We already have text-to-text, text-to-image, and text-to-video models that work well enough. However, I honestly didn’t think we needed a text-to-sound-effects model, until now. Elevenlabs has rolled out an AI sound effects generator, and it’s out there for you to use. So, of course, I couldn’t resist the urge and had to give it a try.
AI Sound Effects Generator by Elevenlabs
To use the tool, you’ll first need to sign up or sign in (if you have an Elevenlabs account) by visiting this page. Once done, you are welcomed with a quota of 10,000 characters to make use of on the free version, with each generation consuming 200 characters. Honestly, more than enough for me to give it a try and come to a conclusion. With that, I started my fun little testing!
For every text prompt that you add, you can hit Settings underneath and take the intensity slider of the tool from More Creative to Follow Prompt. While I can’t definitely say one is better than the rest, keeping this slider somewhere in between for the most part worked better.
For each text prompt, you get four generations and can use up to 450 characters to detail a prompt. Sound effect generation takes only around 3-4 seconds, which is great.
If you’re not happy with either of the four generations, hit Generate again to regenerate. Depending on your prompt, the generated sound effects can be of up to 22 seconds. However, I couldn’t get the tool to produce that big an audio clip. Most of the clips were around a second to 10 seconds long.
In addition, you get a dedicated download button to download the sound effects directly on your device. Right beside the download button, there’s a Store in History button to help you keep track of the good generations, which you can easily access from the History panel of the tool.
Not Just a Gimmick!
With all that out of the way, it’s time to go a bit hands-on with the text-to-sound effect converter. Here are a couple of prompts that I tried with it:
- Sound of two people fighting with lightsabers with rain and thunder in the background
- Sound of someone turning the pages of a book while sipping tea
- Shuffling between objects made out of plastic with hands looking for something
- Throwing a plastic object on a wooden floor
- Softly tossing a bag onto a rubber sofa
- The sound of a pirate giving orders to his crew on a ship through a storm
- Sound of wind flipping the pages of a book
Some of these prompt requests came from the Beebom’s video team as well, editors always have some trouble finding the right sound effects. Who else to judge the quality of generations than them? And, well, they were taken aback by how cool the sound effects were.
While some of them were not exactly what we were looking for like the “shuffling between objects made out of plastic with hands looking for something” prompt, bringing the slider all the way down to More Creative helped a lot.
Meanwhile, I kept the slider in between 35-50% to get some pretty good results. Besides, from among the four generations that you get right off the bat, one is bound to match your imagination. At least for me, it did.
However, if you feed it too much information as I did with the “Sound of two people fighting with lightsabers with rain and thunder in the background” prompt, while it got the first half of it right, it completely ignored the second bit. Other generations of the same prompt gave me a hotch-potch of sound effects that could sound like a gazillion other things.
Keeping the prompt simple is the best way to go about it, my experience with it suggests.
SFX Made Easy Using AI
While generating and listening to funny sound effects that match your imagination is super fun, the practicality of it is what’s important here. As I said, our video team and I’m guessing any video editor out there will relate to the lack of reliable sources for sound effects. Being able to have control over exactly the kind of sound effect you want is, in my opinion, a big lifesaver for them.
Besides, having gotten a big thumbs up from our video team itself made me realize that, wait, this is it. Gone are the days of having to rely on third-party sources for templates. We are now moving towards an era of being in charge of it all and creating things from scratch with AI. The best thing is that this tool is completely royalty-free too.
Most importantly, the 10,000 character limit per month on the free version is more than adequate in my opinion. At least, for a free version that works this well. Another thing that I noticed is that the tool gets confused when you feed it too much information in a prompt. Instead, giving it simple prompts gets you better results, which you can then stitch together easily on any audio editor out there.
Moreover, the paid plans starting at just $5 per month don’t just get you access to the AI sound effects generator of Elevenlabs. You also get access to their AI text-t0-speech, speech-to-speech, Voicelab, Dubbing, and the complete package, which is beyond impressive. There are also annual tiers starting at $50 that give you two extra months of usage too, which is thoughtful. You can check out pricing for Elevenlabs AI sound effect generator on their official website.
All things considered, this is one hell of a cool tool that can come in very handy if you use it the right way! I have 4,472 characters left and boy am I going to go crazy on those.
What about you? Do let me know in the comments what you think about Elevenlabs’ new AI sound effects generator!
