Have I been encoding my DV captures using the wrong field order setting?

[rkr1958]

I have a Sony TRV-320 DV camera. I been using TMPGEnc-Plus for 4+ years now & CCE-Basic for 2+ years to encode.

Up until the last year or so I've cut out the portions of the avi video that I've wanted to encode with Virtualdub and either done a direct stream copy or frameserve to encode.

In the past year in a desire to learn AVISynth and basic to intermediate filtering techinques I've gained some understanding of details such as video frames versus fields & field order. I've also discovered that my TRV-320 DV camera records and transfers DV video interlaced with bottom field first (BFF). I've just recently discovered that the advance video settings in both CCE-basic and TMPGEnc-Plus for NTSC 4:3 default to output top field first.

Have I been encoding my DV video wrong for the past 4-years?

[redwudz]

If your encoded video isn't jumping up and down when viewed, probably not. Simple test, encode a short test clip with the box unchecked. You should easily notice if you have it wrong. AFAIK, all DV is BFF.

One of our members more experience with CCE can probably explain the behavior of field selection with CCE.

[rkr1958]

@redwudz, thanks for your reply. I did analyses of two previous DV encodes (one w/CCE-Basic & one w/TMPGenc-Plus). Based on this analyses (which is summarized below) all appears well. Also, it appears that using the defaults CCE-Basic encodes BFF interlaced to TOP field first and TMPGEnc-Plus encodes BFF interlaced to BOTTOM field first. It would be nice thought is others could confirm or refute this and the following analysis results.

I took the first vob from the authoring of each encode and copy them to my harddrive. I used DGIndex to generate a d2v file for each. I then found a fast action segment in each and used the following AVISynth script is virtualdub to see which playbacks were smooth and which were jerky to confirm if the outputted field orders of each encode played correctly.

# Import / Load needed plugins
Import(plugin_dir+"mvbob.avs")
LoadPlugin(plugin_dir+"DGDecode.dll")

# Load D2V video
MPEG2Source("video.d2v")

# Set field dominance (try both & see which one is smooth and which one is jerky)
# AssumeTFF()
# AssumeBFF()

# Seperate into "fields" using a simple bob
# Note video is treated as full-height frames at twice the original framerate
eedibob()

# Get Info (Load file in VirtualDub)
# info()

1. CCE-Basic encode.
CAP Method: VHS > DV passthrough > PC (Panasonic DV Codec)
CAP Source: AVI, 4:3 NTSC 720 x 480, interlaced with BFF
Encoded Output: MPEG-2 720x480, interlaced with TFF
TFF Default in AVS (i.e., both AssumeTFF() & AssumeBFF() commented out): smooth playback
AssumeTFF() Enabled in AVS script: smooth
AssumeBFF() Enabled in AVS script: jerky

2. TMPGEnc-Plus encode.
CAP Method: VHS > DV passthrough > PC (Panasonic DV Codec)
CAP Source: AVI, 4:3 NTSC 720 x 480, interlaced with BFF
Encoded Output: MPEG-2 720x480, interlaced with BFF
BFF Default in AVS (i.e., both AssumeTFF() & AssumeBFF() commented out): smooth playback
AssumeTFF() Enabled in AVS script: jerky
AssumeBFF() Enabled in AVS script: smooth

EDIT: The TMPGEnc-Plus results are incorrect. I just did a short encode of a BFF interlaced video and the playback is jerky. So, somehow when I originally encoded the video in TMPGEnc-Plus field order must have been set to BFF. I'm not sure how though.

O.K. it's time to do a more controlled analysis. I'm going to take two small avi files, one interlaced w/TFF and the other interlaced w/BFF and encode both with CCE-basic & TMPGEnc-Plus using their default advance settings and see what I get. I'll post what I find.

[rkr1958]

Looks like I was lucky (very lucky) with both with CCE-Basic & TMPGEnc-Plus.

If you use the wizard in TMPGEnc-Plus to set up your encode it determines with field order to use based on your input. However, I don't recall using the wizard (maybe I did) or maybe I was just lucky.

With CCE-Basic, it appears that the default advance settings are for input interlaced source w/BFF and it encodes to TFF. That's because in the default CCE-basic advance settings (show above) the offset line = 1, which is below the Output top field first stream (which is checked).. I found out that with the default advance settings if the input source is interlaced w/TFF that the encode is incorrectly set to TFF. There's no way, which I see, that CCE-Basic tries to determine the correct field order.

Maybe I'm missing something here ... but it looks to me that CCE-basic requires a much more knowledge user than does TMPGEnc-Plus w/wizard mode.