I output my MJPEG back to TV it looks pretty close to the orig. (in w98) In W2k, it looks like some sort of interlacing problem with fuzziness.
                            I was curious about something. what is the difference in quality between capturing via composite and TV tuner

                            Comment


                            • #15
                              Updated the page with comparison pictures.
                              http://www.geocities.com/deluminite/...on/compare.htm

                              Comment

                              Working...
                              X