 Hey, it's Anfa. This video is a extended cut of my segment that I've recorded for another Linux YouTuber called TechHut. He's been making a video about tools for content creators, specifically free and mostly open source ones too. And he has kindly invited me to talk about audio software. So I've sent him my raw footage and he did a great job and cut it down to four minutes. But we both agreed that it would be nice to publish an extended cut because a lot of good tips has landed on the cutting room floor. So I took the same footage and made my own version. And this is that video. In case you haven't seen his video yet, there should be a card in the corner. What is that? Oh, come on. Not that kind of card. Not that kind of either. Stupid video editing monkey, I don't know. Yeah, go watch that video first and this one will make more sense then. Yeah, that's all I wanted to say. Enjoy. Hi, I'm Anfa. I'm a music composer, producer and sound designer. I work with audio production for video games. And I also make videos about the ins and outs of the Libre slash Linux audio production ecosystem. TechHot has asked me to recommend you two essential pieces of software that I use. And I'll do that. But first, I need to say something else. TechHot has told you before in this video that kid in life is the way to go. But I will say, try all the video editor. For me, it does a much better job than kid in life, as it's fully GPU accelerated, more stable and has better editing tools, at least for my needs. Now I'll go back to the audio stuff. So of course, both programs I'll recommend are free software, both run on Linux, Mac and Windows, and both have the names that start with the letter A. It's Ardor and Audacity. If you've never used them, you might think that they look pretty similar, but they are indeed very different. And they are both very useful and kind of complimentary. Ardor is a full blown digital audio workstation. That means you can make music in it, score films, record bands or produce sound effects for games like I do. It's a heavy tool, however, and there's a steeper learning curve than with, say, Audacity. But once you learn Ardor, there is little limits to what you can do with it. I can recommend you a video I've made, which is a all in one beginner tutorial for Ardor 6. On the other hand, we have Audacity, which is an audio editor, not a digital audio workstation like Ardor. What's the difference? In Audacity, all the editing is destructive. That means if you remove a piece of audio, you can't get it back unless you use the undo function or you just import a different copy of the source material. In Ardor, however, if you delete a piece of audio, you can just grab and drag an edge and you get it back. It's all there. Editing in Ardor is non destructive. In Audacity, it replaces what you had before. What I mostly use Audacity for is checking my final exports for levels. For example, if I exported a video file with these levels and I would like to get this file ready for uploading to YouTube, what I would do is go to effects, loudness normalization, and choose perceived loudness. Go to negative, maybe 16 or 14 L UFS. Negative 14 L UFS is the loudest you can have on YouTube. But for spoken audio, negative 16 or 18 would be fine. And now our sound is louder. However, if there was a very loud peak somewhere here, for example, if we normalize to negative 13 L UFS, you can see that parts of our recording touch the 1.0, or if we zoom out, you can see that our audio is crossing the maximum available level. And that's no good because if we export that, it's going to clip and it's going to sound terrible. So after I apply loudness normalization, I also apply a limiter. I use an effect called Fast Lookahead Limiter. It's a Latspa or LV2 plugin. And we don't need any input gain. And it's a good idea to limit to negative one decibel because that makes sure that nothing clips later in the process when the audio is re-encoded on YouTube after you upload it. And the release time of 50 milliseconds is usually sufficient. Now I can do apply. And you can see that our peaks have been reduced and they are no longer clipping, which is great. Now you can see that applying effects in audacity is just like editing. It's destructive. This is now all I have. The only way to go back is to undo. And unless I made a copy. That's why often when I do some things I'm not sure about in audacity. I duplicate my tracks, edit duplicate. I mute the original and I work on the duplicate. So now if I mess something up, I can just go back to the original, duplicate it again and try once again without worrying about possibly losing my work and not being able to go back. Audacity is surprisingly powerful and it's really easy to denoise 50 files if you need so. You can just drag them all in and then go, now select all audio, apply an effect and then export, export multiple and use, for example, the tracks. And it will re-export all the files. And yeah, now as you remember, processing and editing in audacity is destructive, so this is all we have. In order, however, all the effects are applied in real time. So if I add the fast lookahead limiter plugin right here, you see I have the same settings, input gain, limiting, let's go negative one decibel and release time we want 0, 0, 0.5. It's just 50 milliseconds. And it's going to do the same thing, given we have increased the volume. I'm pressing alt and six right now. Alt seven decreases the volume. It's here in region, gain, boost gain and cut gain. These are the shortcuts. There's also normalize, which you can also constrain to RMS, which is negative 14. And now we have this similar thing, which I also use for like processing if you're a submit in music for my live streams. I use this normalization function. And Ardor applies all these effects all the time. So if I re-record part of this, oh yeah, woo. And let's say I make it louder. Now if I re-export my session, my project by going session, export, export to audio files, type in a name, choose a format, blah, blah, blah. I don't have to reapply the effect like I would have to do it in Audacity, because Ardor is non-destructive. It's all about having things be separate and being applied automatically, and you can change any part of it in the chain. And all the rest stays the same. So you can be very productive with it. But it's a heavy tool. Audacity is not a heavy tool. It's easy to pick up. It's easy to do things. You can even record audio in here. And if you really want it, you could make some music, but it would be like trying to shave yourself with a shovel. It's a great tool, but for a different task, I'll drop in one extra thing for you. And that is FFmpeg. So once I have increased the loudness of my exported video and I export that to a separate audio file, now I need to replace the new audio and put it into my already rendered video file. But I don't want to mess up my video project or create a new video project in my video editor just to replace the audio. That would be really silly. So FFmpeg is the perfect tool for the job. All you have to do is run the FFmpeg command. First, type in dash i, which specifies an input file. And here will be our video. I use MKV files. Now we'll specify the new audio file. So dash i, new underscore audio. Now these names are all placeholders. You will replace them with your actual file names. And what this does is grabs two input files, but we can have only one output file. So what do we do? Well, we shuffle the streams around and we do that with the map parameter. So I type dash map. And now we specify which file, which input file we want to take the stuff from. And I want the first one, which is in programmer's terms, there's zero one. So I type zero colon V for video. So now we have map zero colon V. That means take the video part from the first specified file, which is video dot MKV. Now we need another map parameter. And that is map. Now we need one, because we want to use the second specified uh, input file, which is new audio, but programmers count from zero. So we need to use one instead of two. So one colon A for audio. And that will take the audio stream from the second specified file. So now we told ffmpeg to swap around the streams. We take video from the first input and audio from the second input. Great. Now we need to specify the codecs. And it's very simple. Dash C for codec, copy, which will take the data and just remix it. So interleave the binary data without re-encoding anything. So we are sure that the quality stays 100% perfect. And the final step is specifying the output file name. I'm going to call this new video dot MKV. Now you run this command and you have a new video file with swapped out audio. Much faster than creating a new video project. Okay, that's all from me. Big thanks to Tech Hut for inviting me for this segment. If you'd like to learn more about audio production on Linux with free software, check out my channel. Bye. That's all. Thanks for watching. I hope you learned something. Also huge thanks to all the people who support my work financially. If you, dear viewer, would like to join them and help keep this show going, please go to patreon.com slash ANFA or liberapay.com slash ANFA. You can also check out my beautiful, beautiful website at anfa.xyz or XYZ. I also have a community chat. Actually, it's a rocket chat instance self-hosted at chat.anfa.xyz. So if you want to meet some people, get inspired, ask me about stuff, get help with Linux audio or liber audio programs. Check it out.