Radio Effect Voice overs

[NightFlier]

Hi guys,

I'm looking for some good hints on how to make realistic voice overs.
I've been using Audacity for a while now and I made some voice overs using filters but I'm still looking for the best results.

I really like the Zone 5 Campaign voices and I'm trying to replicate the effect but it's not easy and I'm not an expert sound editor.

Any suggestion is appreciated

Thanks

[cfrag]

18 hours ago, NightFlier said:

I'm looking for some good hints on how to make realistic voice overs.

Hmmm. I think that's a somewhat loaded question, so let me interpret that as 'good-sounding, entertaining' audio from a game perspective (realistic audio is bland, fraught with 'um', 'uh', 'hmmm' and other fillers you don't want in game audio unless scripted). It's a good question, and since you have already tried your hand at it, you've found that it's very, very difficult to create good audio. There's a reason that audio production has it's own academy award section: it's an art form. As such, there are multiple aspects involved, some under your control, others not so much.

• Hardware:
  You need a decent Mic setup. A headset won't do, but for a few hundred bucks you can get a good setup that can last a lifetime: a microphone, a stand, some filter and audio reflection absorber (cheap, audio-absorbent foam for the room) - now let's call it 'The Studio' when your room is configured for audio recording. Many modern mics come with USB, which allows you to skip the amp stage. All this is easy enough, and usually for around 300 USD you are set up.
  Also easy: being prepared. Have your studio set up correctly and tested for recordings ahead of time, have all text printed out and rehearsed. Make sure no major outside activities (construction work!) ruin your recording session because editing background noise out is really difficult. Move noisy hardware (computers with loud fans) out of the studio. Have enough food and drinks on-location that you do not have to make long breaks, and have everyone agree not to eat and drink in the "Hot" room (unless it's scripted, of course). When in a remote location, make sure to have enough energy, and try not to use noisy (both audio and electrical) generators - use batteries. With these simple preparations, you have already covered a good 1% of the way. Yup, sorry. 1%. 
   
• Software: 
  You already use Audacity, which is a very capable audio processing app. Since you aren't interested in multi-tracking, perhaps don't even need stereo, in theory it should cover all your needs, even if there are other, more capable audio processors available, most of them are geared towards music, and treat voice as just another instrument (meaning you pay for other instrument features). I've used Logic Pro in the past (and still own it), but still found me using Audacity in the past years for voice production. 
  
  And here comes the rub: what you are probably primarily interested in are 'filters' or effects that transform a recorded mediocre audio source into the beautiful sound scape that you are looking for: adding radio static, hiss, make it fuller, deeper, or tinnier, etc. And that, unfortunately, is where we leave the realm of tech and enter the world of art. There are - as far as I know - no standards, every production uses their own 'secret sauce'. Sure, it's easy to simply strip the audio spectrum below 300Hz and above 3k3Hz to mimic a telephone call (and similar for FM radio which clips above some 12kHz. So a basic radio voice can be created with some static overlaid on a voice that is clipped to 2k-12kHz). But that's just the beginning and getting to the right combination of filters and settings is art. Even if you find someone who can create the filters for you, using them correctly to produce good results still is art. To give an analogy: I have Adobe Photoshop and a pen-based high-end graphics tablet. All my paintings are crap - I have no discernible artistic talent, and no amount of high-end technology can bridge that gap. What you need is talent, and if you don't have it, the best you can ever hope for is 'not terrible'.
  And that's before we get into mixing for even quality of sound and levels. You can do it yourself, but if you are aiming for real production level quality, you'll need someone (a friend perhaps) who is on a pro or semi-pro level. But it's certainly doable.
   
