All right everyone calm down. There IS a bug in trimming the out point of
clips in 3:2 Expand containers, Spencer mentions that you step into a 3:2
Expand and trim the out point but when you step out, DS doesn't show the
additional frames, but rather shows black. Frustrating and repeatable and
still not fixed. But the 'inaccuracy' in 3:2 contract is the result of math
and not voodoo.
DS 3:2 Contract is accurate, that's for sure. Although one might think it
would be trivial to contract and expand and retain cadence it's not as easy
as you might think. You have to remember that with 2:3 you have 2 NTSC
frames out of 5 that contain parts of two separate film frames and two other
frames that last 1 field longer than the others (3 fields for one frame with
one field repeating). So when you have a cut that falls on a split-field
frame that contains info. from 2 separate film frames which frame do you
preserve and which do you throw away? Do you add the field from the frame
before or not?
Put another way, lets say you have a 3 NTSC frames with 3:2. One whole frame
and 2 split field frames. You have B,C, and D. Do you contract to B and C or
do you contract to B,C, and D? How can you contract 3 frames into 3 frames?
But contracting into 2 frames doesn't work either as you're missing half a
frame.
So who cuts with 3-frame clips? Everyone! Because every clip is made up of
five-frame 3:2 'sections' and then at the head and tail of the clip are the
1-5 frame 'leftovers' which are what cause the trouble.
You have to decide what the rules are for those segments. DS says, if your
segment at the head is less than 3 frames, DS will err on the side of
shorter. If the first frame of your clip contains either split-field frame
or the last whole frame (xxSSW) DS might APPEAR to be cutting a frame off
the head. It's not really but it's not what you expect.
In my opinion, DS gets over-zealous with it's rounding on the end-frame as
if your clip ends on the with the frame before the split-field frames or
with either split-field frame (xWSSx), DS will give you the shorter clip. I
personally think that the first whole frame shouldn't be included. I also
think the last whole frame shouldn't be included at the head for that
matter.
So although the results aren't what you might expect, they are technically
accurate. 3:2 Expand is a topic for another time, and I know this email was
long on explanation and short on useful tips but I may have some more long
DS renders tomorrow to fill in some blanks.
I know defer to Bob Maple, Howard Chasteen, and others to make corrections
and additions then move the final article be submitted to the documentation
team for addition in the DS docs. Then they can edit it for the Symphony
Nitris docs and put our names on it as devoted Symphony fans...
-Todd
Union
> -----Original Message-----
> From: owner-ds(at)Softimage.COM [mailto:owner-ds(at)Softimage.COM] On Behalf Of
> spencer
> Sent: Thursday, November 17, 2005 18:36
> To: DS(at)Softimage.COM
> Subject: Re: 3:2 contract issue
>
> I can't remember the time when DS actually could pull the 3:2 out of
> that 'laboratory' clip and put it back together again without some sort
> of issue.
> I even believe that 3:2 issues were brought up on this list by a few
> in past years and no one seemed to really understand the issues. Maybe
> they didn't care. Maybe everyone just does "workarounds". I don't
> remember... so I do "workarounds".
> I have noticed the 3:2 issues on version 7.0 systems 7.6 Equinox
> systems and on 7.6 training stations so far all versions deal with that
> 'laboratory' clip the same way. Basically I don't trust DS to do 3:2
> correctly... ever.
> What I mean by all of this is:
> Yes, DS will pull the 3:2 out. Yes, you can build your effect. And yes,
> you can put 3:2 back into the shot.
> But to be sure that the new affected clip is frame/field accurate a lot
> of manual tweaks must be preformed to wrangle DS into putting the ABCD
> frame into their correct positions. Now if the clip's source if tape
> based then I have noticed that DS has less of these following issues.
> As or the 'laboratory' clip well I think I have seen what your seeing.
>
> Try this: (I know you have probably done 3:2 work alot or you wouldn't
> have noticed the issue. I'm listing this out so those who are new or
> haven't seen the issues can try... maybe their DS are working and only
> the systems I work with are "buggy". NOTE: I'm also not in front of a
> DS and writing this from memory)
>
> 1. Place 'laboratory' clip on V1
>
> 2. Place a copy of 'laboratory' on V2. Be sure they line up.
