UPDATE: January 29, 2021
Since the innitial post back in October of last year we have made a ton of progress and some of the details originally posted here are no longer acurate. I have updated the post to be more accurate to the state of things today.
Here are some important changes to be aware of
Thanks to everyone for testing and for all your awesome feedback! - Francis
The captioning workflow in Premiere Pro is getting a complete makeover and you can be the first to try it out! We have a new text panel for editing and viewing your captions as text. Captions now have their own track on the timeline so they can be edited just like video. Styling uses the power of the Essential Graphics panel so you can make fantastic looking captions.
Best of all – we will be adding speech to text and auto captions powered by Adobe Sensei – coming later.
Some important things to know before you start testing this feature
So, if you’re good with all that, read on . . .
Enabling the Feature:
Download this sample project to follow along PR New Captions Workflow quickstart.zip
Play around with visual styles in the Essential Graphics panel and try editing the text in the Text panel. Read on for more detailed info on the various pieces of the workflow.
The Text Panel
Open the Text panel from the Window menu. This is where you interact with text and it has two tabs – Transcript and Captions
The Caption track - working with captions in the Sequence
We have completely reimagined the way captions work in Premiere Pro. If you have used captions previously in Premiere Pro, this is totally different – much better we think. Please tell us what YOU think by leaving a public comment below.
Create and delete caption tracks
There are a few ways to create a new caption track. You can drag an SRT file onto the Sequence or use the “Create new captions track” button in the Captions tab of the Text panel. To delete a caption track, right-click on the caption track header and choose “delete track”
Drag an SRT file on the sequence
Dragging an SRT caption file onto the Sequence will automatically create a new caption track at the top of the Timeline.
To see all the keyboard shortcuts available for captions, open the keyboard shortcuts menu from the Premiere Pro menu and search for "caption". Here are some notable ones. There are many other which do not have default assignments.
This can hide the caption track area or solo just the active caption track if you have more than one. It is useful if you want to declutter your timeline. It does not turn off the active caption track from rendering in the program monitor, however.
This will enable/disable Program Monitor viewing of the Caption track selected. Only one Caption track can be active at a time, so when you make one track active, the other tracks will automatically be disabled. You can also choose to disable all Caption tracks. Inactive Caption tracks will dim making it easier to note your active Caption track.
By default, this will show the caption track format (eg Subtitles, CEA-708, CEA-608, etc.) You can right-click the Caption track header and choose Rename to choose a custom name (same functionality as A/V tracks)
Additional caption tracks can be added by right-clicking in the caption track header – choose “Add Track”. You can also delete or rename a caption track in the same way.
Caption tracks will always be at the top of the sequence. Additional caption tracks are added on top. This will continue until you have reached 25% of the Timeline vertical space, then the tracks will stop “growing” and scroll bars will appear to the right. Caption tracks will never take up more than 25% of the upper space of the Timeline. This is dynamic, so if you increase or decrease the Timeline panel height, the 25% will adjust accordingly.
Lock the track to prevent editing. This functions the same as A/V tracks.
The Caption track header has a ‘Toggle the track targeting for this track’ button with same functionality as A/V tracks.
The Caption track header has a ‘Toggle Sync Lock’ button with same functionality as A/V tracks.
Editing timing of caption items on the track
You can use all the familiar editing tools that you are used just like editing video and audio.
Styling with the Essential Graphics panel
Open the Essential Graphics panel from the Window menu. This is where you choose things like font, size, color and position. Make sure you have at least one caption selected. This will activate the Edit tab of the Essential Graphics panel. If you have created and stylized text in Premiere Pro before, the following should be very familiar to you. However, for captions we are adding small enhancements such as zones, and vertical text alignment.
Align and Transform
Transcribing the dialog into captions
You have three options for transcribing your audio and creating captions: speech to text, use a third-party service, or do it by hand.
Speech to text
Use a third-party service
If you already have a workflow that uses a third party transcription service, you can bring in that file. SRT is the best option.
Do it by hand
For short sequences, doing it by hand may not too much trouble.
