Adobe Updates Premiere Pro for NAB 2025

If you edit using the Adobe Creative Suite, you’ve doubtless heard about the updates released just prior to NAB 2025 (officially the 25.2 release). This article notes the key new features available in my video editor of choice, Adobe Premiere Pro.

Generative Extend

Generative Extend lets you extend video clips by up to two seconds and audio clips by up to ten seconds. To use it, select the new Generative Extend tool on the toolbar, and click and drag the end of a clip to extend it (Figure 1). The tool uses artificial intelligence, specifically Adobe’s Firefly Video Model, to analyze elements like the scene composition, motion, and audio characteristics of the original footage. It then generates new frames and sounds that closely match the original in style, ambiance, and continuity.

It’s not a clip generation feature as much as an alternative to rate-stretching your clip, which slows the frame rate. You see the feature in Figure 1 (click to see the image in full screen). You select it in the toolbar on the upper left, then click and drag the end of the clip.

To test this new feature, I applied it to four clips. Twice, I did what you see in Figure 1, which allowed me to compare Generative Extend (on the left) with the actual additional frames in the actual clip. For those who haven’t had their morning caffeine, this means that I:

  • Added the clip to the timeline twice
  • Trimmed off two seconds from the top clip while leaving the original frames on the bottom clip
  • Extended the top clip with Generative Extend
  • Cropped and moved the frames to compare Generative Extend with the original frames.

Two other times, I applied the filter and just eyeballed the results.

Figure 1.  Premiere Pro’s new Generative Extend feature (click to see the image in full resolution.

In two of the four cases, the result were quite good, the two seconds generated by Premiere Pro looked very natural, and it was tough to tell when the real clip ended and Gen AI started.

I tried to extend a soccer clip by two seconds and found the motion artificial and a bit jarring. I tried to apply the effect to Sintel, the animated movie, and got a “Can’t generate extension” error message. If you find yourself in need of those extra two seconds, particularly for a transition that might hide any unnatural movements or pixels, it’s definitely worth a shot.

That said, Generative Extend isn’t free, at least not for long. It’s available through standalone Firefly plans or in select Adobe Creative Cloud business plans. Adobe offers a limited number of complimentary uses for Generative Extend, but continuous access requires a subscription to one of these plans.

Here’s Adobe’s video on how this feature works.

AI-powered Caption Translation

I had better luck with Adobe’s AI-powered captions, which can translate subtitles into 27 languages in seconds. Auto speech to text is one of Premiere Pro’s most valuable features, and this extends it tremendously for those serving multilingual audiences.

To use the feature, you create your captions as normal. Then, then in the Captions tab, click the new Translate Captions button and choose your language(s). Premiere Pro puts each language on a separate track which you can export for uploading into your content management system or YouTube, or burn into the video (Figure 2; click the figure to see it at full resolution).

Figure 2. Translate your English captions to 27 different languages. Click to see at full resolution.

To check the quality of the German translation, I sent a short clip with German subtitles to my German daughter in Bonn. She responded that “everything made sense and was properly translated.” Tough to generalize this to 26 other languages, but AI has been working this problem for a few years now; I’m sure the results are competent today and will be better tomorrow.

Still, what’s suitable for social media may be inadequate for mission-critical business communications or premium content. Be sure to match your level of review to the importance of the content.

Media Intelligence

Adobe’s new AI-powered Media Intelligence makes it easier to find clips in complicated projects, and will be a godsend for many producers. Working via the new Search tab shown on the upper left in Figure 3, this feature automatically identifies the content within clips, including objects, locations, and camera angles. You can search using natural language like “animals,” which picks out real and animated animals (and not humans) in the project.

Figure 3. A new Search panel makes it easier to find clips in clip-heavy projects. Click to see at full resolution.

This tool also recognizes spoken words and metadata like shoot dates or camera types, and all your transcriptions. In Figure 4, I’m searching through the transcript to find all mentions of the word injunction in this single clip. This is a killer feature if you’re working with interviews and need to identify relevant sections.

Figure 4. Media Intelligence found all mentions of injunction in this interview.

Media Intelligence requires no manual tagging. The analysis is fast and local, so it doesn’t require an internet connection. Organization for even moderately complex projects is always a bear, and this feature should reduce that significantly.

Here’s a YouTube video that explores this feature.

Other Features

Other features were less applicable to how I use Premiere Pro, but included Updated Color Management, which automatically transforms log footage from nearly every camera into HDR and SDR.

For a full list of new features, click here. These include:

  • Added support for import of MKV (H.264/AAC) files
  • Add Content Credentials into exported video files
  • Dynamic audio waveforms
  • Hardware acceleration for Canon Cinema RAW Light
  • Support for ARRIRAW SDK 8.3.1
  • New and updated sample media

About Jan Ozer

Avatar photo
I help companies train new technical hires in streaming media-related positions; I also help companies optimize their codec selections and encoding stacks and evaluate new encoders and codecs. I am a contributing editor to Streaming Media Magazine, writing about codecs and encoding tools. I have written multiple authoritative books on video encoding, including Video Encoding by the Numbers: Eliminate the Guesswork from your Streaming Video (https://amzn.to/3kV6R1j) and Learn to Produce Video with FFmpeg: In Thirty Minutes or Less (https://amzn.to/3ZJih7e). I have multiple courses relating to streaming media production, all available at https://bit.ly/slc_courses. I currently work as www.netint.com as a Senior Director in Marketing.

Check Also

How AdCurate Helps Publishers Improve Ad Quality and Boost Revenue

In this conversation, Jan Ozer from the Streaming Learning Center interviews Gabe Thomas, VP of …

How Witbe Measured Super Bowl Streaming Performance — Insights on Latency, QoE, and 4K Quality

On March 11, 2025, I (Jan Ozer) from the Streaming Learning Center interviewed Yoann Hinard, …

Review of Multi-Resolution Encoding for HTTP Adaptive Streaming using VVenC

In their paper entitled, Multi-resolution Encoding for HTTP Adaptive Streaming using VVenC, Kamran Qureshi, Hadi …

Leave a Reply

Your email address will not be published. Required fields are marked *