Sound Editing for Video: Level Up Your Audio in 5 Steps

1. Introduction: In the film-making world, we have a saying which goes, audio is a king of video. This means that the perceived quality of your video is more influenced by the quality of your sound than the quality of the visuals. Despite this, sound editing is usually taught in a very complicated, confusing way. There's a lot of clashing advice out there. Then, there are all of these intimidating knobs and frequencies. While I am here to clear things up, this class presents a clean, step-by-step, and streamlined process you can use to enhance the quality of the sound in your video. My name is Dandan, and I am a top teacher here on Skillshare, and an Emmy Award-winning filmmaker. Often as creatives, we are visually biased and ear blind. Besides teaching you the technicals, this class will also teach you how to hear sound in new ways. Whether you're looking to post videos on YouTube or create a commercial for your company's latest products, this class will help you deliver your message with high-quality sound. Even if you're creating podcasts, the principles learned in this class will help your episode stand out in that noisy world. For this class, you'll need Adobe Premiere Pro and Adobe Audition, both of which can be downloaded as free two trials. However, all of the steps demonstrated here can be applied to any video editing program with comprehensive sound capabilities. You'll also need a good pair of headphones and a quiet space, so you can really hear the nuances of sound. Without further ado, let's dive right into this fascinating world of sound. Welcome to the class. 2. Class Project & Tip for Success: For your class project, you have two options. You can either edit a short video clip with the principles learned in this class. Then post a before and after version on the Projects page on Skillshare. Or you can share on how you're planning on applying the principles learned in this class to your project. Tips for success; if you're not used to hearing sound, it can take a while before you're able to hear the nuances. It's just like color grading. When you start learning how to edit color, it can be difficult to see the subtle color tones of the highlights or shadows. It's the same with sound editing. So don't get discouraged if you can't easily hear the sound differences I mentioned in this class. I will actually exaggerate them so they are clearer. The important thing is to keep practicing and listening with a good pair of headphones. Get excited. You are developing a new superpower of listening. 3. Prep: Organize Your Timeline: Before we start, let's get set up and organized which is going to streamline everything and save you a lot of time. In this part, we're going to separate each sound source and type on its own track. For example, here in this timeline for my documentary short Dottie, you can see how Dottie, my main character, has a dialogue on this dedicated track named Dottie's voice. You can rename the track by right-clicking. If you use different mics to record your character, make sure you separate each mic on its own track. Same if you recorded your character in different settings that sound vastly different in terms of their ambience or room tone. Then, you'll want to put the music, each song with its own track. Below the music tracks, you'll want the sound effects. You have the choice here of separating each of your sound effects according to their own track. This can be helpful if you have multiple sound effects at the same category, like lots of footsteps. That way, you can globally change all of them from one command instead of having to adjust each clip one by one. Or if you don't have so many sound effects like my case here, you can just place them on two main tracks since it's not so much work to adjust each clip. If you need to add more tracks, you can right-click and select "Add a Track". As you can see, you can also tell where do you like to place these tracks. I have also color-coded the sound clips according to their category, just so it's easier to see what I'm working with. To do this, you can right-click and select ''Label'' with the color you'd like. [MUSIC] Now, I want to show you the audio panel, which will be our main hub in Premiere Pro for the sound editing. To do this, we're going to open the audio panel and then the audio track mixer. If you can't find it, go to Window and then click "Audio Track Mixer''. Make sure not to get this confused with the Audio Clip Mixer. You can undock the panel to make it bigger. As you can see, all of your labeled audio tracks are here. Every change you make here will affect all of the clips on the track. Now, hit the little arrow on the left to open the effects rack. This is where we will add our effects. Below this, you have the fader, which will control the volume of that specific track and save you from having to go in and changing the levels of each clip. Now that we're all set up and oriented, let's take actions. 