URGENT: Major EDL Bug

You guys may have read posts (from CreativeCow) relating to the problems many of us are having with creating EDLs.

I'm not going to go on in a venting rage at Apple or anyone, although not to is taking much self control.
I don't know if FCP 4.1.1 is to blame, but that's the version I'm running - can't vouch for anyone else.

The problem is very simple to replicate:

Simply create a timeline and lay down some shots. - I'm using Targa sequences as I use Cinéwave.
Place dissolves between the shots.
Export an EDL - I'm using CMX3600
Now import that EDL.

Notice that the first dissolve is (in my experience) accurate, but the following dissolves have all gone crazy.

What's happening is the EDL has created a cut in the outgoing clip 1 frame before the dissolve, and this new tiny clip contains a timecode from a random point earlier in the out going clip.

I discovered this by being abused from a Director for making him waste his money in a Flame suite cleaning up my EDL for an hour.

This bug is preventing us from including dissolves in our edits and it MUST BE RESOLVED asap.


Please guys... whoever has the clout to escalate this problem to someone who will ACTUALLY LISTEN at Apple, please help us.


?
Steve

Unbeleivable!
How did somehting this major go undetected for so long!

In my tests, I dont get any disolves at all, just the strange smal clip that is half the length of the transition.
Some were 1 frame shy of being half the length, and on these, there was a one frame gap before the next clip started.

Making the EDL with the "Omit Transitions" option selected gives a cleaner list.
However, when I treid that, the last clip, instead of starting at the center of the disolve, starts at the HEAD of the disolve.
If my transition had been more than 1 second, I wouldnt have had enough handle to do the disolve, and would have had to recapture if it had gotten that far.
Did you get that Steve?

Nick

I think what your describing is ANOTHER bug with EDLs!!!

There are two items in the EXPORT EDL window that relate to transitions.
One is OMIT TRANSITIONS, the other is a checkbox named TRANSITIONS.
The checkbox item, when checked, creates additional text description in the EDL, namely "EFFECT NAME - CROSS DISSOLVE" under the event that the dissolve takes place.
From what i can gather, the OMIT TRANSITIONS feature will not provide any details regarding transitions (obviously) in the EDL. However, when FCP imports an EDL generated with the OMIT TRANSITIONS feature DESELECTED, but with the other TRANSITIONS checkbox marked it doesn't bring in the transitions... even though when viewing the EDL list in text edit I can clearly see Dissolve events & information (just not the "EFFECT NAME - CROSS DISSOLVE" info).

The problem I'm referring to, is that when FCP edl's are imported (either to FCP or to Discreet Flame) they include dissolves, but the dissolves are wrong.

The additional edit created by the EDL for the dissolve to take place has an offset timecode relating to the duration the dissolve is meant to be.

This sounds horribly confusing, I know.

In other words - shot A dissolves to shot B over 20 frames, right?
What happens is the EDL creates a cut in shot A one frame before the dissolve, and that clip's timecode, instead of being the next chronological frame of shot A is in fact 20 frames earlier - so when played back you get a sudden jump backwards in time before you dissolve to shot B.


This bug is easily replicated - please someone in the know, hear our cry!!

Hi Steve,

Using FCP 4.1.1, OS X 10.3.2, QT 6.5, PAL DV Sequence and CMX 3600 EDL

I can replicate the timecode offset error on a second dissolve in a CMX EDL but the timecode offset on my system/sequence is "always" (at least in the tests until now) one second (25 frames).

First of all, the extra edit-event inserted at the starting point of the dissolve is not an error.

Lay down 3 clips in the timeline next to each other: A, B and C.
Add dissolves between cut A-B and B-C.
Export EDL from FCP and open the EDL in TextEdit.
On my system, the edit-event describing the second dissolve (B-C), the starting clip's timecode (B) is exactly one second early but the clip that is dissolved to (C) is correct. It doesn't seem to matter what length the dissolve is, wether it's centered on start, center or end, or if it is a default effect with a different length: the source timecode (B) is one second less that it should be. It also seems to happen with audio tracks as well. This is clearly a bug! And a most annoying one as well!!

The short edit-event on the dissolve start, however, is not a bug. EDLs were designed for linear tape-based editing. If Clip A and C were from the same reel, the source-VTR would need to shuttle to the new in-point after the first dissolve was done, and would in most cases not be ready for the dissolve from B to C. Hence the need for an extra edit-event to allow the VTRs to catch up. If you insert a new clip D in the middle of clip B, the second dissolve in the EDL does not have the extra edit starting at the dissolve.

