Automatically clean and boost your audio with Auphonic and DaVinci Resolve

If you want to automate your post-production of the audio portion of your videos, here's a way to do it using a service called Auphonic. You can use it for free for processing two hours of audio per month. You can also get paid plans if you need to process more audio than that per month. But for someone running a YouTube channel where you publish a video once a week and that video is 10 to 15 minutes, two hours obviously is plenty. So, let's show you how to do it. I'm here in DaVinci Resolve on the Fairlight tab. This is the audio that we're starting with to show you some of the problems that we have right now that Auphonic is going to process and solve for us. So, at the typical distances you would normally use a shotgun microphone, 18 inches, 45 centimeters. Plosives are not really a problem. You're going to use this for podcasting and get really close to it. That's when you can start to run into plosive issues. You should use a foam cover in that case. Okay, that's where we're starting. A couple of issues, obviously the music and the dialogue is not balanced. Auphonic can take care of that for us. The dialogue is not nearly loud enough. And also there are some pretty prominent breaths in the audio there as well. And there's a little bit of noise. It was probably hard for you to hear that there because it is so much quieter than the music was. Let's go ahead and hop on over. Well, before we hop on over, how do I get the audio out of here? What you can do is come up to the timeline menu and choose "Bounce Mix to Track." We're going to take this audio mix and bounce it out and put it on a new track. So what we want to do is take bus 1. That's our main mix in this case. It's a stereo format. That's great. And we want to go ahead and add that to a new track. Once we've done that, click OK. And boom, there we go. If we come back to the start here and just kind of fit this on screen. All right. Now that we've bounced the mix to a new track, which we have down here, you can even see visually that we have a problem here. Here's the music. Here's the dialogue. Obviously, this is going to come through quite a bit louder. But what we can do now is actually export this particular clip right here and get that over into Auphonic to do the audio post-processing. A couple of things you want to remember to do. Number one, we're replacing these tracks right here now. We can go ahead and mute them because we don't want them to play back at the same time this is playing back. That's one thing. And then also we need to export this audio right here. I'm going to come up here and choose the range tool. Go ahead and click on this clip right here, which is what we want to export. Right click, choose export audio files. These are the settings we want to use. Go ahead and choose where you want to save it. We're going to go by track name. We are going to prompt for each duplicate. These are kind of some of the default settings here. Broadcast wave, same as project, 24, interleaved, consolidated files, and then use only the selected tracks. We can leave these at their defaults, and we leave no normalization right here. So go ahead and click export, and there we go. Come on over into Auphonic. This is the service that we're going to use. You could just drop your file right here. I actually already have a production. They call them productions. It's basically a project. So we're going to come over here into the productions, and I'm going to edit this one that I've already set up just to make things a little bit quicker. First thing we need to do is go ahead and upload our file. So I'm going to choose the file that we just exported. Choose upload. Now, these first several sections, intro and outro, basically, you don't need those. Just leave them at their defaults. We need to choose an output file. I only want one. I want WAV 24-bit PCM, optimal stereo, and I want it to end with WAV. We can give it a name right here. That's what we're going to call the export file. Speech recognition and publishing to external services. We'll go ahead and just skip those, leave those off for now. If you do get a paid account, you can have it automatically create show notes or do speech recognition for you on the free account. You don't get that. Here are the things that this will do for us. It'll do adaptive leveling. So if I click here, I want to set the leveler strength to 100%, which is the default. I want the mode to be default leveler, and I want the compressor set to auto. Filtering. Now, the reason-- actually, I should just say, the reason that we use an adaptive leveler is first of all, we're going to even out the music and the dialogue, but it will also do some compression. So what that means in practical terms, if you remember when we were in DaVinci Resolve and we tried to normalize, first of all, we weren't able to achieve the loudness that we wanted to because it couldn't boost it enough without clipping some of the audio, without making some of it distort. A compressor will actually