Announcement

Collapse
No announcement yet.

The JVC JY-HD10's Nasty Little Secrets

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

  • The JVC JY-HD10's Nasty Little Secrets

    Ok, guys, I thought I'd pass along a few observations that have been confirmed by others.

    You may remember I bought a JVC JY-HD10 HDV camcorder when I was still working for Ulead... in 2004.

    I used the camcorder to test the then-new Ulead HDV plug-in for the version 7 of MediaStudio Pro.

    The Ulead plug-in seemed to work okay.

    However, I did run into a very unexpected problem with the files captured from the JVC camcorder, which I have sold.

    I used MediaStudio Pro's NTSC DVD templates to down-convert the 720/30p 1280 x 720 HDV video to be compliant with 29.97 SD DVD standards.

    I still have native HDV files and the SD MPEG-2 files that resulted from the down-conversion.

    The big surprise comes when you burn these down-converted files to DVD and then play the DVD on an interlaced SD TV set.

    Whenever there is motion, there is an unbelievably annoying moire pattern that manifests itself on the interlaced display.

    I've down-converted the MPEG-4 files generated by the SANYO HD1 and I've burned them to a DVD and I think I'm seeing the same problem, but due to the nature of the scenes, and their lack of motion, it's more difficult to see.

    But the point is this:

    If you think you're going to be able to take a PROGRESSIVE output from an HD consumer format camcorder and then expect the down-converted video to look great on all types of TV screens, particularly SD interlaced screens, you may be in for a big, nasty surprise.

    The video produced by these camcorders looks fine on progressive displays.

    Jerry Jones

  • #2
    By the way, a fellow named Tom Roper made the same observation in May of 2004, which is when I bought my JVC JY-HD10.

    Here are his comments:



    Jerry Jones

    Comment


    • #3
      The motion artifacting problem is being reported all over the Web.

      Jerry Jones
      Last edited by Jerry Jones; 30 September 2006, 21:54.

      Comment


      • #4
        Jerry,

        What software are you using to do the downconverting? There is a pretty wide variation in the interpolation quality from one software to another. I have been using Vegas for about a year or so and have excellent results when down converting Z1 video to SD for TV display.

        Mark
        - Mark

        Core 2 Duo E6400 o/c 3.2GHz - Asus P5B Deluxe - 2048MB Corsair Twinx 6400C4 - ATI AIW X1900 - Seagate 7200.10 SATA 320GB primary - Western Digital SE16 SATA 320GB secondary - Samsung SATA Lightscribe DVD/CDRW- Midiland 4100 Speakers - Presonus Firepod - Dell FP2001 20" LCD - Windows XP Home

        Comment


        • #5
          Originally posted by Hulk
          Jerry,

          What software are you using to do the downconverting? There is a pretty wide variation in the interpolation quality from one software to another. I have been using Vegas for about a year or so and have excellent results when down converting Z1 video to SD for TV display.

          Mark
          Hi Mark,

          You have this model...

          Change your country or language settings


          ...right?

          So when you down-convert material from that camcorder, are you down-converting 720p?

          I don't see 720p listed in the specifications for this particular model of camcorder.

          The problem I'm seeing does *not* result from down-conversion of 1080i.

          The problem I'm describing results from down-conversion of 720p.

          I do not believe the problem results from use of the Ulead MPEG.Now codec.

          Why?

          Because another fellow posted similar observations to my own.

          He had JVC's consumer model... the GR-HD1...



          I had JVC's professional model... the JY-HD10...



          Both models are similar... with only minor differences... but they both record natively to 720p... not 1080i.

          The fellow who observed similar problems with down-converted 720p video made his observations on this Web page:



          His name is Tom Roper.

          He used Vegas and Womble and observed the same results I observed.

          Tom then sold his JVC GR-HD1.

          He now owns a Sony HVR-Z1U... identical to your model.

          He recently wrote me with the following:

          "The Z1U is better at downconverting because it doesn't have the unusual 2:2 30p cadence of progressive video that many DVD players have issues with."

          "60i interlace has the familiar 3:2 cadence that DVD players with pulldown handle better."

          I am, therefore, inclined to believe that 1080i is more easily down-converted to 480i for viewing on standard DVD players; 720p - on the other hand - apparently poses some challenges.

          My read of the specifications for your HVR-Z1U suggests your model can't record to 720p.

          So I'm not surprised that you haven't run into the issues that Tom and I encountered.

          Jerry Jones
          Last edited by Jerry Jones; 11 February 2006, 12:01.

          Comment


          • #6
            Jerry,

            I stand corrected. Sorry about that.

            Still I'm not quite sure why this is happening. It's not really that difficult to take a progressive frame and split it into two fields.

            I'm not sure where the 3:2 comes in unless you are shooting 24p.

            Can you upload a few seconds of a bad section of this video? I'd like to fool around with it.

            Mark
            - Mark

            Core 2 Duo E6400 o/c 3.2GHz - Asus P5B Deluxe - 2048MB Corsair Twinx 6400C4 - ATI AIW X1900 - Seagate 7200.10 SATA 320GB primary - Western Digital SE16 SATA 320GB secondary - Samsung SATA Lightscribe DVD/CDRW- Midiland 4100 Speakers - Presonus Firepod - Dell FP2001 20" LCD - Windows XP Home

            Comment


            • #7
              Yes, Mr. Roper's explanation regarding DVD players may be flawed.