4. STEP 1: Clean Up Your Sound: The first thing you can do to make your sound better is to clean it up. Besides general noise, which you probably know can be fixed with noise reduction, there are five common unwanted sounds you want to fix. The first is clipping. Clipping happens when your sound peaks hit or surpass zero, and what it sounds like is distortion. As they get more and more, a visual analogy are those hotspots in your image where the areas are so bright they look white. Information is lost there and it doesn't look good. Similarly in sound, you want to avoid your peaks hitting or surpassing zero. Even when you set your recording levels to give yourself headroom, sometimes some transient peaks will slip through and clip. This often happens when someone is laughing or making an exciting exclamation. Here I have an example of something that clearly sounds distorted from the clipping. As they get more and more you can see how the top is just shaved off, and it hasn't even hit zero decibels yet. If I bring the volume down, it still sounds distorted. As they get more and more. To fix this, it's simple. We'll need to send over a clip to Adobe Audition, from our Premiere Pro interface. To do this, right-click on your clip and hit Edit Clip in Audition. Now these two programs are linked and talking to each other. Whatever changes you make an audition will automatically be applied to your clip in Premiere Pro, when you hit Save. We're going to use the Declipper effect. To add this, we're going to go to Effects, down to Diagnostics, and then select Declipper. Then we're going to highlight our entire clip and hit Scan. On the bottom it'll show us the number of possible clipping problems that were detected. You'll see in this case there are 61. Now hit Repair All, and if you play again, you'll see that I removed the distortion. As they get more and more, if you even look at the waveform, you can see how the top is now more rounded out compared to before. Amazing, isn't it? I'm using the default setting now. But if you still hear some distortion you can try the other settings, like Restore Heavily Clipped. The second thing we're going to listen for are mouth noises, also known as clicks. In this example you can hear these mouth noises. [inaudible] I left home when I was white. If I isolate them, they sound like this. To remove these, it's also very simple. In Audition, you can go to Effects, Diagnostics, and now you hit Declicker, highlight your entire track. Just like we did for declipping, you're going to hit Scan and Repair All. [inaudible] I left home when I was white. As you can hear, the clicks are still there. So I'm going to try the heavy Declicker setting and see whether that works better. [inaudible] I left home when I was white. I got turn green on the fields, and I return red to the house. It does. [inaudible] Likewise you can do this in Premier, by going to the effects rack and adding Automatic Click Remover. If you still hear some clicks coming through, adjust the threshold until they are minimized. The third thing you're going to watch out for are plosives, which are the popping sounds that happen when the rush of air from saying consonants hit your mic from your mouth. Often this happens with p, h, and t words. This is what it sounds like. Plosives, plosives. Most of the times your plosives will be more subtle than this. Listen to the words two and effect in the next example. Hit Preview to hear what it sounds like without effect. This is something that you want to fix by addressing each plosive one-by-one, instead of adding an effect to the entire track. Here I have the clip with a plosive. Hit Preview to hear what it sounds like with the effect. Go to Effects, find the FFT Filter, add it to your clip, and then choose, Kill The Mic Rumble. As you can see, the plosive goes away. Hit Preview to hear what it sounds like with the effect. Similarly you can do this in Audition by highlighting the problem word. Hit your peaks. Then go to Effects, hit FFT Filter. Hit your peaks. You can also move the curve a little bit downwards to grab more of that plosive, if any remains with the default setting. Hit your peaks. Avoid your peaks hitting. The fourth thing you're going to watch out for are any distinct unwanted noise passing through your vocals, like glass breaking outside or airplanes flying overhead. Here I brought an example of an interview I had, where there is a bus passing outside, since my main character lived in Lisbon, a busy city. At that moment they began a printing factory. To fix this, this is where the spectral display on waveform mode in Audition is needed. To turn it on, go to this button here which says Spectral Display, and boom, this beautiful display comes on. What this does is basically map out all the energy of the different sound frequencies with brighter colors showing more intensity at the frequencies located on the right. You can see I have one from my left channel and one for my right channel. The yellow means a concentrated build-up of frequencies, going down to purple, which shows less concentrated frequencies, usually scattered noise. What's cool about this is that you can get really targeted with your sound selection and sound removal. For example, if I select this portion, it will allow you to hear just that portion. You can even get nit-picky and use the lasso tool, just like in Photoshop, to select a specific area to reduce or remove. In this example you can see where the bus sound is concentrating. You can even test by moving the selection above and hearing whether there's bus sound in there. There's not. I'm going to bring down my