Upgrade Caption to Graphic does not work with Presets with Type: Scale

[KYLER HOLLAND]

Hello,

I am experiencing an issue involving Upgrade Caption to Graphic.

Watch this video to see the bug:

I have created a simple preset that scales the width up and I set the Type: Scale so that the keyframes automatically adjust to the layer length. However, when I apply the preset to the Upgraded Caption to Graphic text layers, the preset does not scale. To test, I created a new text layer and applied the same preset and the preset worked (scaled accordingly) I will attach 2 photos to show the difference. Please fix this as soon as possible becuase this is not allowing me to use Automatic Captions if none of my animations work.

The photo below is showing how the preset should work. This is a text layer I made without automatic captions and upgrade caption to graphic.

The photo below is showing that the preset is not working on Upgrade Caption to Graphic text layers.

Is anyone else experiencing this issue? Please let me know what can be done. Thanks!

-Kyler

[KYLER HOLLAND]

Lemme know if you know anything about this! Thanks! 

[KYLER HOLLAND]

I've talked with other creators and they are experiencing the same issue.

[KYLER HOLLAND]

Still having issues with this. Totally slowing down my workflow. Would love the ability to use presets on subtitles. Please let me know if anyone from Adobe has seen this. Thanks!

[jstrawn]

I'm sorry, I'm having a really hard time following what you are trying to do. What version of Pr are you using?

Could you write out some steps for me to follow? You probably don't have to include the parts about upgrading caption to graphic or comparing that to native Pr Text. Also, sending me your preset might help. You can do that using the direct message feature of this forum. You're looking for a file like this in your Profile... Effect Presets and Custom Items.prfpset

Also, is there a reason you need to use the transform effect for this? Generally speaking, animating Scale for a Text Layer or for a Whole Graphic is a more reliable method.

[KYLER HOLLAND]

@jstrawn 

Watch this quick video for an overview of the issue: Upgrade Caption to Graphics BUG in Premiere Pro 

Thank you for replying! I appreciate looking into this. I am using the latest version 23.5 on both Mac + PC with the same results.

-- "Also, is there a reason you need to use the transform effect for this? Generally speaking, animating Scale for a Text Layer or for a Whole Graphic is a more reliable method."

The main reason for using the transform effect is to have special effects (not just scale) such as position/scale/scale width/ etc. I can animate the intro and outro of the animation (save preset as Type: Scale) and have it automatically adjust to the length of the text layer. But with upgrade caption to graphic text layers, the animations do not adjust with text layer length. This is not just a bug with transform effect but any preset created with Type: Scale. I've done testing as well and even "Anchor to Out Point" does not work. The keyframes will not automatically adjust to text layer length. The overall goal is to be able to create subtitles and then apply motion presets/effects to all of the subtitles I created but since the presets are not locked down to the intro + outro, the keyframes are all over the place since the subtitle layers are different lengths.

 -- "Could you write out some steps for me to follow?"

GOAL: To add effects/presets to all Upgrade Caption to Graphic Text Layers.

1: Create any preset you would like with the keyframes at the intro and the outro of the layer. (ex: transform effect - keyframe at start Scale 100%, keyframe at end Scale 120%)

2: Save the preset with Type: Scale. (We want the preset to adjust any text layer and adjust length of keyframes automatically.)

3: Now create a captions track based on any video with audio you have.

4: You are unable to apply effects directly to the subtitles so we need to upgrade captions to graphics.

5: After upgraded, we can now drag and drop any effects/presets on to the captions. However, none of the keyframes on those text layers are correct.

6: You can recreate this by creating your own text layers and compare to the ones created with upgrade caption to graphic. Just drag on the preset you made to All of each and you will see why the keyframes are incorrect for upgraded graphics.

I hope this provides some insight. Let me know if you need anything else @jstrawn 

[Stan Jones]

@KYLER HOLLAND,

It has taken me several days to get caught up enough to respond to this. I have done previous testing on the nature of Caption to Graphic (CTG) items, and it becomes mind-numbing. So until I had a chance to test, I had nothing to offer.

Aside: For example, CTG inherits the “paragraph text” property, including position properties of the Text Anchor Point and the Text/Transform to the upper left of the bounding box. Point text is bottom left. This confuses and frustrates users who want to zoom and cannot figure out how to get the CTG anchor point into the middle of the bounding box. Or they move the anchor point correctly in the Text/Transform section, and then zoom based on the Vector motion or Video motion sections. One strategy is to set the captions to the center, then apply the Transform effect and nest in the main timeline to move the animated “captions” to the bottom of the frame. I wonder if there is any alternate workflow for you.

You have discovered, and I have confirmed, that something is wrong with the CTG when a preset is applied. I simplified this by eliminating the transform effect.

