I am very new to this forum. So, please excuse me if I’m not in the correct forum section.
Is it possible to hire an audio developer from this forum?
I need to hire a developer to create a plugin to process narration audio.
I will gladly pay for the plugin and also provide the plugin for free for others to use.
Here is the requirement. You see, I’m not a gifted narrator. So, when recording narration, I often repeat words. A 3-minutes final audio can be as long as 15 minutes due to the repeated words. It is very time-consuming to delete these repeated words.
Basically, I need a plugin that will scan the audio backward. If it finds repeated words, it will keep the last one and remove all the others or the previous ones. Because usually, the last one is the best audio that I want to use.
Is this hard to do? And how much will it cost and how long?
As described, yes it’s extremely hard to do, and would be very expensive.
The major problem is that it would require some form of speech recognition to know when you are repeating a word. Speech recognition is very hard to do - Google, Microsoft, Apple and Nuance (Dragon software) have each spent $millions and thousands of developer hours developing their speech recognition technology.
There’s lots of companies / people that will happily do the entire job of narration and editing for a fee. For example, Amazon runs a market place for narrators, producers and authors, which is called “ACX” (Google it).
Audacity is good software for doing it yourself, but producing high quality narration is, as you have discovered, is very time consuming. As a rule of thumb, editing typically takes around 10 times longer than the running time.
This is partially manual. Sorry. But it’s something and you replied to my thread so I may as well offer it.
Get a youtube and google account. Convert audio to video using https://www.onlineconverter.com/audio-to-video . Upload to youtube under your account. Set language to English. Upload and set to unlisted. Wait a while. View video. Go to transcript. Control-a, control-c. Go to docs.google.com. Open new document. Control-v (paste). Duplicated words should show up in red along with the timestamp. Now you need open two version of youtube with your video. One for editing and one for the transcript to give you the timestamp. Manually go to the editable video, and click on the dupe 1 area of the transcript by searching in the other window for that entire line with the dupe. Magnify, highlight, and snip. Go to preview. Do next dupe. But, it’s manual. Also if you are narrating this will only delete not silence. I am unsure if youtube allows muting yet.
It’s worth shooting nuance this question since they purchased Dragon Naturally Speaking and that’s where the feature should be.
Another idea is to use the edited google docs file as a source for a TTS engine.