This IS the Avid L2: Re: [Avid-L2] Best method of converting a 23.976 file and or Avid timeline to 29.97P file?

Monday, November 9, 2015

Re: [Avid-L2] Best method of converting a 23.976 file and or Avid timeline to 29.97P file?

When I said I was viewing on a broadcast monitor, I meant just that: HD-SDI feed to a Sony PVM. I'm well aware of the playback differences of computer screens and broadcast monitors, which is why I waited until I was in work with access to a broadcast setup before I was able to properly test this. (Playback on a computer screen will show you either both fields at once, or a deinterlaced preview, so field order is irrelevant.)

I'm not sure how I can describe this any better than I already have. When you set the field order interpretation for a file, you are telling your computer which half of the frame to play first: Upper or Lower. You're NOT telling it to take either half of the frame and shift it into the position of the other frame.

However... if you then take the incorrectly interpreted video (now set to play Lower-Field-First) and put it in a Premiere timeline set to play back Upper-Field-First, THEN there has to be transposition (switch) of fields. Comparing the LFF version with the original (correct) UFF version appears to show that *both* fields (i.e. the whole frame) has shifted down by one line as part of this process. The result is that the clip still plays back with reverse field order (i.e. horrible shuddering) in a UFF sequence, so that if I were to export this, I would have a UFF export containing a clip that plays back with a horrible reverse field-order problem. That's how it gets baked in and, sadly, broadcast by people with a lack of properly trained QC department. (Or a newsroom with no time to properly QC at all.)

The diagonal line is still NOT aliased in any version though.

On 9 November 2015 at 17:41, bigfish@pacbell.net [Avid-L2] <Avid-L2@yahoogroups.com> wrote:

What monitor did you view the diagonal line on? Was it an external sdi feed monitor or the computer screen. It was mentioned in this thread that you can't see the reverse field order on the computer monitor so I wonder it that might mask something. I just don't see how if field 2 plays back first it has to be scanned at a field one time that mean the video information is filling the lines for field on which are offset from field two. Not sure how the half line start would play into it.

---In avid-l2@yahoogroups.com, <speckydave@...> wrote :

Just tested this on a broadcast monitor by adding a big diagonal line across some upper-field HD footage. Exported as QuickTime ProRes upper-field, which I them imported into Premiere and played back: smooth motion and no aliasing while interpreted as upper-field. When I switched the interpretation to lower-field, I got all the horrible shuddering from fields being played in the wrong order, but the static diagonal line showed no aliasing at all.

I would say your assumption about fields being re-positioned is wrong. When someone mis-interprets upper-field-first footage as lower-field-first, they just get the fields played in the wrong order. That's it. No re-positioning. Just (in this case) the lower field played first, then the upper field.
I would have thought that to take the upper field and shift it down a line to the lower field position and vice versa would be quite a deliberate process. I can't right now think of a way to *accidentally* tell the system to actually reposition the fields like this (though since there generally tends to be a way to easily screw anything up, I don't doubt that it can be done with a single bad keystroke...).

D.

On 9 November 2015 at 01:24, bigfish@... [Avid-L2] <Avid-L2@yahoogroups.com> wrote:

If the fields are in reverse order then they are not playing in the same field the are intended for. Think of a diagonal line. I am assuming that with wrong field order Field 2 plays first and is positioned on the lines intended for Field 1 and Field 1 plays second in the lines intended for Field 2. That would mean the horizontal scan lines are inverted too. To me that would create the stair stepping even on a static frame.