Exporting your sequence with captions
There are three options for exporting captions: burned in, sidecar file, or embedded into the video file.
Happy captioning! Once again, we welcome your feedback. Please leave comments below.
We need the ability to import captions as multiple different types without going into the Import Settings window each time. An Import Captions As dialog would suffice.
Please add to Prelude the ability to disable caption import.
Please speed up the caption reading process. I can provide source media.
Also, please add H.264 and H.265 MXF as export options so we can embed captions.
Thank you for the feedback!
For Import, are you looking for the ability to import any single Caption format and have options to bring it into Premiere Pro as a number of selected formats? eg: Import a single CEA-608 and have it automatically create Open Caption, CEA-608, CEA-708 and Australian OP-47 in the Project panel.
If you import multiple Captions at one time, would you expect specific options for each Caption imported or options apply to all Captions?
Also, would you expect warnings for changes that had to be made to conform to Caption specifications?
Hey Trent, thanks for the quick reply!
Yes, Extreme Reach wants both 608 and 708
Both bulk options and the ability to change specifics would be great
I would always expect warnings when warranted
Please please please allow MXF H.264 and H.265 exports!!!
Please check back over time as we continue to add more closed captions support to the new reimagined workflow. You can test the new subtitle editing and styling workflows now, but closed captions support is coming later.
OHhhhhhh YES please :
"Also, please add H.264 and H.265 MXF as export options so we can embed captions. "
It's not only necessary to embed captions...
It's also to "transport" timecode and multiples audios tracks with the configurator panel of AME and the quality of H264/H265 compress like we have !
Please ... DO THIS !
The implementation is looking good so far, congratulations! It definitely has a modern feel to it.
Stuff I like:
- Easy editing of the timing of the subs like regular clips!
- The ability to easily change text style for an entire track of subs!
- The text search and replace option!
- The ability to finally natively import .srt files!
Stuff that I'm missing or which feel a little counter-intuitive:
- Somehow I keep wanting to click inside of the program monitor to edit the text. It will need some getting used to for me to switch working in the new text-panel. This is probably just muscle memory on my part.
- I like having control over where I keep the sentence breaks on both lines of a caption, as to make it look good. Right now, if I wish to move a section from the top to the bottom row and use the Enter-key to do so, it shows an Enter symbol in front. Only way to make a sentence on the top line move down is to add generic spaces before it. Is this a limitation of the standard or?
- I'd love a way to export out an srt file from a caption track. This is great for doing multi-language version for instance where I can have a translator do their work directly on the SRT text file and then bounce it back into premiere.
Keep up the good work!
Appreciate the feedback. Many things are still in progress for the new Captions workflow, so expect additions with future builds.
- Editing Caption text in the Program Monitor should be coming later.
- We are also working on editing in the Text panel. We hope we can make this better in later builds.
- Sidecar exports (eg: .SRT) will be coming in later builds.
Will new caption tool acept mixed text styles srt files? Many times srt files are prepared outside premiere by specialised corporations, many times they have mixed format - regular and italic. Old caption tool dont recognize italic, or if sometimes it does, after changeing font all mixed text becomes the same - italic or regular depends of users choice. We need to have abbility to change font without losing mixed type of srt. What will be the proper srt format for italic lines?
Italic „marker” at the begining and at the end of every line
Hello! I'm running into an issue where I'm able to get a transcription of a seuqence, but when I go to "Create Captions," the wheel spins, it creates an empty captions track and nothing happens.
Are IMSC subtitles supported for import and export?
No current support for IMSC, but it is on our radar.
Thanks, the new UI looks great btw.
First of all, thank you for giving captions some much needed attention!
This update looks extremely promissing and hopefully addresses all of the issues I pointed out in my "Everything wrong with Captions" FR (https://adobe-video.uservoice.com/forums/911233-premiere-pro/suggestions/38080990-everything-wrong-w...)
OPEN QUESTION TO THE BETA COMMUNITY – what would you expect this button should do? Should it turn off captions from rendering?