selection until I isolate the bus sound as much as possible. Now I'm going to tell Adobe Audition that this is a sound print I'd like to remove from my track. To do this, I'm going to go to Effects, Noise Reduction, Learn Sound Model. Now Audition is going to say, "Okay. I'm going to listen for this particular sound." Then select a portion of the speech where the sound is present. At that moment they began a printing factory. Go to Effects and hit Sound Removal Process. Play your sound to hear how the effect is working. At that moment they began a printing factory. As you can hear, there's no more bus going through his voice. At that moment they began a printing factory. The last thing we're going to address is reverb, which is basically echo. If you recorded in a large room that has a lot of hard surfaces, chances are there's going to be some echo present. Here I have an example where I recorded in a hard surface room where there's lots of reverb present. Days when you don't hear or feel anything from your inner intelligence. To minimize this reverb, it's simple, you add a dereverb effect. I'm going to add it now to the rack, so it applies to my entire track. As you can hear, if I push it all the way up, my voice sounds distorted. Chances are that if you haven't been in touch with this intuitive voice, it's going to be a little shy coming out. So there can be days when you don't hear or feel anything. I'm going to push it until I start hearing my voice distorting. From your inner intelligence, and that is totally normal and okay. Chances are that if you haven't been in touch with this intuitive voice, it's going to be a little shy coming out. So there can be days when you don't hear or feel anything from your inner intelligence. So there can be days when you don't hear or feel anything. It's going to be a little shy coming out. As you can hear, it took away some, but not all of the reverb. To do this in Premiere Pro, you can also add a dereverb effect to your entire track, which will affect all of the clips on that track. 5. An Important Note on Using Audition: When you edit in Audition, there are two modes. The first is waveform mode, the second is multitrack mode, which looks like an editing timeline. The big differences between the two are that with waveform mode, all edits and effects are destructive. That means whenever you apply some effect, like the DeClipper, and hit "Save", it will make the changes into your original source file. There's no going back. With multi-track mode, the edits and effects are non-destructive, meaning that even if you hit "Save", you can remove them later. Why use waveform mode at all then, which you'll see me using throughout this process? There are three benefits. The first is that you can really look at your waveform clearly and make micro-adjustments. For example, you can bring down some peaks with control. The second is that it offers a spectral display, which, as we just learned, allows you to clean up any unwanted sound or noise with precision. The third is that there are some effects, the ones labeled process like the DeClipper we used before that are only available on the waveform mode. When editing with Adobe Audition, I recommend saving a copy of your original source file just in case you hit "Save" by accident and bake an effect that you didn't want. I know destructive sounds scary, but a lot of the times, you just want to remove the noise and be done with it. Rest assured, when using waveform mode, you can apply effects and Control Z to undo them. However, if you hit "Save", there's no going back. 6. The Essential Frequency Ranges: The second action you can take to make your audio sound much better is to use EQ, which stands for equalization. But before we dive in, we'll need to learn about sound. Before we start editing sound, let's first understand what we're working with. Sound is made up of air disturbances that vibrate at different frequencies. For example, when you pluck a guitar string, it sends disturbances through the air that vibrate at a certain rate. The higher the frequency, the faster the sound wave moves up and down, the more sound waves there are per second. The higher pitch the sound, the lower the frequency, the slower the sound wave moves up and down, the less sound waves there are per second and the lower pitched the sound. The frequency of sound, or the number of sound waves per second is measured in Hertz. Just like how the visual spectrum can be divided into different values, sound also has a spectrum based on its frequencies. You can divide the sound spectrum into six zones for mixing. For video editing, you just need to focus on five of these areas. Note that these are ranges, so every sound engineer defines them slightly differently. The exact numbers may vary depending on your unique situation. Why do we focus on these ranges? Each of these ranges have specific characteristics, including good and bad qualities. This is similar to your exposure curve and color grading. You have the blacks, shadows, mid-tones, highlights, and whites. When you're able to see something off balance, you know which area to address to fix the issue. Let's go through these ranges one by one, talking about the good and bad characteristics of each. The first range, 0-50 Hertz, is your extreme low-end. This area is