Ahh Youtube, the world’s largest library of video. An astonishing collection of humankind’s creativity and knowledge, including everything from how to grow an avocado tree to insanely catchy songs about people who live in the forest and wear tight clothes. If you want it, it’s on YouTube.

Sometimes it can be a little too much though. If all you want is a quick recipe for Pad Thai, sitting through a monologue about a YouTuber’s trip to Bangkok and five minutes of sponcon from a meal-in-a-box vendor is a little annoying. Google recognised this and introduced highlighted clips within their search results to take people straight to the clip they need, wherever their algorithm can identify a helpful video.

But what if you want to save this clip, download it, or remix it for social? Well, you’ve got a few options.

Creating clips

If all you want to do it create some clips for sharing elsewhere, you’ve got plenty of options.

YouTube’s built-in clipping

If it’s enabled on a YouTube video, you can create a quick clip right from the YouTube interface. Hit the “Clip” button, choose your start and finish times (either with timestamps or by dragging the handles), and you’ve got a quick clip ready to go. Here’s a clip I created from Jose’s recent video on humans and AI working together.

However, as with video transcripts on YouTube, you’re a little limited with what you can do with this clip. You can’t download it, you can’t remix it, you can’t edit it any further. Youtube wants you to stay on their platform to watch more delicious advertising.

Recording from your computer

A more DIY option is to record the clip locally using screen recording software. If you’re on a Mac you can use Quicktime (although you’ll need to use something like BlackHole to record your system audio – good explanation here), and on Windows you have a plethora of options.

This approach is infinitely flexible, and has the advantage of giving you a hard copy of all your clips, but requires a little more fiddling than browser-based solutions.

Recording from your mobile

Another simple and quick option (especially for people who edit videos primarily on their phone for TikTok or Instagram Reels) is to record your clip using your phone’s built-in screen recorder. Both Android and iPhone will let you screen record with audio easily (Android instructions here, iPhone here).

The most important thing to be aware of with this method is, depending on the options you choose and your phone operating system, you may record your entire screen. That means you’re not just creating a clip of your YouTube video, but also your notifications and your battery level, which is probably embarrassingly low.

You can crop the video once you’ve finished recording, but this can take a long time due to the limited processing power available on mobiles.

Editing for social

If you want to use your clips in your Tiktoks, Reels, or somewhere else, the next step is to edit them. Most YouTube videos are in horizontal, or 16:9, format. Conversely, most social apps are built for mobile screens and so use vertical, or 9:16 video. So, you need to crop them.

Pro editing software

If you have access to it, something like Adobe Premiere will let you make all the edits you need and them some. Adobe is doing a great job of infusing AI features into Premiere, so you can use the auto-reframe feature to switch you video to landscape and keep the subjects centered.

However, Premiere is a pro tool, built for pro people. So if you’re just getting started with video editing it can be a little (or very) intimidating. A nice middle-ground is Premiere Rush, a stripped-down version of the editor built specifically with creators in mind.

Consumer apps

For more basic editing tasks – cutting montages together, adding music etc. – there are plenty of free apps that will get you most of the way there.

iMovie for Mac, Lightworks, and plenty of others are a great way to get started with editing, and won’t cost you a penny. However, each has its limitations and none are built specifically for recutting clips for social channels, so won’t have AI reframing, auto-captions and all the nice features you’ll get in a pro package.

Dedicated social clipping apps

If you’ll be doing a lot of clipping, cropping and reframing for your social channels, you should check out a dedicated social clipping app. These don’t come with all the bells and whistles of a full editing suite, but have features specifically built for social video. Cropping from horizontal to vertical, AI reframing, captions, and file export in popular formats.

And, as you’ve probably guessed, Imaginario AI is one such platform! In fact, we can handle the whole process of taking a video from YouTube and clipping it either for download or for social channels. Simply import a YouTube video directly from your channel, choose your clip, and reframe. Give it a try, you can create 5 clips per month totally free!

---

---

Article credits

Posted by

Jon Norris

Jon has been working in video, journalism, marketing and social media for over 15 years.

Originally published on

June 4, 2024

Filed under

How To

With image generation from

OpenAI

And TikTok creation from

Imaginario AI

Sure, watching a video is nice, but sometimes it’s just easier to work with text. Maybe you need a quote for a paper you’re working on, you’re trying to work out some song lyrics, or you’re just the kind of person who prefers reading. Transcripts are great, we love transcripts.

Unfortunately YouTube – the place that 9 out of 10 adults watch the majority of their online video – does not offer video transcripts, even though they auto-generate a transcript for every single video uploaded to their platform. This is because those transcripts are used for subtitles; YouTube doesn’t want you to have access to the full transcription in a useful format, they want you to watch the videos (so they can show you ads!).

Aside from the subtitles, YouTube also allow you to access a video’s subtitles as a navigational aid. Click “…more” to expand the video’s description, and find the “Show transcript” button to open up a side panel with timestamped lines from the video. Click on any line to jump to that section of the video. Unfortunately there’s no way to export this text other than manually copy and pasting it, so this is where your built-in transcript download options begin and end with YouTube.

But, if you want to download the full transcript of a YouTube video in a useful format, for whatever reason, you have a few options.

The traditional way – commission a transcript

For those of us who have gotten used to automation and AI this might seem incredibly quaint, but there are still lots of services out there such as OG online transcribing service Rev which create transcripts the old-fashioned way, by getting a human to watch the video and write down what they hear. But wait! There are good reasons for this approach. AI is maturing fast, but there are some things it can’t do just yet.

For example if you want captions including audio descriptions, not just subtitles, humans are still the best way to go. AI can’t figure out what is contextually important for storytelling, so although it might be able to correctly identify audio within a video, it won’t know if wind howling in the distance or a car door closing is an important story beat, or just some random background audio.

