A new Speech to Speech tool that can change the voice and even clone someone else’s voice using an additional VoiceLab tool has been added to the elevenlabs.io neural network. Let’s see how it can be applied in affiliate marketing and traffic arbitrage on the example of working with creatives.
Examples of voice change
Neural Network can be used for voice acting of the creatives from scratch. There is no need anymore to hire actors, or to replace the voice of existing creatives. Below there are a couple of examples.
Text to Speech + voice cloning example
Here you can listen to the cloned original Morpheus voiceover and Text to Speech speech synthesis. We also used the deepfake technology from this article: https://cpa.rip/en/services/facefusion-deepfake/.
Speech to Speech example
Voice replacement for the creative:
- 00:00 – Original voice.
- 00:34 – Voice 1.
- 00:55 – Voice 2.
Elevenlabs neural network features
Elevenlabs neural network is a complex system for working with voice: it can clone (VoiceLab), synthesize speech (Text to Speech), and dub into different languages (dubbing).
Our review of Elevenlabs.io’s dubbing neural network, which translates and dubs video and audio recordings into different languages while saving the original voice: https://cpa.rip/en/services/elevenlabs-dubbing/.
Speech to Speech voice-changing neural network is an addition to the existing Text to Speech tool, which greatly expands the capabilities and scope of the Elevenlabs neural network.
Eleven Multilingual v2 currently supports 29 languages, including Russian:
There is also a large number of ready-made voiceovers. Each of them has tags with information about the language, style, and intonation:
Tariff plans
- Free – 0
- Starter – $5 ($1 for the first month) +includes quick voice cloning.
- Creator – $22 ($11 for the first month) +includes professional voice cloning.
- Independent Publisher – $99
- Growing Business – $330
The free tariff plan includes 10,000 characters (credits) to generate.
For payment you can use a card issued by card services mentioned in our article: https://cpa.rip/en/finance/bank-cards/.
Working with Text to Speech
To synthesize speech from random text, use Text to Speech mode.
- Select “Text to Speech”.
- Select one of the standard voices or your own in the VoiceLab tool.
- In the “Text” field type the text for the voiceover.
- Click “Generate”.
Generation takes only a few seconds. The finished voiceover can be listened to and downloaded in mp3 format by clicking the download button:
With each new generation, the accent and intonation change slightly, so you can repeat the generation until you get the desired result. You can also change additional settings:
Working with Speech to Speech
Speech to Speech is a new and more interesting tool, it changes the voice of the original voiceover.
- Select “Speech to Speech”.
- Select one of the standard voiceovers or your own in the VoiceLab tool.
- Upload a sound or video file in the “Audio” field. The maximum weight is 50Mb. Or record your own via microphone. For best results, use clean audio with no extraneous background noise.
- Click “Generate”.
The generation also happens in seconds and you can immediately download the changed voice.
Copy voice in Elevenlabs
Let’s move on to the most interesting tool – VoiceLab. Here you can quickly copy someone else’s voice ( DeepFake voice), synthesize a voice by setting various parameters, or use voices from the huge community library.
Let’s take a look at how to clone a voice. Go to the VoiceLab page by clicking “Voices” or “Add Voice” in the left side menu near the list of standard voices on the Speech page:
- The VoiceLab tool is opened
- Click “Add Generative or Cloned Voice”
- Previously created voices can be used by clicking the “Use” button or by selecting from the list of voices on the Speech page.
- Select “Instant Voice Cloning” for instant voice cloning
- Enter the name.
- In the field, upload a reference of the voice you are going to clone or record your own via the microphone.
- Fill in additional text description if necessary.
- Agree to the terms and conditions and click “Add Voice”.
The voice recording must be at least one minute, free of extraneous noise and the uploaded file must be no bigger than 10Mb. We tried cloning a voice based on a 4-second recording and got a good result too.
The cloned voice can be used for a voiceover in Text to Speech or for changing the voice in Speech to Speech.
Do not use cloned voices for illegal or fraudulent purposes or if it violates copyright.