ElevenLabs Review: Sounds Amazing — But Not Perfect

Ever really feel like your content material is lacking that particular one thing? Like your phrases are simply… falling flat?

Within the age of TikTok and podcasts, audio is king. However let’s face it — not all of us are blessed with a voice that would soften butter. And hiring voice actors? That is an entire different headache (and expense) most of us would quite keep away from.

And that’s why AI text-to-speech providers are rising in popularity. ElevenLabs appears to be the subsequent in line in AI audio technology. As somebody who’s examined extra AI instruments than I care to confess, I used to be skeptical. However I wished to provide it a strive, and let me let you know, I favored it. Quite a bit.

So, what made me change my thoughts? Let’s speak about it.

What’s ElevenLabs?

Ever wished you may have Morgan Freeman narrate your grocery record? Effectively, ElevenLabs may not get you that far (but — although they do have Deepak Chopra), however they’re actually pushing the boundaries of what is potential in AI-generated speech. We’re speaking about an AI firm that’s making content material accessible in any language, with any voice you’ll be able to think about.

ElevenLabs is extra than simply taking part in round with robotic voices. Their analysis group has AI fashions that may create (or extra appropriately, voice out) reasonable and context-aware speech throughout 32 languages.

Their essential characteristic is their text-to-speech fashions (which we’ll speak about intimately later) however in addition they have fashions for altering your voice, sound results, and audio isolation. They’re additionally giving API entry to their fashions if you wish to use it for content material creation and combine it with different apps.

Who’s ElevenLabs For?

When you’re a content material creator who’s uninterested in listening to your individual voice or making faceless content material, ElevenLabs must be a fantastic various to hiring voice-over actors. It doesn’t matter when you’re on YouTube or TikTok, ElevenLabs gained’t copyright strike you for utilizing one in every of their voices.

Nevertheless it’s not only for content material creators. Recreation builders and indie filmmakers might additionally profit from utilizing ElevenLabs. Think about having the ability to prototype character voices with out hiring a single actor, or localizing your recreation into 32 languages with out breaking the financial institution.

And when you’re an creator or journalist, ElevenLabs might additionally flip your articles or books into professional-sounding audiobooks. College students may also use ElevenLabs to make their displays and movies extra fascinating.

How Does ElevenLabs Work?

I’ll be sincere — I haven’t actually used any audio AI fashions earlier than ElevenLabs, however I didn’t actually wrestle utilizing their platform. This speaks quite a bit (no pun supposed) to their user-friendliness. 

Right here’s how their text-to-speech software program works:

You simply want an enter script — stuff that you really want the robotic to say. Actually something beneath the solar. However since I don’t actually have a script readily available, I’m going to make use of ChatGPT to write down a brief true crime story excerpt. 

Now, I’m simply going to repeat that and paste it into ElevenLabs’ text-to-speech subject.

All you must do now could be choose a voice that you simply like most from their choice (there’s quite a bit they usually additionally embrace a brief observe of what the voices are finest for) and press “Generate Speech.”

Right here’s a fast pattern of what it seems like. 

Utilizing their settings, it’s also possible to select how secure you need the output to be (extra secure sounds extra robotic, much less secure is extra emotive however can typically sound glitchy), similarity, and elegance exaggeration. By tweaking a few of these settings so as to add some exaggeration and emotion, I feel I’ve hit the candy spot on this model utilizing these settings.

Oh, and after I say that you are able to do all kinds of issues, I imply it. I’ve been making an attempt their voices out all day and I’ve been having enjoyable testing them in numerous eventualities. Right here’s an ASMR pattern.

Right here’s a dungeon grasp introducing his new steampunk-themed world.

Or perhaps I might curiosity you with an audiobook narration?

And like I mentioned, it’s multilingual, so you’ll be able to feed their mannequin scripts in numerous languages (French, Italian, German, Filipino, Spanish) and also you’ll nonetheless get a high-quality recording.

The extra I exploit ElevenLabs, the extra I prefer it. Don’t get me mistaken — their outputs nonetheless have that “uncanny valley” feeling to extra discerning ears, however for most individuals, I don’t suppose they’ll be capable to differentiate it from common audio system, particularly if there’s ambient sounds and background music taking part in.