usually more important for music mixing then for video. Since this is where the frequencies of a nice full base in punchy kickdrum reside, the sounds in film usually do not vibrate at this lower frequency range. If they do, they are probably unwanted rumbles from cars passing by or accidental microphone bonds that sound boomy. I've isolated this region with my multi-band compressor just so you can hear what it sounds like. Two options. You can either re-edit this. As you may notice, there's not much you can hear in this range from my speaking. However, I have a sample here with some outside rumble Not so nice. In fact, we usually just remove the sounds in this range by applying something called a high pass filter, which I will get to later. In summary, a nice extreme low-end brings out a nice punch. Typically in music instruments, if this range is not balanced properly, it can make your mix sound boomy or rumbly. Next is the 50-250, 300 range. Here I have an example of my voice. I'm going to emphasize this range so you can hear what it does. Share how you, it's like color grading. When you start learning how to edit, the sounds in this area, govern warmth and thickness, like what you often hear in those nice podcasts, voices that sound like melted chocolate. This area is especially useful for vocals if they are sounding thin and you'd like to bring out their warmth. However, if there's too much of this range, your mix will sound muddy. These spoken words will lose their clarity and sound muffled, almost like you are speaking through a pillow. Here I have an example of my voice. I'm going to emphasize this range so you can hear what it does. Because it'll be hard for you to hear the nuances of the sound processing that I do. However, you'll need a pair of earbuds. You can hear that if I remove this range, the voice sounds thin as you sound later on. You'll need Adobe Premiere Pro and Adobe Audition, both of which you can download as free two-week trials. Third, we have the 250-1 kilohertz range. Here we have the central frequencies of voice for intelligibility. Hello, it can be difficult to see the subtle color tone, of the highlights of the shadows. It's the same with sound editing. The more you practice the more you train your ear to. If this range is well balanced, the mix will sound full present and open. If the frequencies are not well balanced in this range, your vocals can sound boxy, which describes the effect of recording in a place with lots of drywall and tile. Here is what that sounds like. The class resources section. Last but not least, you'll need a quiet space to learn and edit. For your class project, you have two options. You can either re-edit. If this range is not balanced properly, your voice can also sound too congested or nasal. Here is what that sounds like. Clip onto the projects page on Skillshare or you can share your reflection of how your perception of sound has changed after taking this class. If this range is not present enough, the voice sounds not present or full, because you don't have those frequencies that make up the vocal body present. I do here can be applied. It can be very helpful. It can also be very helpful to have the sound frequency map printed out in front of you, located in the class resources section. Last but not least, you'll need a quiet space to learn an edit. Fourth, we have the 1-5 kilohertz range, what's known as the upper mid-range or high mids. Here is what that sounds like. To pick up on the things you didn't hear before. Don't get discouraged if you can't easily hear the sound differences I mentioned in this class. I will actually exaggerate them a little so they are clearer. If the sounds are well-balanced here, your mix will have clarity. When this range is too high, the voice sounds shrill. You have two options. You can either re-edit the sound from. You can either re-edit this. For your class project, you have two options. You can either re-edit the sound from a short video clip using the principles learned in this class. When this range is too low, the voice sounds stall and you can boost this range to brighten up the mix and add some clarity. Or you can share your reflection of how your perception of sound has changed after taking this class. Or you can share how you're planning. Finally, we have the five kilohertz and beyond range. They are unclear. The important thing is to keep on practicing and listening with a good pair of headphones. Get excited, you are developing a new superpower of listening. This is where sibilance is found. Sibilance is the "s" sound in speech, which helped bring out this nice crisp clarity to your spoken words. The exact location of these sounds depends on the unique properties of your voice. A big mistake I see people make is having too much sibilance where the "s" sounds in their speech sound painful like they're going to chop off your ear. Besides that, too many frequencies in this range can make your audio mix sound aggressive. I will actually exaggerate them a little so they are clearer. The important thing is to keep practicing and listening with a good pair of headphones. Get excited. When this range is too low, your speech can sound inarticulate because you don't have that sibilance to add definition to your words. Your mix can also sound not present or forward, like it's playing in the distant background. Thing and listening with a good pair of headphones. Get excited, you are developing a new superpower of listening. The anatomy of sound. I know, this might seem like a lot of information. Don't worry about them too much yet. We're going to work with this frequency ranges in the next two lessons, where you'll get to learn by doing. 