@jstrawn, re your comment "You probably don't have to include the parts about upgrading caption to graphic or comparing that to native Pr Text." You may have simply meant that the steps were clear. Because of the results of my earlier testing, I do not assume that CTG and native text are necessarily the same. But I also keep thinking they MUST be and this is some kind of time calculation oddity. And, of course, when I get into the weeds like this, I could just be wrong!

I tested using 23.5.0. I used a single-clip timeline. I realized after the fact that this is probably VFR but nominally 29.97. I cut it into clips of varying lengths. I used a 3:09:00 section to create gaussian blur presets. "Anchor in" set keyframes at zero and 1 second, blur 20 to zero. Anchor out, end minus one second and end, blur zero to 20. And Scale was zero and end, blur 0 and 20.

All the presets work correctly for video clips, and typed graphics text created as point or paragraph text.

CTG of all durations works for Anchor in.

CTG of duration 4:29 works for all conditions.

CTG of any other duration does not work for Anchor out and Scale. I tried 2:29, 3:00, 4:00, 5:29. I did this by setting the caption length, then applying the Upgrade Caption to Graphic. I was copying the CTG to my test timeline, but it was matching.

There is an interaction with Edit -> Preferences -> Timeline -> Still Image Default Duration.

With the usual default of 5 seconds (4:29), a CTG of 5:29 places the end point (for Anchor Out or Scale) at 4:28. With the preference set to 4 seconds (3:29), the Presets set the CTG 5:29 end point at 3:28.

Next step; update Beta and test there.

Stan

[KYLER HOLLAND]

@Stan Jones 

Appreciate the detailed analysis. I will continue testing but I'm glad you have validated the bug as well. I would love for the CTG's to work like text layers as that seems to be the intended use. The ability to apply ANY effects for to captions would be beneficial, hense why there is an option to "Upgrade Caption to Graphic"

"--Aside: For example, CTG inherits the “paragraph text” property, including position properties of the Text Anchor Point and the Text/Transform to the upper left of the bounding box. Point text is bottom left. This confuses and frustrates users who want to zoom and cannot figure out how to get the CTG anchor point into the middle of the bounding box. "

--I do agree that the bounding box may be confusing and complicated for some presets. The CTG seems to be a bigger box than actually around just the text presented. In my testing, only using Anchor to In Point (since that is the only one that works atm), the scale is way off due to the anchor point issues. Maybe this bug will also present the issue with the bounding boxes as well. We could theroetically prevent this issues by adjusting the anchor point on the preset but that adds another level of unnecessary complications.

We can only hope all of this makes sense and Adobe will update Beta so we can test further. Currently, My presets and animations will have to be locked down to Anchor in. 

Thank you for your input. I greatly appreciate it.

[Stan Jones]

Noted. I did file a bug report in the Beta forum:

https://community.adobe.com/t5/premiere-pro-beta-bugs/preset-applied-to-caption-to-graphic-uses-wron...

Stan

[Stan Jones]

Re workarounds. I'm stumped, but here are some options I'm working with for single word "pop-in" style captions.

Create captions to create single words. You can maximize this with max 7 characters, min dur 1.2 seconds, single line.

Then style them while they are still captions as Center Middle Zone, center/middle paragraph alignment, and any non-animated font properties.

Once upgraded to graphics, they are centered etc. Use the Transform effect, as you are, but yes, limited to Anchor In  preset. Many of the pop-in styles tend to do that anyway. Now select all the caption to graphics, and nest. Then use the nest's properties to move it where want on your screen (e.g. back to lower middle captions position).

For this workflow, this avoids the bounding box and anchor point issues. 

I realize now that I ran into this Anchor Out doesn't work, but I just thought I wasn't using it correctly.

Stan

[jstrawn]

Thanks for that video, @KYLER HOLLAND, that helps so much. It does look to me like some unexpected behavior is going on there, but I need time to run this by the Graphics team to see what it should be doing. If we can identify a bug there, I will report it internally and try to get it fixed. Even if there technically is no bug, there still may be a feature enhancement request we could make from your report. I'll let you know how that goes.

In the meantime, you should know about "Responsive Design - Time" which will protect your intro and outro and scale the in-between keyframes as needed. Give me some time think about a way to maybe proliferate that to other Graphics without having to make a mogrt and re-apply it everywhere.

@Stan Jones There was a lot there to parse. The only reason I said Kyle could skip the steps about upgrading to Graphic and comparing to Pr is because I wanted to save him some writing time. I already understood what he did for those parts and did not need detailed steps for those things.