With the CC button turned off, captions should remain visible in the Program Monitor and continue rendering and exporting as normal. After all, if a user wants to hide them, that's what the (temporary) eyeball icons are for.
Ideally, the CC button should have 3 stages rather than 2:
1. CC off (greyed out)
2. CC on (blue)
3 CC off but with captions present (perhaps red, or another color?) This button state would be useful for warning an editor that, although the captioning interface is not visible, it does contain captions. Viewing the Program Monitor is not a good gauge for whether a sequence contains captions, since a caption track's visibility can be turned off in addition to the CC button being turned off. Thus the usefulness of having the presence of captions communicated through the CC button!
Without having tried the new captioning feature yet, I noticed 2 issues from the explanations above:
1. You can select multiple Caption track items by Shift-selecting each item
Please NO!!! Rather than further entrenching Pr in convention-busting workflows, please change this to respect world-wide computing conventions: Ctrl-select = items selected one by one, Shift-select = all media between the first and last selected item gets selected. I wrote a FR about Pr's selection inconsistencies here: https://adobe-video.uservoice.com/forums/911233-premiere-pro/suggestions/41277514-fix-selection-inco....
Here's the chart I made for reference:
Please fix this rather than add to the problem with yet another selection inconsistency!
2. when you razor a Caption item, both items will have the same text which you can then modify in the Text panel
This adds more work for the user and is therefore less than ideal. The text in captions should act like the content in video clips: If you razor a clip in half, each half contains half the clip's content, and rolling the edit point adds/removes content from each part simultaneously. Ideally, the same concept should apply to captions. If a caption with the text "This is a caption" is razored near the middle, the first half should contain "This is" and the second half "a caption". Cutting around a quarter into the caption would result in two captions with "This" and "is a caption", and so forth. This is much preferable than ending up with two captions with the same "This is a caption" text, which would force the editor to always correct not just one but both captions.
This concept should go one step further: Rolling the edit point between two side by side captions could simulteneously remove words from one caption and add them to the other. For example, if a user has two captions, a and b, rolling the cut between them to the left would simultaneously substract words from caption a and add them to b, and vice versa. When editing captions, I don't decide where to cut the text between two captions based just on the text itself, but also on the video content. I will often cut a caption shorter or longer to match the cut between two video clips. When I adjust my video clips, the captions often need to be adjusted accordingly to achieve the best visual sincronicity between caption and video cuts. In other words, it usually looks better if the cut between 2 captions happens at the same time as the cut between 2 clips, rather than just before or after. Users often need to adjust the point at which two captions cut to best match the visuals. This is why being able to roll-edit the content between two adjoining captions would be so useful!
I'm looking forward to testing out the new features and providing additional feedback! The current Pr Beta version that's available to download from the Creative Cloud Desktop app is v14.6. How can I download Pr v15? Thank you!
Thank you for your detailed comments. Your suggestions on the CC button are pretty much in line with what we are thinking, but we want to gather more feedback before making a definite call on the matter. The shift-select issue is totally valid – I'll look into that for sure, see what we can do. We will be adding the ability to split and merge captions in the caption panel which should give you a lot of control somewhat similar to what you are describing. If the transcript was created using the Adobe speech to text service we could potentially have word-level timing which could potentially make the kind of rolling edits you talk about posible. But what if you just imported an SRT or typed them by hand? For those, we only have timing at a caption block level so it would be a challenge to get the word-level timing correct when splitting or trimming in the sequence. For now, it is likely that we will keep splitting operations to the captions panel where you can place the cursor where you want to split.
Thank you for your reply!
You wrote "But what if you just imported an SRT or typed them by hand? For those, we only have timing at a caption block level so it would be a challenge to get the word-level timing correct when splitting or trimming in the sequence."
It would indeed be hard to get perfect results, but less than perfect results are certainly better than guaranteed results no one ever wants (i.e. a cut caption being duplicated, or rolling a cut while the text in each adjoining caption remains the same). One way to go about solving this would be to corelate the length of the caption in the Timeline to its # of words 'length'.