7. STEP 2: EQ Your Vocals: Back to EQ. What an EQ tool does is allow you to raise or lower certain frequencies in your sound. Like a sculptor, we focus first on lowering frequencies that are building up too much in the voice. This is also known as subtractive EQ. For example, if your voice is sounding muddy and unclear, you'd want to dip the presence of frequencies around the 200-300 hertz range. Then, we'll boost any frequencies that are not present enough. This is known as additive EQ. Most times you don't even need to boost other frequencies, because cutting the problem frequencies solve the problem. Before we get started, go ahead and highlight a short portion of your dialogue by creating in and out points. To do this, move your play head to the start of your selection, mark in or I if you're on a Mac, then move your play head to the end of your selection and mark out or O for the short part, then press a "Loop Playback" button. What this does is it'll keep replaying this clip once it reaches the end of your selection, just so you don't have to keep moving your mouse back to the same starting point. I lived was the world, but it was a very small community. We grew up. To do EQ, first add the parametric EQ effect onto your effect rack. You can undock the audio track mixer to make it bigger. Now double click on the effect to pop up its window. As you can see, this parametric equalizer maps out all of the frequencies present in your clip, showing you the value in hertz on the x-axis and the loudness in decibels on the y-axis. What this tool allows you to do, is shape the frequencies by creating bands. As noted by the dots L, one through five, and H. For this attractive EQ part, there are four steps. The first is to add a high pass filter. What this will do is remove any low-end rumble that we don't need, that aren't found in voice. To do this, click on the "HP" button to activate it. You can see how now a new point appears called HP with an accompanying slope down mine. Before the line was even, now the line is sloped down. The frequencies on the left of the line will not be allowed to pass through this mix, while the ones to the right of the line will. As you can see, it'll show you at which frequency value the threshold center's on, here in frequency. With your timeline plane, move this high pass threshold, 250 hertz as a starting point, and keep moving it to the right until you start hearing it distort your voice. We grew up with those riddles and stories because we had no television, no books hardly, but pictures, no cartoons of course. I thought where I lived, was the world, but it was a very small community. We grew up with those riddles and stories because we had no television, no books. Usually, for males, you can keep it at 60 hertz. For females, you can bump it up to 80 hertz. It really depends on the pitch of your voice, so keep your ears open. The second step in this part is adding a low-pass filter on the other end of the spectrum to cut out any unwanted sounds on the high end that we don't really need. Usually noise from the room or high pitched static. Set this at 18 K and move it to the left until it starts distorting your voice. Generally, you don't want to move below 14 K. Riddles and stories because we have no television, no books hardly, no pictures, no cartoons of course. I thought where I lived was the world, but it was a very small community. You can also change the slope of this line. In this clip, there was a lot of noise on the high end, so I'm going to make the low-pass filter steeper, so it cuts out more of that noise. No cartoons, of course. I thought where I lived was the world, but it was a very small community. We grew up with those kind of riddles and stories because we had no television, no books hardly, no pictures, no cartoons of course. I thought where I lived was the world, but it was a very small community. The third step is to listen for any problem areas. Does your voice sound nasal? Does your voice sound dull? Then go to the EQ chart, identify the potential problem area, select the band in that area, and pull the frequency slightly above to hear where the source of that area is coming from. For example, in this clip it sounds muddy. Most of the time when you're recording an interview subject, for example, to start off a little bit on the quiet side because they're a little shy. But then as they get comfortable, their voices start to raise and during the enthusiastic parts they get really excited to share their story. I know muddiness happens around 200-300 Hertz, so I'm going to select "Band 2" and raise it, moving it back and forth until I can isolate the muddy area. But then as they get comfortable, their voices start to raise and during the enthusiastic parts they get really excited to share their story and sometimes there are laughs like ha, ha and you can see that when you have this variation, it's good to even things out a bit. Sometimes they