• Voice Acting
  This is what is going to hurt most: voice acting also is a real talent, and mere mortals like me can only ever hope to achieve 'does not suck entirely' levels. Listening to my own voice is a real turn-off. Listening to my own pathetic attempts at infusing my voice with faux emotion is vomit-inducing. Plus, you need a decent innate voice to begin with (guess what?). Not to mention accent. You can get tutoring (breathing, intonation, enunciation, modulation, tonal exercises etc.) but that only goes so far. If you lack voice talent (as I do), it can be a bad experience. Really bad.

So, in short: you can get some good hardware, you already have OK software. You'll need talent for both acting and audio processing, which can be difficult. After working (never as talent) on a number of productions, I have great respect for what people achieve in something that seems as minor as audio.

This is not to discourage you - the exact opposite is true: try, improve, try again. Just don't expect parity with professional productions unless you yourself have a similar kind of support.

My advice: if you are as talentless as I am, either bite the bullet and simply speak your text, and refrain from acting. Or use a text-to-speech synth. I thought it can't be worse than me. I was wrong, but your mileage may vary. Afterwards, apply the audio filters as best you can to produce audio. Or get friends to help. On both issues. Just don't give up.

Good audio is hard. To paraphrase Kennedy: that's why we do it

 

Edited May 30, 2022 by cfrag

[NightFlier]

1 ora fa, cfrag ha scritto:

Hmmm. I think that's a somewhat loaded question, so let me interpret that as 'good-sounding, entertaining' audio from a game perspective (realistic audio is bland, fraught with 'um', 'uh', 'hmmm' and other fillers you don't want in game audio unless scripted). It's a good question, and since you have already tried your hand at it, you've found that it's very, very difficult to create good audio. There's a reason that audio production has it's own academy award section: it's an art form. As such, there are multiple aspects involved, some under your control, others not so much.

• Hardware:
  You need a decent Mic setup. A headset won't do, but for a few hundred bucks you can get a good setup that can last a lifetime: a microphone, a stand, some filter and audio reflection absorber (cheap, audio-absorbent foam for the room) - now let's call it 'The Studio' when your room is configured for audio recording. Many modern mics come with USB, which allows you to skip the amp stage. All this is easy enough, and usually for around 300 USD you are set up.
  Also easy: being prepared. Have your studio set up correctly and tested for recordings ahead of time, have all text printed out and rehearsed. Make sure no major outside activities (construction work!) ruin your recording session because editing background noise out is really difficult. Move noisy hardware (computers with loud fans) out of the studio. Have enough food and drinks on-location that you do not have to make long breaks, and have everyone agree not to eat and drink in the "Hot" room (unless it's scripted, of course). When in a remote location, make sure to have enough energy, and try not to use noisy (both audio and electrical) generators - use batteries. With these simple preparations, you have already covered a good 1% of the way. Yup, sorry. 1%. 
   
• Software: 
  You already use Audacity, which is a very capable audio processing app. Since you aren't interested in multi-tracking, perhaps don't even need stereo, in theory it should cover all your needs, even if there are other, more capable audio processors available, most of them are geared towards music, and treat voice as just another instrument (meaning you pay for other instrument features). I've used Logic Pro in the past (and still own it), but still found me using Audacity in the past years for voice production. 
  
  And here comes the rub: what you are probably primarily interested in are 'filters' or effects that transform a recorded mediocre audio source into the beautiful sound scape that you are looking for: adding radio static, hiss, make it fuller, deeper, or tinnier, etc. And that, unfortunately, is where we leave the realm of tech and enter the world of art. There are - as far as I know - no standards, every production uses their own 'secret sauce'. Sure, it's easy to simply strip the audio spectrum below 300Hz and above 3k3Hz to mimic a telephone call (and similar for FM radio which clips above some 12kHz. So a basic radio voice can be created with some static overlaid on a voice that is clipped to 2k-12kHz). But that's just the beginning and getting to the right combination of filters and settings is art. Even if you find someone who can create the filters for you, using them correctly to produce good results still is art. To give an analogy: I have Adobe Photoshop and a pen-based high-end graphics tablet. All my paintings are crap - I have no discernible artistic talent, and no amount of high-end technology can bridge that gap. What you need is talent, and if you don't have it, the best you can ever hope for is 'not terrible'.
  And that's before we get into mixing for even quality of sound and levels. You can do it yourself, but if you are aiming for real production level quality, you'll need someone (a friend perhaps) who is on a pro or semi-pro level. But it's certainly doable.
   