              But I used the Ulead MPEG.Now encoder's three options... frame-based, upper field, lower field... to down-convert.

              Each of the three outputs... on DVD playback... showed the dreaded combing effect on a SD interlaced TV.

              I have sold the JVC JY-HD10.

              But I may still have a raw 1280 x 720 progressive frame clip on my hard disk.

              If I find it, I'll try to upload a small section to my Web site.

              If you can successfully burn it to DVD with Vegas, fine.

              But Mr. Roper had Vegas, too, and he failed.

              So either his workflow was flawed... or... it is more difficult than one might think to down-convert 720/30p progressive video to 480/30i.

              By the way, the only reason I got curious about doing this in the first place was to resolve the question of whether the JVC JY-HD10's standard DV mode could equal the quality of down-converted HD mode output.

              If you simply connect the camcorder to an interlaced TV using composite out cables, it doesn't matter which mode you select... 720/30p, 480/60p, 48030i... the picture looks the same because the camera does the down-converting internally.

              My curiosity had to do with whether there might be a difference between 720/30p computer files down-converted to 480/30i and the camcorders standard DV material.

              I do have a question for you, Mark.

              If you were working with 720/30p clips, wouldn't you want to specify "frame-based" for your editing template?

              That is what Ulead has done in connection with the HDV plug-in templates.

              The template that is specifically designed for the JVC JY-HD10 specifies FRAME-BASED.

              Now, when you down-convert this video using the Ulead MPEG.Now encoder, which option should you choose?

              I tried all three.

              But my first choice was to keep the FRAME-BASED selection.

              What would be your first choice?

              Would you choose FRAME-BASED?

              Would you choose FIELD A (LOWER FIELD)?

              Would you choose FIELD B (UPPER FIELD)?

              Jerry Jones

              Comment


              • #8
                As a rule when making DVD's I don't use the MPEG encoder in DVDWS2. I like to encode from the NLE. I feel I have a bit more control there and then I can verify that WS2 is not re-encoding the video. I realize now that you are talking about encoding from MSP.

                I would probably go with Field B. One of the things I like about Vegas is that it will not only tell me if a clip is frame or field, but it will also tell me field order.

                I have found that with Ulead using frame based you generally MUST check the anti-flicker tickbox to get acceptable results for TV viewing. With field based encoding I have found it to work either way (field A or B) depending on the source video. I was working with a school project one where they have me some analog footage and I had a terrible time with field order. I finally just used the anti-flicker setting and that produced acceptable results.

                I was always a bit confused by the various way one can mess with frame and field in MSP. You can do it from templates, by right-clicking on clips, and from the encoder settings. There might even be more I'm not remembering. I used to have it kind of worked out as to how they affected one another but have long since forgotten.

                - Mark
                - Mark

                Core 2 Duo E6400 o/c 3.2GHz - Asus P5B Deluxe - 2048MB Corsair Twinx 6400C4 - ATI AIW X1900 - Seagate 7200.10 SATA 320GB primary - Western Digital SE16 SATA 320GB secondary - Samsung SATA Lightscribe DVD/CDRW- Midiland 4100 Speakers - Presonus Firepod - Dell FP2001 20" LCD - Windows XP Home

                Comment


                • #9
                  Mark,

                  Do you have a snail mail address?

                  I'll burn the HDV file on my hard disk to a CD or DVD-R disc and mail it to you.

                  Jerry Jones
                  Last edited by Jerry Jones; 30 September 2006, 21:55.

                  Comment


                  • #10
                    Problem explained: http://tinyurl.com/2ojg4f

                    When shooting HD, the camera captures 30 progressive frames per second — half the temporal rate of 720p HD broadcasts. The NTT “SuperENC” MPEG-2 decoder/encoder chip is primarily responsible for the low frame rate. Some shooters will like the low rate because it is close to 24fps, thereby providing what they consider a “filmic” look. Others will dislike the look, as rapidly moving objects — or non-moving objects when one pans too quickly — appear as “double objects.” The name for this visual artifact is “eye tracking,” and it is generated within our eyes. The double images are not recorded to tape. Our eyes create the artifact from moving objects within a series of images where every frame is repeated — as it is when 720p30 is converted by the camcorder to 720p60 for display. (Just as when film is projected using a double-bladed shutter.) Although the artifact can't be eliminated, you can minimize it by locking the shutter-speed at 1/60 second — a speed equivalent to a film camera set to a 180-degree shutter. JVC recommends locking a 1/30 shutter speed that masks the artifact by creating so much motion blur — from the very slow shutter — that the two objects blur into one. While I prefer the former solution, my testing showed that any shutter speed from 1/30 to 1/60 second is equally acceptable. Another alternative is to shoot 480p60, native 16:9, SD video because it is free of eye-tracking artifacts. While image resolution is visibly lower, SD (like HD) is free of both interlace and NTSC artifacts.
                    I got screwed.

                    Jerry Jones

                    Comment

                    Working...
                    X