Progressive or interlaced?

[sergiovalentino]

My question is: does Marvel g400 capture progressive or interlaced frames? If interlaced which is field order, upper first or lower first?

thanks

sergio

[sergiovalentino]

Then, why Premiere preset is "no field"?

thanks

sergio

[This message has been edited by sergiovalentino (edited 11 May 2001).]

[hippie91]

well you're right about the field order b thing, however, field order b is the lower field, not the upper one

[sergiovalentino]

Still confused.

sergio

[Dr Mordrid]

Every card is different as far as which field order it uses.

Premiere doesn't know what card you're using so it defaults to "no field", which means frame (non-interlaced) mode.

The Matrox cards have all used field B since the Marvel G200.

The old Rainbow Runner Studio used field A.

Dr. Mordrid

[sergiovalentino]

Thanks doc,
this mean i've to set "lower field first" in Premiere, right?
However this setting change or affect performance and quality?

sergio

[Dr Mordrid]

Setting the proper field order will give you the best quality. If you're using the wrong one you get all kinds of playback glitches.

Using "FRAME" is also OK, even preferable if you're using a lot of overlays, but this should be assigned on a per-clip and as-needed basis in the editor and not in the project settings. When this is done the export should be to the proper field, be it upper or lower.

The idea is that using "frame" for the clips in an overlay prevents field artifacts in the effect. These look like a combs on acid and are the result of a line of one field order in the foreground overlaying and interfering with a line of the opposite field order in the background. Using frame prevents the interaction and cleans up the result.

Dr. Mordrid


[This message has been edited by Dr Mordrid (edited 13 May 2001).]

[sergiovalentino]

Thanks!

[KuroNeko]

For grabs that will wind up as mpeg2 later (progressive), what would be better, Frame or Interlaced and have a deinterlace filter applied later?

On that subject, is there a good adaptive motion deinterlacer software?

Neko

[FM]

Interlaced, upper field first (filed order B).

[hippie91]

i'd say either edit it as interlaced, then at the final render deinterlace it, or better yet, deinterlace it before doing anything to it, then edit it as frame. if you get a good program to deinterlace (i use progressive dv so i can't really suggest one, or comment on the quality of premiere's) then the final product should turn out excellent.

[SebastianX]

The Marvel uses Upper field dominance. I'm sure of this because there's a test to determine this using After Effects. I did that test and it came pretty obvious that the dominant field is the upper one. It's also referred to as Even field or Field A.

Sebastian

[Dr Mordrid]

Sorry to disagree, but regardless of the test used in AE the Marvel's are all field B dominant. This has been the case since the G200 Marvel beta first started and has carried forth to the G400 Marvel.

This is easily shown in MSPro by opting for field A in project settings and starting a preview. This forces a full-project render and produces playback problems when printing it to tape.

Using field B allows MSPro to smartrender and results in a smooth playback.

Dr. Mordrid


[This message has been edited by Dr Mordrid (edited 16 May 2001).]

[Frankie G]

Then could someone please publish a table of all the different field order terms:

Field Order A/B
Upper/Lower Field Order/Dominance
Even/Odd Field

Which go with which? I seem to remember TMPGenc changing
its terminology between versions as well -- confusing to
say the least! (I go with Even when doing MPEG-1 and it
looks OK, guess that's what G200 Marvel is).

[SebastianX]

Answering Dr. Mordrid post, I'm a little puzzled now. I tried the test you mention in MSP6 and it would look like the lower field is the dominant one, however, the After Effects test is one of the most widely used to determine the field dominance. It's simply inserting the movie file into a composition and interpreting the file trying both fields: in the wrong setting, the movie will go back one frame every ten frames or so, so the field order that doesn't show that anomaly is the dominant one. I did that test in After Effects with the HuffYUV codec, the PicVideo Mjpeg codec and the Matrox Mjpeg codec, and the three of them show the upper field is dominant. I couldn't say why MSP6 sees it the opposite way though, but there must be a reason.

------------------
Sebastian