> (work only on the V2 clip so that you can deactivate and check it's
> frame/field accuracy after you have pulled the 3:2 out and put it back
> in. Also take note that the clip starts on a B frame)
>
> 3. Park the position indicator on the first of 2 split frames. (The
> frames near the end where the green goo is moving up the tube are easy
> to identify the split frames)
>
> 4. Select Clip. Then "Time Effects>>3:2 Contract."
>
> 5. DS will pull the 3:2 out of the 'laboratory' and pad the container
> with 1 red frame at the head. The clip will look shorter due to the
> fact that your now viewing the original 24fps footage
> At this point DS is working correctly the red frame is a padded frame
> DS creates to recreate the missing A frame. Since the 'laboratory' clip
> comes from file there are no additional heads for DS to pull cadence
> from. if the clip had come from tape the it might show a frame from the
> heads or a media not available depending on your personal situation.
> Don't trim the red frame off... I imagine most people do thinking its a
> mistake. If this frame is trimmed at this stage then your cadence will
> not match later.
>
> 6. At this point you could do your work on the 3:2 contract container
> but I usually fix the 3:2 first and make sure that's working and lined
> up. then I can step back in and work on the clip after everything is
> lined up. So for now just place a DVE on the 3:2 Contract Container and
> leave it unaffected as a placeholder for a future effect.
>
> 7. Select 3:2 Contract Container. Then "Time Effects>>3:2 Expand."
> (It doesn't matter where you park your position indicator. DS inserts
> Cadence baised on the first frame. Hence the 1 red frame.)
>
> 8. DS will reinsert 3:2 into your clip. You'll notice that it has gone
> back to being the exact same length as it was before. Seems good.
> But here is where the inaccuracy starts. The first frame will still be
> red. Although this is a good thing for cadence, Your clip is now 1
> frame out from where you were before.
> The 'laboratory' clip started on a B frame. But your clip starts with a
> red A frame. You'll need to trim or slip the footage to remove this red
> frame.
>
> From here I will take the "Trimming" route and not the "Slipping". (The
> "Slipping" route seems to give the same frustrating results.)
> 9. Select the front Edit point and trim off the 1 red frame. Good now
> your clip starts with a B frame... but it's also now 1 frame short.
>
> 10. Drag the clip 1 frame to the left. Your clip will be in the right
> place and if you check the frame/fields they should line up with the
> shot underneath. But your still one frame short.
>
> 11. So select the end edit point and trim out the clip to match it's
> original length.
> Here is where I really notice issues with DS.
> You'll trim out the clip and at first you'll think it worked. If you
> look up at the ribbon you'll see that that 1 frame doesn't need to be
> processed. It's because DS is showing the frame for the clip on V1. Not
> the footage for V2.
>
> 12. Solo V2.
> You'll see that DS doesn't show that 1 frame at the end. Why?
> Because DS did pad the front of the 3:2 Contract Container with 1 red
> frame. But DS did not add 1 or more frames to the end to match. So your
> 3:2 Contract Container is simply 1 or 2 frames short.
>
> 13. Step into the 3:2 Expand Container.
>
> 14. Trim the 3:2 Contract Container out by a frame or two. You'll see
> that the missing frame has simply been trimmed off by DS and any
> additional frame you trim out DS will pad as red frames. (also if you
> had done any animation work at step 6 like Roto you would have do 1
> more frame possibly 2. Not that that it is a hugh issue.
> And don't forgot to turn on cropped curves for your effect(s) or your
> animation will be scaled out with the trim. In either case this is why
> I get the 3:2 part working first then go back and do the work when the
> frames/fields are all line up.)
>
> 15. Step up and back to the top timeline. Notice DS has NOT put that 1
> frame into the 3:2 Expand Container.
>
> What I do from here varies on a machine to machine basis.
> 16. Step back into the 3:2 Expand Container and although you have
> already trimmed out the end of the 3:2 Contract Container try just
> re-trimming the edit out by 4 frames maybe 5. Maybe trim some frames
> off instead.
>
> 17. Step to the top timeline and view results.
>
> 18. Repeat Steps 16 and 17 until DS puts media into the last frame.
>
> 19. When DS does finally put that last frame in take a quick look at
> the first frame.
> Alot of the time for me DS simply slipped the footage which has placed
> that 1 red frame back at the head of your 3:2 Expand Container clip.
>
> 20a. If this is so: repeat Steps 9 through 19 until DS puts the
> frames/fields in the correct place.
> 20b. If DS did put the frames/fields in the correct place then step
> into the 3:2 Expand Container and preform your work on the 3:2 Contract
> Container Clip.
