Text to Speech (TTS) awareness

[Mistermann]

How many of you wish you could put together compelling missions with deep voice acting, but just don't have the bandwidth to gather friends & family to voice all the parts?  Or maybe you've experimented with more traditional text to speech (TTS) tools resulting in voiceovers that sound like the WOPR from that old movie "Wargames"? 

Well, if you have, this thread is for you.  Keep reading.

TLDR
Give AI TTS tools a try if you haven't yet.  Elevinlabs works great for me. https://elevenlabs.io/ 

Details

For those still with me, thanks.  

A little background on me.  I am a LONG time DCS guy with a decent command of the ME.  I've made hundred's of missions over the years 99% of which  ever get posted/shared.  Not because I am selfish, but because I just can't create the kind of mission that I believe stands out from all the other posted missions.  Missions that stand out to me are those that tell a story.  Missions that are experienced, not simply just flown or played.  That experience is often heavily influenced by voice acting, radio traffic and creative ME randomness.  To me, that's what often separates the memorable missions from the average ones that we all forget after a couple days. 

Until recently, I thought there were just two Voice Acting options for us mission designers/makers:

1) Leverage volunteer voice actors to get a variety of content and eliminate hearing the same actor over and over, or

2) Use TTS (Google, Microsoft, others) to create a limited set of voices that if I'm honest about it, still sound artificial and are stoic and rigid.

As someone who's lent their voice to several official DCS campaigns and understand some of the planning/work/effort/patience, I can say I just don't have the time, nor the social network, to go with option 1.  I've never really experienced a TTS mission that immersed me.  I really try to overlook the robotic sounding audio content because I know how hard the creator worked to even try TTS.  

I'm just always looking for a middle ground. Something that sounds real, but doesn't require the work that your favorite campaign creator puts in ... I need something "good enough" to immerse and "better" than traditional TTS.

Here's my answer ... AI voice tools. https://elevenlabs.io/ 

Please note, I do not represent this service in any way whatsoever.  I am a happy customer - that's it.  If you're interested in rich voice acting without the need for actual actors, here's a compelling option to consider.

I've subscribed for almost a year and during that time have seen significant improvements to their offering.  I was amazed by their initial offering of TTS, but have been absolutely BLOWN AWAY with their latest feature they call, Speech to Speech.

Now, the proof here is in how these voices sound, so let's jump to a different medium where you can actually hear what I am talking about.  For those that have interest, I direct you to a brief video below.  

Hope you find this information useful.

 

Edited January 9 by Mistermann
Formatting

[rob10]

Elevenlabs has been recommended before and it really is amazing what it can do.  Another happy user.

[Mistermann]

It definitely has, but I had never seen a short video showing the capabilities.  I am hoping this gives people an idea of the power outside of just reading it on a forum.  Plus the fairly new Speech to Speech addition has been a game changer for me and I wanted to socialize.

 

[Priest]

This is pretty cool. Anyone try to recreate Wags via speech-to-speech yet haha

What's a good tool for adding some effects? Radio static, gunfire, like in yours. "Good" in this case means easy to use and ideally free.

[Elphaba]

I've been using it for 2 years for my missions and my Superscript, but tbh, this kind of thing should be built into DCS.

Edited March 26 by Elphaba

[Mistermann]

11 hours ago, Priest said:

What's a good tool for adding some effects? Radio static, gunfire, like in yours. "Good" in this case means easy to use and ideally free.

Audacity.  And yes, it's free. It will take a little time to learn but I don't find it overly complicated.  There are lots of videos explaining how to do what you describe.  Have fun.  The rabbit hole is deep.

[cfrag]

On 3/26/2024 at 1:45 AM, Priest said:

What's a good tool for adding some effects? Radio static, gunfire, like in yours.

I'd recommend Audacity (free and complete) or Logic Pro (Mac only, Pro-Level).

On 3/26/2024 at 1:45 AM, Priest said:

"Good" in this case means easy to use and ideally free.