can even speak very quietly. Now I'm going to adjust the Q value or the width of my band to make sure that I isolate the muddiness as much as possible without affecting other frequencies. I'm then going to pull the frequency slightly down and hear the difference. Sometimes they can speak very quietly or sometimes they get so excited about sharing their story that their voices get louder and louder. To start off a little bit on the quiet side because they're a little shy, but then as they get comfortable, their voices start to raise. During the enthusiastic parts, they get really excited to share their story and sometimes there are laughs like ha. As you can here, now the voice sounds much clearer. You can turn the effect off by clicking this button here and turn it back on to hear the difference you've made. After you've made some cuts or realize you don't need to cut any frequencies. Now is the time to boost any frequencies that are not present enough, also known as sweetening the sound. For example, if your voice sounds thin, you can increase the frequencies in the 200-300 hertz range to add more warmth. If your voice lacks clarity, you can increase the frequencies in the 1-5 kilohertz range. For example, in this clip, my voice sounds dull. It takes some practice to know what problem areas sound like. The key here is to play around and trust your ears. I'm going to boost the 1-5 kilohertz range. It takes some practice to know what problem areas sound like. The key here is to play around and trust your ears. If your voice sounds good to you, then no need to boost any frequencies as that can bring in additional sound artifacts that you don't need. It takes some practice to know what problem areas sound like. The key here is to play around and trust your ears. You can start with Band 1, narrow the Q and do a sweep, hearing if any bad sounds come out. Experiment, what does the voice sound like if I make a gentle cut to this range? What happens to the voice if I narrow or widen the cut? It's going to take you practice to hear the differences and frequencies. But I promise, the more you practice and listen, the more you'll be able to hear these differences, clearly. Use the EQ map as your starting point and friend. 8. STEP 3: Compress Your Dialogue: The third action you can do to improve your sound is compression. Pay attention to this section because compression is the most challenging element, in my opinion, to understand in the sound editing process. But have no fear. I'm here to break it down in the clearest, most simple way possible. Compression is a form of processing that reduces the dynamic range of your audio clip. Which means the difference in loudness levels between your loudest parts and your softest parts. When you record someone speaking, for example, in an interview, there's usually some variation in the loudness levels of your wave form. They can start off shy and quiet. But as they get comfortable and are talking about something that really excites them, they can start talking louder. Sometimes there's a laugh which just really pops out in your waveform. While it's good and natural to have some dynamic range, in video and film, it often sounds better when you even things out a bit. This makes it easier for the audience to hear what is being said. They don't have to lean in to hear the quiet parts. In fact, if you watch The Matrix movie, which has excellent sound editing, you'll hear how the whispering parts are not so different in loudness from the normal talking. By using a compressor, you even out the dynamic range of your audio clip by bringing down the loudest peaks and raising up the lowest ones. This makes a voice sound tighter, more present, more forward, which cuts through the mix more. To make the clearest example, I have recorded myself. I'll play it now so you can hear how it has soft parts and loud parts. Sometimes your interview subject can start off quiet because they're shy. But as they get comfortable, they get more and more excited and enthusiastic about sharing the story. Sometimes, hahaha, there can be laughter and transients like, yeah. Now, I'm going to add a two-modeled compressor to my effects rack. Can you see how the waveform now looks more condensed? Or in some editing terms, fat. This is without compression and this is with. Now hear the difference, this is without compression. Sometimes your interview subject can start off quiet because they're shy. But as they get comfortable, they get more and more excited and enthusiastic about sharing the story. Sometimes hahaha, there can be laughter and transients like, yeah. This is with. Sometimes your interview subject can start off quiet because they're shy, but as they get comfortable, they get more and more excited and enthusiastic about sharing the story. Sometimes hahaha, there can be laughter and transients like, yeah. Let's get into these settings. A compressor is made up of four elements: the threshold, the attack, the release, and the ratio. The threshold is the decibel level at which the compressor begins to work. If I set the threshold at minus 12 decibels, that means that any peaks bouncing below minus 12 will not be touched by the compressor. Only those that break above minus 12 will be compressed. Next, we have the ratio, which tells the compressor by how much to