No matter the length and # of words in a caption, calculating the point at which a word should jump from one caption to another could be achieved with this simple formula:
CAPTION LENGTH divided by #WORDS = amount of rolling needed before a word makes the jump.
Therefore, a 10 second caption with 10 words would need to be rolled back/forward by 1+ second for one word to be swapped, 2+ seconds for 2 words to be swapped, and so forth. By doing it this way, users would be free to make small roll-edit adjustments of a few frames at a time without any word swapping occuring. The swap would only be triggered when rolling past a certain point as determined by the formula above.
The same formula could determine what words go where when cutting a caption in 2. For example, cutting a 10 second caption with 10 words between the 3-4th second would result in the first caption having the first 3 words and the second one having the last 7 words.
A really nice feature would be the ability to split a caption directly in the Text panel! With the cursor placed between any two words in a caption, the user could use a shortcut to split the caption in 2. All words before the cursor would remain in the current caption, while all words after the cursor would appear in the newly created caption. The cut in the Timeline could occur at the location of the playhead, whenever it is present somewhere on the caption being cut, or default to cutting in the middle whenever the playhead isn't over the caption.
The same concept could be used to move words from one caption to another. Rather than cutting words from caption A and pasting them into caption B, users could simply use a new 'roll-edit' tool in the Text panel to modify the wording 'cut point' between two adjoining captions. As for how much the cut between 2 captions should move in the Timeline when swapping words, perhaps a similar formula to the one above could be used.
Hopefully this makes sense. Would this always work perfectly? No. But it would almost certainly get users much closer to the final result they want compared to the way it works now.
To help you state your case to the rest of the Pr team with regards to fixing how captions get selected, here's an updated chart for v14.6:
Hopefully this red, orange and green chart can be turned to all green, for the sake of us all. Best of luck stating your case for this to be sorted out! 🙂 If there's anything else I can do to help, please let me know!
Caption length, some languages count words, some letters, f.e. in Finnish we use letters just because our compound words can become quite long: "apulaismyyntipäällikkö = Debuty sales manager", this kind of compound are not recommended to be hyphenated in captions.
That roll edit, well, it would be nice...
For exporting the captions: We NEED for TV BROADCAST the ability to EMBED the captions into MPEG2 files and h264 files. Not burn-in. Almost all broadcast televisions stations and cable use mpeg2 format. They do not want a sidecar file.
You can do MPEG2 with MXF, but you cannot do H.264 or H.265 currently and I request that functionality
Do you have any timeline as to when you can add MPEG2 output with embedded captions??
Thank you very much!
Sorry, I do not have an ETA, but the feature request is stil open and I see you added your comments to UserVoice. Thank you.
Do you have any timeline as to when you can add MPEG2 output with embedded captions??
Thank you very much!
We have been exporting video with the Mpeg-2 codec and embedded captions from premiere for years for our television broadcasts at our station. We have used Snell, Miranda ITX/Grass Valley playout servers just fine with files from premiere. XDCAM in MXF container is 50Mbit Mpeg2 video and a standard in the television industry.
The most recent problem I have run into is that I have a new religious network that wants MPEG2 with embedded captions in SD / 720x480. There is no option under MXF to do this. I have gone through them all and the size of the video is locked. I have hit a brick wall in that they will not accept anything else.
Ahh, I'm sorry we haven't messed with SD files in a long time. However although they are not technically pure Mpeg-2 streams, premiere does let me export MXF with DV25 or IMX 30 which are both essentially mpeg2 variants and accept embedded Captions.
I'd ask what playout server they are using, but i don't want to derail this thread too far.
Yeah, I tried those two but the IMX30 has the wrong file size of 720 x 486 and there does not seem to be a way to change it to 720 x 480. I had hoped that one would work.
As for SD, I was very, very surprised when they sent me the file format requirement sheet. I even called them to double-check that. They only broadcast SD! It has been a very long time since I did anything with SD too. No worries, I APPRECIATE the feedback! I just need to find a solution to this asap. I have five 30 minute shows ready to go but no way to get them encoded.