[jstrawn]

Update: I have talked to the Graphics team about this and we have identified problem with upgraded captions. I created an internal report for this and a fix is forthcoming. You can try it in a Beta build if you don't want to wait for the next major release. Note that the fix won't affect existing Graphics that were upgraded from Captions. You will have to upgrade things again to benefit from the fix.

Also, you should consider using Vector Motion for animating Position instead of the Transform Effect. Vector Motion was specifically designed and implemented to work well with Pr Graphics.

[KYLER HOLLAND]

@jstrawn + @Stan Jones 

Thank you for the support on this. This is great news. Please keep me updated on next steps. I will try out the Vector motion and see if it will work!

[KYLER HOLLAND]

@jstrawn + @Stan Jones It looks like the latest build of Beta has fixed the issue with applying presets to upgraded captions to graphics. However, now the issue lies within the bounding box of the text layers. On upgraded captions to graphics, the box is much bigger than regular text layers thus offsetting the Anchor Point. When we apply presets such as scale up, the upgraded captions do not scale from the center of the text layer. Any thoughts on this?

[Stan Jones]

I'll test this when I get a bit of time.

> On upgraded captions to graphics, the box is much bigger than regular text layers thus offsetting the Anchor Point. When we apply presets such as scale up, the upgraded captions do not scale from the center of the text layer. Any thoughts on this?

I would start by looking at my comment above where I suggest setting the caption track style to center etc, so it is easier to  animate/position after the upgrade to graphic. You can use the transform effect to move on screen. Since the box is centered, you can use the non-text scale etc.

https://community.adobe.com/t5/premiere-pro-bugs/upgrade-caption-to-graphic-does-not-work-with-prese...

Stan

[jstrawn]

@KYLER HOLLAND  I'm very happy to hear that the original problem that you reported is fixed for you in the latest Beta build. You can expect it to be fixed in the next public release as well.

When a Caption segment is upgraded to a Graphic, we respectt the text-in-box size and location that the caption had. If you want a smaller text box, then you can make the box small in the Program Mon before upgrading the Caption. You can resize the text box after the fact too, but as you noticed, the Anchor Point will be in the upper left at that point so the text box will not be resizing in a symmetrical way. That is because we make upgraded captions have the same Anchor Pointtas newly created text-in-a-box layers do... with the AP in the upper left. If you want to add something like a Scale Up animation, you'll need to center the Anchor Point inside the text box first. Snap In Program Monitor can help you do that, btw.

I know that's probably not ideal for your workflow, but those are my thoughts on the matter, which is what you asked for, right? 🙂 

[Stan Jones]

@jstrawn KYLER HOLLAND,

My testing is a bit slow. I enjoyed our 10 days in Colorado a bit too much. So for now, a general comment: I have been tracking, and working on improving, several workflows in this area for some time. A lack of some features has been a problem, improved greatly by the Upgrade Caption to Graphic. Bugs were another problem, including the one in this thread - not having a preset with Anchor to Out - as well as problems with the Create Captions Presets - also with recent fixes. This opens possibilities that I will be testing. The entire captions workflow change introduced in PR 2021, speech-to-text, and text-based editing have introduced important features. And with each of those, more feature possibilities, inevitable bugs, and new directions. Users do get confused, and we can second-guess UI decisions, etc. Overall, however, I give very high marks to staff for continuing to move this forward very effectively. I like the early introduction of major features in Beta, and the responsiveness of staff to suggestions. Shout out and FYI to the several of you I have flagged on various posts: @TeresaDemel @AnnikaKoenig @Kerstin Ebert @Alexander Riss @NicoBecherer  @Fergus Hammond @Francis-Crossman @Kevin-Monahan @Trent Happel 

> ...you can make the box small in the Program Mon before upgrading the Caption.

This is the approach I'm testing, and described in my earlier post in this current thread:

https://community.adobe.com/t5/premiere-pro-bugs/upgrade-caption-to-graphic-does-not-work-with-prese...

Maximize formatting (centering etc) and styling (font/font size etc) before upgrading to graphic. The attempt is to limit odd sizes and positions from the upgrading. Many users then use the Transform effect and nesting to make adjustments over multiple items. I also tried making very specific adjustments to the bounding box via EGP Caption Track Styles, but put that on the back burner since simply centering it all (with varying box sizes) and then using the Transform Effect avoided this.

My context is often looking at how much of an effect, such as pop-in captions, can be accomplished without having to manually do one graphic at a time. You can't paste attributes for text without pasting the text itself. Is that true?

Feature issues: toggle paragraph vs point text? Paste attributes to a graphic without pasting the actual text? Admittedly, for all the text-related focus I have, I have a lot to learn about Mogrts. Can you apply a mogrt to multiple text graphics, keeping the text?

Is some of this already possible?

jstrawn, Thanks for the reminder about "Program Monitor Snap." I had it turned off.

Stan