Creative Communities of the World Forums

The peer to peer support community for media production professionals.

Activity Forums Adobe After Effects please critique my capture/editing process

  • please critique my capture/editing process

    Posted by Graham Wellington on May 7, 2007 at 12:49 am

    1. Capture camcorder DV in VirtualDub as “direct stream”
    2. Edit > Save > Close
    3. Open file with AviSynth – SeperateFields
    4. add filter “Deinterlace Smooth” (by Gunthar Thallin)
    5. Save as .avi, Compressed with MainConcept DV codec
    6. Open in After Effects
    7. Add various effects
    8. Render .AVI using XVID for compression

    Is there anything I could change here to have better quality? I know the interlace smooth makes a double sized file but for motion there seems to be no better.

    Kevin Camp replied 19 years, 1 month ago 2 Members · 3 Replies
  • 3 Replies
  • Kevin Camp

    May 7, 2007 at 2:51 pm

    i would bring the dv straight to ae… in ae i waould check the interpret footage settings (selec the footage in thte project pane, then file>interpret footage>main…). make sure the separate fields (usually lower-ntsc), frame rate (usually 29.97-ntsc) and pixel aspect (usually .9-ntsc) are set correctly — ae usually gets the settings right, but its good to check. if you are going to render out as progressive, also check the preserve edges button, if you are going to render out back to fields you can leave that unchecked.

    now when you effect it in a comp, it will use progressive frames (de-interlaced). in fact you can set the comp frame rate to double the footage frame rate (usually 59.94) and see the de-interlaced upper and lower fields. or, use frame blending to composite the “missing” data from the other field — this can help smooth some motion, but can create problems too, it just depends on the movement.

    if you are going to render back out to fields, you may want to use the double frame rate method, and then override the the frame rate in the render settings, setting it back to the original footage’s frame rate and enabling fields (usually lower).

    if you are going to render progressive, then the frame blending may help, or it may be unnecessary, you’ll be able to see in the ram preview if you like it or not.

    by-passing your extra de-interlacing step will save you time, disk space and data/quality loss of compressing you footage on more time. if you do think you get better results de-interlacing externally, use a lossless or uncompressed option, rather than dv.

    Kevin Camp
    Designer – KCPQ, KMYQ & KRCW

  • Graham Wellington

    May 7, 2007 at 6:45 pm

    In virtualdub there is a setting “uncompressed RGB/YCbCr”. Is this lossless? A 1:30 minute clip comes out over 5 gigs, but i can always delete it afterwards.

    After rendering as XVID it seemed to be darker, so I had to brighten by 5. Does that uncompressed format change the colors? I thought it was copying frame by frame exactly what is on the screen.

  • Kevin Camp

    May 7, 2007 at 8:02 pm

    the uncompressed setting should be the best quality, and yes, they will be huge. the shift in color may be due to a conversion between yuv (wich your camera and nle work in) and rgb (which ae, and possibly some of your other software work in).

    you may be able to find some setting to try and compnsate for the shift either at import or export/render. in ae, if you were to bring in the dv straight from capture, you can use the interpret footgae window and select expand 601 luma levels, to bring yuv footage up to rgb values. if you can work with rgb or convert it back at the next stage then you are ok… if you can’t convert it, you can create an adjustment layer at the top of your comp and add levels, setting the white output to 235 and black output to 16 to convert back to yuv (or 601) color space.

    Kevin Camp
    Designer – KCPQ, KMYQ & KRCW

We use anonymous cookies to give you the best experience we can.
Our Privacy policy | GDPR Policy