reduce the dynamic range of your audio a little bit so that you can boost it to the level you want to get it to, and this will take care of that automatically for you. We're going to have to leave filtering on, and we'll just choose the voice auto EQ. That's just going to clean up your audio a little bit for you. Loudness normalization for YouTube. For pieces that are mostly dialogue, mostly spoken word, I typically choose -16 LUFS. We want the maximum peak level at -1.5 dB true peak. We have some music that is stereo, so we're going to uncheck dual mono, and we're going to go ahead and choose program loudness here because we have both music and dialogue. If you have just dialogue, you could choose the dialogue loudness instead. Noise reduction. This one here, I want to go ahead and put this to dynamic. We're going to keep the speech and music and then remove any other noise from that. Now, normally this comes in at auto, but I found that's too aggressive. It sounds obviously noise reduced. It starts to sound a little bit robotic. I prefer to do something a little bit more subtle, and so I'm going to choose 12 dB, which is the medium setting. I also have been finding some pretty good results with the remove breathings. I would have said remove breaths, but okay. If you choose the reduction amount at a moderate level like 12 dB, this works pretty well. I found if you set it or leave this at auto and choose remove breathings, it's too aggressive. It literally cuts the breaths entirely out, and it sounds fake, and it's not the sound that I really like, so I wouldn't recommend that. So these are the settings I'm using. If you're doing video, you do not want to choose automatic cutting. This will remove silent portions, but you don't want to change the length of your audio because you have to match it back up to your video. So don't do that unless you're doing an audio only podcast or something like it. So make sure that's unchecked for video. We'll go ahead and click start production, and it will tell us how much we have left in terms of our credits. I've got an hour and 51 minutes left, so that's fantastic. We've got plenty. We'll go ahead and start the production. Now, what it does is it first uploads the file to the Auphonic servers. Once it does that, then it sends it over to its servers that do the actual processing of the audio. Now, this is a 36-second clip, so it's not a really long one, but it can take a little while. For example, if you had a 10-minute video, it may take a few minutes, it will alert you when it's done by sending you an email, so you don't have to sit here and wait for it necessarily. But you can see right now it's doing the audio encoding, so we're in a good spot, and there it is. This is now what the processed audio sounds like. So at the typical distances, you would normally use a shotgun microphone, 18 inches, 45 centimeters. Pluses are not really a problem. If you're going to use this for podcasting and get really close to it, that's when you can start to run into plosive issues. You should use a foam cover in that case, and what you might do is move the microphone off to the side, still aimed at the corner of your mouth, but that way when you're talking and you say P or B or whatever, you're not going to get that puff of air directly into the microphone, so you shouldn't have any plosive issues. Okay, that's actually sounding quite a bit better. We can go ahead and download our WAV file, which is this top one here. You want the WAV. I would recommend not using MP3. MP3 will have more artifacts, and it won't sound as good. So go ahead and download the WAV file, and then let's get ourselves back over into DaVinci Resolve. What I can do now is I'm not going to use this track, so I can go ahead and mute that. We're going to come back over to our media pool, and we are going to go find that new file that we just created. I'm going to go ahead and find it here. It was in downloads, and here it is. Minus 16 LUFS. So there it is. We're going to come back over to the Fairlight tab, and I'm going to go ahead and take that minus 16. Let's see which one is it here. I have a bunch of them going. Here's the new one right here. I'll go ahead and drag that down to a new track just like that. We have already muted these other three originals. Now, once we have this embedded in there, as long as I push this up against the left, let's use our arrow tool here. As long as that's pushed up against the left, you can see it's the exact same length as all the others, which is good. That means it's in sync with our video. You could test it if you need to do that. But here's your finalized audio. We'll right-click and just check to make sure it got our levels where we wanted. It got it right around minus 16 LUFS, which is what we wanted, and the true peak is at about minus 1.4 dB full scale. So we're good there. Now we can go ahead and export our video as usual. I hope that was helpful for you. If you have any questions, go ahead and leave that down below. Get out there and make some great sound, and we will talk to you again soon.