• Voice Acting
  This is what is going to hurt most: voice acting also is a real talent, and mere mortals like me can only ever hope to achieve 'does not suck entirely' levels. Listening to my own voice is a real turn-off. Listening to my own pathetic attempts at infusing my voice with faux emotion is vomit-inducing. Plus, you need a decent innate voice to begin with (guess what?). Not to mention accent. You can get tutoring (breathing, intonation, enunciation, modulation, tonal exercises etc.) but that only goes so far. If you lack voice talent (as I do), it can be a bad experience. Really bad.

So, in short: you can get some good hardware, you already have OK software. You'll need talent for both acting and audio processing, which can be difficult. After working (never as talent) on a number of productions, I have great respect for what people achieve in something that seems as minor as audio.

This is not to discourage you - the exact opposite is true: try, improve, try again. Just don't expect parity with professional productions unless you yourself have a similar kind of support.

My advice: if you are as talentless as I am, either bite the bullet and simply speak your text, and refrain from acting. Or use a text-to-speech synth. I thought it can't be worse than me. I was wrong, but your mileage may vary. Afterwards, apply the audio filters as best you can to produce audio. Or get friends to help. On both issues. Just don't give up.

Good audio is hard. To paraphrase Kennedy: that's why we do it

 

 

Thanks for your reply

What I'm aiming for is just to create a good Radio Effect.
I'm not looking to act or even talk with my own voice I use Text to Speech websites to make voice overs so what I need is a good Radio (possibly UHF) effect to get a good result when listened in game

[cfrag]

8 minutes ago, NightFlier said:

What I'm aiming for is just to create a good Radio Effect.

All I can offer is a very basic radio effect (i.e. not necessarily good). The first part is stripping out frequency bands to make the voice sound a bit tinny: filter out 0-2kHz and everything above 12kHz. That alone should get you a creditable 'radio effect'

Then create a second track that has white noise, and then reduce the amplitude (volume) for 99% of the track to very low intensity, with the occasional 0.1s random peaks to 50% or even 90%. Overlay both, and you should be set. If you want to be fancy, edit the static bursts so that they run over the worst of the synth voice's artifacts (that they almost always have, even after carefully editing the source text to vocalize the oddities of written language).

You'll need a 'click' and 'click-click' sound for the start and end of the sound; those you may be able to get from the web, just make sure to not infringe on anyone's copyright if you intend to share your work. overlay them to your existing audio. Master to some 75% even volume, and you should be done. Keep the project for later edits

 

Edited May 30, 2022 by cfrag

[NightFlier]

23 minuti fa, cfrag ha scritto:

All I can offer is a very basic radio effect (i.e. not necessarily good). The first part is stripping out frequency bands to make the voice sound a bit tinny: filter out 0-2kHz and everything above 12kHz. That alone should get you a creditable 'radio effect'

Then create a second track that has white noise, and then reduce the amplitude (volume) for 99% of the track to very low intensity, with the occasional 0.1s random peaks to 50% or even 90%. Overlay both, and you should be set. If you want to be fancy, edit the static bursts so that they run over the worst of the synth voice's artifacts (that they almost always have, even after carefully editing the source text to vocalize the oddities of written language).

You'll need a 'click' and 'click-click' sound for the start and end of the sound; those you may be able to get from the web, just make sure to not infringe on anyone's copyright if you intend to share your work. overlay them to your existing audio. Master to some 75% even volume, and you should be done. Keep the project for later edits

 

 

Many thanks!
I will try but I'm at a very low level in audio editing so I would need "step by step" instructions on what to select and which filter to apply using Audacity

[Paladin1cd]

I had a great background to combine into one track and can’t not seem to find it. I found that easy to use, hard to find. 