bring the peaks down if they exceed the threshold. For dialogue, set this at three to one. This means that for every three decibels peaked above your threshold, the compressor will only allow one decibel to pass. If you set your threshold at minus 12 decibels and one of your peaks is playing at minus nine, your compressor will reduce that peak by minus two decibels to play at minus 11. If one of your peaks is playing at minus six decibels, your compressor will reduce that peak by four decibels to play at minus 10 decibels. Third, we have the attack setting, which tells the compressor how fast it's going to kick in and compress any signal that exceeds your threshold. Usually for dialogue, you can have a fast attack at two milliseconds. This means that as soon as a peak exceeds minus 12 decibels, the compressor will start compressing the signal down after two milliseconds. If your attack time is too slow, it will let some peaks above your threshold pass through. For example, I'm going to set the attack to 100 milliseconds. Sometimes hahaha, there can be laughter. Do you see how the peaks go above minus 12 decibels? Because the compressor is not grabbing those successive peaks fast enough. Now, I'm going to set the attack back to two milliseconds. You'll see it will not go pass the threshold at minus 12. Sometimes hahaha, there can be laughter. Fourth, we have the release, which tells the compressor how long it's going to hold the signal down after it begins to compress. For dialogue, you can put this at 100 milliseconds. That means that the compressor is going to hold any peaks that exceed the threshold down for 100 milliseconds. Now let's put these settings all together. We already decided to put the ratio at three to one, the attack at two milliseconds, and the release at 100 milliseconds since those are good settings for dialogue. Now is a question of where to set the threshold. To do this, first eyeball a general decibel level at which your average peaks are playing at. In this clip, I see that the peaks of myself, this section here, is playing at around minus 18 decibels. That means that I want to bring down all of the peaks that are playing above minus 18 decibels, so the overall loudness of the entire clip is more even. I'm going to start out with that as a threshold. Listen to how the voice changes as I move the threshold down to minus 18 decibels. You should hear a tightening up. To get more and more excited and enthusiastic about sharing the story, and sometimes hahaha, there can be laughter and transients like, yeah. Now, you may see that there's this meter flushing red. This is known as a gain reduction meter and it only flushes when the compressor is working. Also telling you by how many decibels is reducing the peak by. I'm going to play the clip again and watch how it doesn't really flash red during the quiet parts, but more during the parts that are louder. Times your interview subject can start off quiet because they're shy. But as they get comfortable, they get more and more excited and enthusiastic about sharing the story. Sometimes hahaha, there can be laughter and transients like, yeah. As you can see, the gain reduction meter is only flashing during the times where the peaks are exceeding minus 18 decibels. Showing me by how many decibels they are being reduced by. Hit "Apply" to look at your waveform. As you could see, my waveform now looks more even. Then, listen to the clip. Sometimes your interview subject can start off quiet because they're shy. But as they get comfortable, they get more and more excited and enthusiastic about sharing the story. In sometimes hahaha, there can be laughter and transients like, yeah. As you can hear, the clip sounds more even, the loud parts and not poking out as much, which means we set a good threshold. If you find that there's still a distinct difference between the loud and soft sections of your dialogue, bring down your threshold even more to make them more even. Make sure though not so set your threshold down too much. You remove all of the dynamic range and your voice sounds monotonous or robotic. Sometimes your interview subject can start off quiet because they're shy. But as they get comfortable, they get more and more excited and enthusiastic about sharing the story. As you can see, because I have brought some of the highest peaks down, the overall loudness of the entire clip went down. The clip sounds much softer. This is where make-up gain comes in. This tells a compressor by how much you want to raise everything backup again after you have brought the peaks down. What this will do is raise the loudness of the entire clip, including the quiet parts of your dialogue that were not brought down by the compressor. I'm going to raise it now until my dialogue is bouncing between minus 15 and minus 12 decibels, where it was bouncing before I added the compressor. Sometimes your interview subject can start off quiet because they're shy. But as they get comfortable, they get more and more excited and enthusiastic about sharing the story. Sometimes hahaha, there can be laughter. Sometimes your interview subject can start off quiet because they're shy. But as they get comfortable, they get more and more excited and enthusiastic about sharing the story. Sometimes hahaha, there can be laughter. Sometimes