Similarly, if you want your subtitles translated into another language, a human translator could be a better bet than an AI due to cultural differences and context. An English speaker might describe an easy task as “a piece of cake”, but an AI won’t understand the metaphor and will translate literally, describing the task as “un porción de pastel” for Spanish viewers, which won’t make any sense whatsoever.

So, in many cases, human transcription still makes sense. Or, as is becoming more common, a first pass with AI then a manual review to correct errors by a living, breathing human.

The fast way – generate an AI transcript

If what you’re looking for is an English-language, dialogue-only video transcript, by far the easiest and cheapest way to get this done is with an AI tool. In fact, if the video you want to transcribe is less than 30 minutes, you can do it completely free with Imaginario AI.

AI YouTube transcripts are generated in much the same way as YouTube’s own transcripts, using a voice recognition model to turn speech into text. Depending on the subject matter, you may find a third-party transcription generator gives a more reliable output than YouTube’s own model, which tends to struggle with acronyms, strong accents or technical language. “ChatGPT” often becomes “Chat gbt” and “JSON” almost always becomes “Jason”. And, importantly, a third-party service will let you download the transcript in a format you can use!

Generating a YouTube transcript in 30 seconds

Lets see just how fast we can generate an AI transcript. One of the secrets of AI transcription is that the video is processed faster than it can be watched. AI systems can digest all of the dialogue as a single input, free from the linear passage of time that shackles us fleshy humans. Lets see how we do.

The process is, as the Spanish say, un porción de pastel.

Step 1: Register for an Imaginario account

I’m sure I don’t need to explain this one. Just get to it.

Step 2: Import your video

You can import videos into our platform through file upload or direct connection to your favorite cloud storage, but for the purposes of this experiment we’ll be importing directly from YouTube. All you need is the YouTube link.

Step 3: Choose speech processing

We can also recognize visuals and audio in any video, but if we’re just making a transcript, we only need to process the speech. This saves us a bit of processing juice and saves you a bit of time.

Step 4: Download your transcript

Once your video has finished processing, choose Download -> Subtitles, with your file format as text and hey presto, you’ve got a 99% accurate transcript of a YouTube video in less than a minute!

---

---

Article credits

Posted by

Jon Norris

Jon has been working in video, journalism, marketing and social media for over 15 years.

Originally published on

April 8, 2024

Filed under

How To

With image generation from

OpenAI

Photoshop

And TikTok creation from

Imaginario AI

As a former critic and journalist, I have my own growing collection of classics and evergreen films and TV shows. When I have time to kill, I tend to sift through and re-edit famous scenes to create inspiring clips and build a deeper connection with my audience (if and when I decide to publish these videos on TikTok ).

For example, when discussing themes of entrepreneurship, resilience, and managing a startup, I often find inspiration in scenes from films such as “The Founder” (2016) or “Persuit of Happiness” (2008) or from inspiring figures like Muhammad Ali, Elon Musk, Bruce Lee, Warren Buffet, and Marc Andreessen. Similarly, producers and filmmakers in the process of crafting a new script or re-editing scenes for social media are on the lookout for recurring themes, motifs, and artistic styles.

Searching and rewatching an extensive video catalog with the goal of finding inspiration in pre-production, recycling approved B-roll content for new projects, finding relevant scenes coming from dailies, or creating high-quality compilations for social, can take hours, even days. This is where advancements in video search and analysis technology come into play.

With Imaginario AI, creative teams and production companies have access to an end-to-end cognitive media platform with robust video search capabilities that understand context as a person does: across senses (vision, speech, sounds) and by understanding the passage of time. With a simple text or image query, Imaginario AI can locate exact moments needed for a project in seconds without the need for labels or technical complexity.

For example, you might be looking for a specific scene, such as Tom Cruise running shirtless on the beach, or perhaps finding the moment when the underdog football team wins a match. It doesn’t matter: we can find it. Imaginario AI is not only streamlining the creative process but also opening up new possibilities in post-production.

Use cases creators and production companies are tackling with Imaginario AI

• Finding people, objects, actions, emotions, sounds, topics, keywords, and more across your entire library or a subset of it (specific folders, collections and videos)
• Locating and recycling b-roll content for new projects (and save money in stock footage in the process!)
• Adding subtitles and converting horizontal videos into TikToks, IG Reels, and more in minutes. You can also export your clips to Adobe Premiere for further customization
• Locating any logos and text on screen to monitor brand exposure
• Discovering contextual ad inventory to optimize campaigns
• Identifying sensitive scenes to adapt video content to different markets, platforms, and audiences (e.g. violence, adult scenes, drug abuse, etc)
• Moodboarding in pre-production

Just tell me how it works

1. Create an account here. Use a Chrome or Chromium browser to log in.
2. Video upload: Upload or ingest your videos to Imaginario AI via AWS S3 private buckets or your Google Drive account. Watch as Imaginario AI indexes them in minutes. Say goodbye to pre-defined and inflexible labels, complex queries, and begging your technical team to update your video search tags and systems for more accurate results
3. Search: Go to the Search section and select the type of query you want to pursue: visual, audio, or speech. You can use any type of text query, or you can upload any image to search for specific faces, objects, and more in your videos
4. Save your clips or add them to a collection
5. Download or re-edit clips. You can download them in your drive, re-edit them on Imaginario’s Clip Studio (captions, resizing) or export your project to Adobe Premiere.

---

---

Article credits

Posted by

Jose Puga

Originally published on

March 17, 2024

Filed under

How To

With image generation from

OpenAI

And TikTok creation from

Imaginario AI