>
> Eventually DS will get it right and your frames/fields will be in the
> correct place.
>
>
> Alot of work to keep your 3:2 frame/field accurate.
> I show this "fix" in my 201 classes. 3:2 always fails and basically
> each student has to do steps 9-19 a few times before DS gets it right.
> I have them save once it's right and then continue on with the rest of
> the tutorial.
>
> Maybe it will help some of you.
>
> _spencer
>
>
>
>
>
>
>
> On Thursday, Nov 17, 2005, at 14:26 America/Los_Angeles, Benoit
> Melancon wrote:
>
> > Actually, this is the old "laboratory" clip provided with the DS
> > courseware, and it never gave me any trouble before....
> >
> > Here goes rebooting....
> >
> > Benoit
> >
> > Bill Admans wrote:
> >> This sounds like you could be dealing with Advanced 2:3:3:2 cadence.
> >> DS
> >> cannot correct that.
> >> Cheers,
> >> Bill Admans
> >> Director Creative Editorial
> >> FotoKem Film & Video
> >> 2801 W Alameda Ave
> >> Burbank CA 91505
> >> Main +1 (818) 846-3101
> >> Cell +1 (818) 391-3593
> >> badmans(at)fotokem.com
> >> www.fotokem.com
> >> -----Original Message-----
> >> From: owner-ds(at)Softimage.COM [mailto:owner-ds(at)Softimage.COM]On Behalf
> >> Of
> >> Victor Wolansky
> >> Sent: Thursday, November 17, 2005 2:04 PM
> >> To: DS(at)Softimage.COM
> >> Subject: RE: 3:2 contract issue
> >> You sure the transfer was made at 24? Some time ago I had the problem
> >> and
> >> was because the transfer was made at 25..
> >> Victor Wolansky
> >> DS VFX Artist
> >> WEBsite
> >> Demo Reel
> >> 815 Slaters Lane
> >> Alexandria. VA.
> >> 443-797-3507
> >> -----Original Message-----
> >> From: owner-ds(at)Softimage.COM [mailto:owner-ds(at)Softimage.COM] On
> >> Behalf Of
> >> Benoit Melancon
> >> Sent: Thursday, November 17, 2005 4:59 PM
> >> To: DS(at)Softimage.COM
> >> Subject: 3:2 contract issue
> >> Hello DSers,
> >> Can someone test the 3:2 contract for me in 7.6 qfe3?
> >> I seem to run into problems, as one interlaced frame still persists
> >> after applying the contract container (NTSC sequence, used on the
> >> first
> >> frame of a mixed pair).
> >> (Can't reboot right now, but will later)
> >> Thanks in advance,
> >> --
> >> Benoit Melancon
> >> Avid DS ACI/ACSR
> >> www.nadcentre.com
> >> ---
> >> Subscribe? E-mail Majordomo(at)Softimage.COM with the following text in
> >> body:
> >> subscribe ds
> >> Unsubscribe? E-Mail Majordomo(at)Softimage.COM with the following text
> >> in body:
> >> unsubscribe ds
> >> ---
> >> Subscribe? E-mail Majordomo(at)Softimage.COM with the following text in
> >> body:
> >> subscribe ds
> >> Unsubscribe? E-Mail Majordomo(at)Softimage.COM with the following text
> >> in body:
> >> unsubscribe ds
> >> ---
> >> Subscribe? E-mail Majordomo(at)Softimage.COM with the following text in
> >> body: subscribe ds
> >> Unsubscribe? E-Mail Majordomo(at)Softimage.COM with the following text
> >> in body: unsubscribe ds
> >
> > --
> >
> >
> > Benoit Melancon
> > Avid DS ACI/ACSR
> >
> > www.nadcentre.com
> > ---
> > Subscribe? E-mail Majordomo(at)Softimage.COM with the following text in
> > body: subscribe ds
> > Unsubscribe? E-Mail Majordomo(at)Softimage.COM with the following text in
> > body: unsubscribe ds
> >
>
> ---
> Subscribe? E-mail Majordomo(at)Softimage.COM with the following text in body:
> subscribe ds
> Unsubscribe? E-Mail Majordomo(at)Softimage.COM with the following text in
> body: unsubscribe ds
---
Subscribe? E-mail Majordomo(at)Softimage.COM with the following text in body: subscribe ds
Unsubscribe? E-Mail Majordomo(at)Softimage.COM with the following text in body: unsubscribe ds