[Mistermann]

On 5/30/2022 at 9:23 AM, NightFlier said:

I use Text to Speech websites to make voice overs so what I need is a good Radio (possibly UHF) effect to get a good result when listened in game

Glad to hear you're doing that.  I have been using both Microsoft and Google for text to speech over the past 2-3 years.  The variety of the voices they each support is quite large.  While some voices sound like the computer voice from "Wargames", there are some that are surprisingly humanlike. 

As for a step by step - let me take a stab at it for you.  This is my current approach to Audacity.  Note, I say current.  There are many, many ways to do this.  What I am about to share works for me.  It may not work for you, but it should give you an idea of how to approach.

Step1 – Start with your recorded voiceover.  Let’s say its called example.wav.  Open Audacity and IMPORT that file into Audacity.  You’ll have something like this …

Step2 – optional.  I usually compress the file so the peaks aren’t so big.  It makes the voice output a bit more uniform.  I usually use this setting and only adjust the Threshold if the “quiet areas” are less then -24db.

The waveform is adjusted to look like this when you hit okay.

Step 3 – Amplify to 6db and allow for clipping.  This setting is very trial and error.  DCS doesn't play audio at the level you record it.  I wish there was a scientific way to know the right answer here.  Over the years, I've settled on 6db as it seems to work well for me when in the DCS cockpit (with or without "hear like in helmet" checked.

Step 4 – I run this filter curve EQ over the sample

This results in the following waveform

Step 5 – I use a filter called vinyl next.  It’s a free Audacity plugin that you can find here - https://www.izotope.com/en/products/vinyl.html

Here’s the settings I use.  You’ll want to play around with these controls to get the sound you’re looking for.

I will usually apply this filter twice resulting in the following waveform.

Step 6 – I re-amplify the soundwave back up to 6db (step3)

Step 7 - Export to ogg format and you can include in the ME.

This is my “easy” approach to making voiceovers.  Could they be better?  Absolutely!  Do they “get the job done”?  For me, that’s a yes.  If I was selling a campaign, I’d use volunteer voice actors and really hone the radio background and add mic clicks/beeps as you hear in the professional campaigns.

Note – these are my personal steps gleened for years of trial and error.  Experts in audacity or sound engineers will probably laugh pretty hard at my “technique”.

Hope this helps!

EDIT:  Sorry for the huge text, I copy pasted from a different editor.

Edited May 31, 2022 by Mistermann

[Mistermann]

I realized I should probably upload the audio I was using to give you an idea of the before and after. This way you can judge for yourself beforehand if the "juice is worth the squeeze".

Note, I am completely open to feedback.  If there are better ways to do this, I would love to learn!

 

[sirrah]

I've been using this modified Radio Message Generator from akaAgar in the past. (https://github.com/akaAgar/radio-message-generator or just download the attached "Radio Message Generator.zip")

 

It uses Window TTS and is very easy and superfast in use.

I added all default English TTS voices (which iirc required some regedit stuff) and used Audacity to add various background noise. Sounds like a hassle, but it's actually quite simple (heck, if I can do it, most of you will for sure!)

 

At some point in history, perhaps due to a Windows update, some of the voices stopped working, but this tool really is gold!

 

 

Perhaps one of the wizkids here could pick up this tool and further improve it (make it more easy to add TTS voices and noise effects)

 

I truly advise/urge you guys to try this tool

 

Radio Message Generator.zip

Edited June 1, 2022 by sirrah

[NightFlier]

@Mistermannmany thanks mate! I'll go through that soon

@sirrahI know that, I don't like the robotic voices that comes with Windows. I'm actually using trials on different websites with some professional voices but it would be great if someone able to enhance that generator tried to update it.