Steve, are you working in NTSC? Are you sure the timecode offset on your system is related to the dissolve length? I haven't been able to replicate this.

All the best - Peter

Hi!

I can confirm exactly the same on a PAL-project! The strange thing is that it happens only on some transitions! The dissolve going into a clip might be fine, the one going out of the same causes the same EDL-errors as described above.

From the FCP webpage:

"A new plug-in API, enhanced EDL support and FXScript ? a visual effects scripting language ? make Final Cut Pro 4 the most extensible NLE available."

Hmmm.

[How did somehting this major go undetected for so long!]

Probably because it's something new! I was doing EDL's from version 2.x, which included straight dissolves, and they behaved well-- only issue I saw was a 1 frame latency on straight cuts! That was personally investigated by Brian on a site visit and solved- FireWire latency issue. Never seen the dissolve truncation before-- but haven't thought to look for it and I don't do a lot of EDL's anymore.

- Loren
Today's FCP4 keytip:
Open a bin in its own tab with Option-Enter!

The FCP4 KeyGuide? is the most affordable power tool you'll ever own.
Now available at
www.neotrondesign.com

hi all

just had this same prob occur recently where the flame guy came downstairs and gave the offline guys some trouble hehe ... anyways he says that its a problem relating to copying dissolves from one to another he says if u just grab a regular dissolve and change the duration to fit your needs and not copy an existing dissolve of a specific length then it woks fine ...

Nope, sorry.

Drag, Shift Overwrite, or Control Click all produce the same results.

Shift Overwrite, and Contorl click aply my curent default of 1sec Cross Disolve.
My Drag was a saved 2 second Cross Dislove


Nick

I am receiving EDL's from a FCP 4 user and I too confirm the problem described here. I am working in an AVID Symphony. EDL's are processed in EDL Manager. If I enable "Dissolves as cuts" in EDL Manager the timeline seems to be Ok but I lose all the crossfades. Not usefull. So I have to painfully import 2 times each EDL: one with dissolves and one without. I then laydown each timeline one ever the other (track V1 and V2) and copy each mix from the bad one to the one with cuts only. It is time consuming and I too would like to say some %"$/?%*&?!!! to Apple for doing this so bad!

Anybody knows a better workaround for this?

I too just found about this bug. My online editor never told me he was having problems.
I tested using a different transition, dip to color, and the bug was still there.
So it's not just cross disolve.
I checked some old EDLs from FCP 3 and they are all good. No bugs.
FCP 2 was also fine.
I need to check some EDLs from FCP 4.0 to see if they are ok.
I'm using 4.1.1 now.
Looks like we will have to checkerboard or tracks, what a pain.

Anybody tried exporting using Automatic Duck?
I should try exporting with Automatic Duck into Avid and then
use Avid's EDL Manager to make an EDL.
Sounds like a long shot.

hi, i'm first come here...

I'm using FCP4.1.1, always export CMX 3600 EDL

my another MAJOR EDL probelm is.. the speed of the clip..

example:
I have a 1 second clip, timecode 01:15:01:00 to 01:15:02:00
in the sequence, I slow it down 50%, so it become 2seconds
then I creat EDL... it comes out....
01:15:01:00 to 01:15:03:00 <-- 2 seconds of footage
how could it be? I only applied 1 second of footage only...

another example:
I have a 4 seconds clip, timecode 01:15:01:00 to 01:15:05:00
in the sequence, I speed it up 200%, so it become 2 seconds
then I creat EDL... it comes out...
01:15:01:00 to 01:15:03:00 <-- 2 seconds of footage
how could it be? I used 4 second of footage...

this is the setting that i'm using:
Format: CMX 3600
Reel conflicts: Generic Edits

that's all...
someone help me please....!! >_<

Annie

I have noticed the speed problems too in FCP's EDLs.
To make things worse, try doing a cross dissolve from a
speed clip to something else. The EDL gets more messed up.

Or does it?

On the good side, try importing your EDL back into FCP.
You will notice everything is fine.
The timecode on your 1 second clip at 50% speed
will still retain its accurate timecode numbers even though the edl says otherwise.

Now test this on an Avid or a CMX system to see if it comes out accurate
as well?
It just might.