Be advised that much like using a good video editor, you can't do much with an Audio Editor by itself: you need access to some audio source (like samples for gunfire etc). There are great sample libraries available for next to no cost, and if you purchase one of these libraries you have one incredibly useful thing: peace of mind that you own the rights to use them and aren't threatened by DCMA takedowns (and resulting litigation) should your mission prove to become popular. A library of some 5'000 production quality sound effects should cost you around USD 20, which I find to be acceptable (the price of three decidedly average coffees where I live). And while there are very good free editors (Audacity), you often do get what you pay for. I recommend you go with free first, and only go paid when you find that you enjoy doing audio production or want to create an income stream

 

 

[Rudel_chw]

On 1/9/2024 at 7:20 PM, Mistermann said:

Hope you find this information useful.

for some reason, Its only now that I see this thread, I’m currently a user of https://ttsfree.com but I will give this AI option a try, maybe I can use it on my future missions  

thanks a lot for bringing it to my attention  

[Mistermann]

11 hours ago, Rudel_chw said:

for some reason, Its only now that I see this thread, I’m currently a user of https://ttsfree.com but I will give this AI option a try, maybe I can use it on my future missions  

thanks a lot for bringing it to my attention  

Glad you found some value here, Rudel.  My mission creation quality has really gone through the roof (IMO) because of this AI capability.  I now have limitless voice actors on call who are willing to say whatever I want, in the exact same tone/inflection that I desire.  No more WOPR sounding voices in my missions.

For example, I recorded this the other day for a campaign I am working on using my voice (which sounds nothing like this recording).  I ran my recorded voice through 11labs using one of their premade voices.  A couple of quick adjustments later (using Audacity) and I have this.  Good enough for what I need.

 

 

[Elphaba]

2 hours ago, Mistermann said:

Glad you found some value here, Rudel.  My mission creation quality has really gone through the roof (IMO) because of this AI capability.  I now have limitless voice actors on call who are willing to say whatever I want, in the exact same tone/inflection that I desire.  No more WOPR sounding voices in my missions.

For example, I recorded this the other day for a campaign I am working on using my voice (which sounds nothing like this recording).  I ran my recorded voice through 11labs using one of their premade voices.  A couple of quick adjustments later (using Audacity) and I have this.  Good enough for what I need.

 

 

Lose 'visual' not 'tally' - Tally is for bandits/bogeys not friendlies. 

[Mistermann]

Thanks 

[CL30]

Yup, these AI voice capabilities are pretty cool. I subscribed to Play.ht for this campaign I’m working on. Mission quality, after Audacity tweaks, is soooo much better. 
 

It takes some experimenting to get them to say the right words sometimes. For example, I have to type “Oozey” to hear it say “Uzi.”  Or “are tee bee” for “RTB.”
 

I have also learned that large paragraphs (I.e. lots of precious characters) are best broken into smaller chunks and files, and then combined in Audacity. It’s super annoying to have typed out/used hundreds of characters only to hear it mispronounce a single word… Texaco.  “Tex-AH-co” instead of “Tex-uh-co.” Shoot me, lol.

[Mistermann]

3 hours ago, CL30 said:

Yup, these AI voice capabilities are pretty cool. I subscribed to Play.ht for this campaign I’m working on. Mission quality, after Audacity tweaks, is soooo much better. 
 

It takes some experimenting to get them to say the right words sometimes. For example, I have to type “Oozey” to hear it say “Uzi.”  Or “are tee bee” for “RTB.”
 

I have also learned that large paragraphs (I.e. lots of precious characters) are best broken into smaller chunks and files, and then combined in Audacity. It’s super annoying to have typed out/used hundreds of characters only to hear it mispronounce a single word… Texaco.  “Tex-AH-co” instead of “Tex-uh-co.” Shoot me, lol.

Try using speech to speech.  You record yourself using tone, cadence, emotions, etc.  then the AI converts your recording to another voice.  No more phonetically spelled words and messed up paragraphs.