How to use message banking for voice banking?

How to use message banking for voice banking?

Great job! You created a message bank. That’s a good solution to preserve selected phrases. Now, however, what you really want is a dynamic voice, a speech engine.

You might think the only way to create a personalized speech engine is to record a new, large set of voice samples. That’s not the case.

In fact, it’s enough that you have a set of existing recordings. TheVoiceKeeper will be able to use them to create a personalized speech engine for you. This is a wonderful solution, especially if you’ve already lost your voice, and are not able record anymore.

How to use my existing voice bank in to create a personalised voice?

There two typical scenarios here.

1. You have an existing message bank.

If you created a formal message bank in the past, you should have a good amount of audio files. Such a message bank is usually enough to be able to use the message banking to voice banking conversion feature at TheVoiceKeeper.

The process is simple.

1.Prepare your files. While not critical for the success of the process, there are some best practices you can follow.

  • Name. Each filename should reflect the content of the recording. So, if the recorded text is “My name is John” the filename will be `My name is John.mp3`. As simple as that!
  • Format. You can use any other popular audio file format, for example, .mp3 or .wav.

2. Compress the files. Convert the files into a .zip or .rar archive.

3. Upload. Upload the archive to Message Banking section on TheVoiceKeeper’s website.

When you click “Upload files”, you can simply drag and drop.

What happens next

After you’ve uploaded the files, the voice creation process is fully automatic. You will get a short email to let you know when your new voice is ready. It usually takes just a few hours.

The quality of the final voice depends on the quality and quantity of the uploaded files.


Ready to upload your audio files?


2. You don’t have an existing message bank.

You might have never banked your voice before. If that’s the case, don’t worry. It’s also a common scenario!

If do have any past recordings of your voice, we can still create a personalized speech engine for you.

Examples of recordings could be a speech or talk you gave in the past, an interview, or a recorded call. Our technology can work with all those types of files, but… you’ll have to help us a little bit.

We’ll ask you to prepare the file before uploading.

How to do it?

  1. You’ll need to cut the long recording into short segments, and save each part as a separate file. The text in each segment needs to be as close as possible to a grammatically correct sentence or phrase.
  2. It’s important that the files contain only the sound of your voice, with as little interference form other sources as possible.
  3. Once you have a list of files, you can follow the same process as described in the section above. In short, convert it into a .zip or .rar archive, and upload to TheVoiceKeeper.

An important note about quality.

Recordings that haven’t been created specifically for voice banking have several drawbacks. Most common of them are that they can often include echoes or environmental noises. These interferences can make your voice harder to hear and, consequently, can severely downgrade the quality of your synthesized voice.

There is unfortunately nothing we, at TheVoiceKeeper, can do to improve these audio files— we can’t remove echoes or noises.

The quality of the synthesized voice that you receive, is dependent on the quality of input.

On the other hand, since it costs you nothing to try the system, we recommend to simply…try it with the materials you’ve got!

If the quality of the voice doesn’t satisfy you, you don’t need to buy it.


What if I don’t have existing recordings?

Don’t worry. It’s ok if you don’t have any past recordings. It may also be that the ones you have aren’t of a high-enough quality. That’s not a problem either.

Both of these problems can be addressed with Voice Donation, an option we offer at TheVoiceKeeper. It’s a simple way of asking trusted individuals—friends, family—to record their voice for you. You can read more about Voice Donation on our blog.


Easier alternatives? Yes, but with caution.

It can take a lot of time to prepare a long file for upload. Cutting and renaming short segments can be demanding and tedious.

We can suggest two options to avoid it, but we want to stress that taking an easier way can take toll on the quality of the final voice.

Upload the full recording, without cutting. While it’s better to create short, one-sentence files, if it isn’t possible, you can upload one long recording. TheVoiceKeeper will use voice recognition technology to split it into sentences.

However, because this process is not as accurate as manual file cutting, it will influence the quality of the created voice.

Skip file naming. While it’s better to name each file with the text spoken in the recording, you can also name the files arbitrary. Just like in the case above, TheVoiceKeeper will use voice recognition technology to transcribe the text and automatically name the files.

The higher the quality of your recordings, the better our voice recognition will perform in cutting and naming the files.

That’s it! We hope this post has given you clear steps on what to do next. We’re looking forward to helping you create a high-quality personalized voice.


Creating a personalized voice with TheVoiceKeeper is free.

Try today, no strings attached.


Cover photo by Oleg Laptev on Unsplash