Options of ElevenLabs

Textual content-to-Speech

ElevenLabs’ Textual content-to-Speech characteristic is not your on a regular basis TTS. With 32 languages and greater than 40 voices to select from, it creates eerily human-like performances. It isn’t simply studying textual content — it is bringing it to life with tone and cadence. Good for turning blogs into podcasts or giving voice to your newest story.

One factor I seen although is that it typically cuts the primary phrase of the immediate. This appears to be a glitch because it solely occurs round 50% of the time in my expertise.

Voice Changer

Wish to sound like actually anybody else? ElevenLabs’ Voice Changer can try this for you. It is like having a vocal shapeshifter at your fingertips. Content material creators can voice a number of characters with out hiring a solid. Novice filmmakers can create a complete animated collection solely utilizing this. There’s numerous potential in utilizing this characteristic.

Or so they are saying. So, I examined it. Right here’s my very own voice:

And right here’s the output utilizing one in every of ElevenLabs’ voices:

One factor I like about it’s that it doesn’t simply change your voice, it fully captures the context of what you’re saying and makes use of that to affect how the output sounds with out straying away from the way you mentioned it.

Sound Results Creator

Like I mentioned, there’s slightly little bit of one thing for everybody with ElevenLabs. For sound designers, the Sound Results characteristic creates customized results in seconds. No extra Wilhelm Scream and no extra looking for hours on finish for the correct audio. Now you can let ElevenLabs create it for you as a substitute.

For every immediate, ElevenLabs will generate 4 completely different results so that you can select from. That is my favourite utilizing the immediate “creepy footsteps from afar.”

I’ll say although — out of the whole lot ElevenLabs has to supply, this one left me probably the most unimpressed. It’s good for brief prompts, however after I tried utilizing a immediate with plenty of context or one thing that’s already layered, it ended up ignoring some elements of my request. Right here’s one I made for “the sound of waves on a crowded seashore.” 

Voice Isolator

When you don’t have the cash for an expert mic setup, this one’s for you. ElevenLabs’ Voice Isolation characteristic removes the background noise from an audio enter. As somebody who’s been utilizing Adobe Premier’s audio clean-up options for video enhancing, I can truthfully say that ElevenLabs’ shouldn’t be solely simpler to work with, but in addition quite a bit cleaner.

As soon as once more, and I apologize for this, right here’s my voice:

And right here’s its remoted audio:

ElevenLabs’ Pricing

I’m going to reply the query that I’m positive is already in your thoughts: sure, there is a free tier. It is fairly restricted — but it surely does offer you 10,000 credit to work with. So, how does ElevenLabs’ credit work? I’m not too positive with their voice isolator and changer (for reference, a 6-second clip prices 93 credit) however for the opposite options:

• Textual content-to-speech: one credit score per character.
• Sound results: 320 credit per immediate.

For the extra severe customers, right here’s an summary of what they provide per tier:

The Professionals and Cons of ElevenLabs

So, What’s The Verdict?

As somebody who’s new to the AI audio technology world, I had numerous enjoyable utilizing ElevenLabs. It’s efficient, fast, user-friendly, and reasonably priced — what extra are you able to ask for?

Seems, the reply to that query is perfection. 

Right here’s the factor: if this was launched 4 years in the past, I’d don’t have any points with it. However we’re now residing in an period the place AI is an on a regular basis factor, a lot in order that we’ve grown accustomed to it. We all know what’s written by AI or not, what’s drawn by AI or not, and (because of hundreds of hours of faceless TikTok content material) what’s mentioned by AI or not.

ElevenLabs’ audio high quality is close to human, but it surely’s simply not there but. And we, as people, can inform. 

So, do you have to use ElevenLabs? By all means. In spite of everything, like I mentioned, it’s fairly wonderful. However when you’ve got entry to assets that permits you to rent an individual as a substitute, that can at all times be the higher possibility. No quantity of AI development can exchange human emotion and expertise.