your interview subject can start off quiet because they're shy. But as they get comfortable, they get more and more excited and enthusiastic about sharing the story. Sometimes hahaha, there can be laughter. In summary, to do compression, first, set your threshold, ratio, attack, and release. Then use makeup gain to bring the loudness of your clip back to your target level. Just a note on the effects rack. You don't need to hit apply for it to be working. If the switch is toggled on and green, that means you're previewing the effect. However, it's only when you hit Apply that the waveform will change its shape. You can always undo this. You can go back and tweak your compression settings before it makes the changes into your original source file. The effects on Premiere Pro, on the other hand, are non-destructive. They will not change your original source file. 9. Compression Example 2: That was more of a dramatic example of loudness fluctuations that I use to show the effect of the compressor more clearly. Let's now take a look at Dottie's case where she spoke for the most part evenly. As you can see from the waveform, it's pretty even. There are just these transient peaks that come through that I'd like to bring down a bit. For example, in this section, you can hear how the no books, no pictures, is hard to hear and how the no cartoons jump out. We grew up with those riddles and stories, because we had no television, no books hardly, no pictures, no cartoons, of course. Likewise, in this section, you can hear how walked and five miles jumps out a bit. I was pondering depression in 1933 and so when I was little and we used to go to school we walked about five miles. Like the previous example, I'm going to eyeball a threshold to start with. It looks like on average, things are peaking at minus 12 decibels. I'm going to use that as my starting threshold and hear what it does. We grew up with those riddles and stories because we had no television, no books hardly, no pictures, no cartoons, of course. As you can see from the gain reduction meter, there was not much compression going on. If I hit "Apply", you can see how there's not much change. The higher peaks still pop out. You can also hear how some words still jump out. We grew up with those riddles and stories because we had no television, no books hardly, no pictures, no cartoons, of course. I was pondering depression in 1933 and so when I was little and we used to go to school, we walked about five miles. I'm going to increase the threshold a bit more until I hear a difference. We grew up with those riddles and stories because we had no television, no books hardly, no pictures, no cartoons, of course. This sounds good to me at minus 17 decibels. As you can see, if I hit "Apply", things now look more even but we're still preserving some of the dynamic range, so it sounds natural and not robotic. We grew up with those riddles and stories because we had no television, no books hardly, no pictures, no cartoons, of course. Now, because I've reduced the decibels of many peaks, things sound quieter. I'm going to push the makeup gain to bring everything back to where it was before the compression started. When her voice was bouncing around minus 12 to minus 15 decibels. We grew up with those riddles and stories because we had no television, no books hardly, no pictures, no cartoons, of course. We grew up with those riddles and stories because we had no television, no books hardly, no pictures, no cartoons, of course. We grew up with those riddles and stories because we had no television, no books hardly, no pictures, no cartoons, of course. I was pondering depression in 1933 and so when I was little and we used to go to school, we walked about five miles. I was pondering depression in 1933 and so when I was little and we used to go to school, we walked about five miles. Her voice sounds more present, tight, and even, which will also help it come through the music playing underneath. The two modeled compressor works the same way in Premiere Pro. You won't really be able to see the difference it makes to the waveform as clearly as you can in Audition. However, it's a great way to train your ears. 10. STEP 4: De-ess Your Dialogue: The fourth action we're going to do to improve our sound is soften any harsh siblings found in our speech. To do this it's simple. I have a clip here with harsh siblings. Make sure you reduce your volume now, so you don't hurt your ears. Perhaps you're in a transition, or you just don't feel like where you are now is authentic to being. I'm going to add a DeEsser effect. As you can see, it's the compressor just for one specific region. First, isolate where your siblings is happening by playing your clip and hitting pause when the harsh siblings, as being said. Then look at where the sibling's peaks in the frequency chart. You can adjust the endpoints of the frequency window to isolate it. Then adjust the threshold until it softens, but doesn't become too dull. Perhaps you, perhaps you, perhaps you, perhaps you, perhaps you, perhaps you, perhaps you, perhaps you, perhaps you, perhaps you, perhaps you're in a transition or you just don't feel like where you are now is authentic to your being. Perhaps you're in a transition or you just don't feel like where you are now is authentic to your being. 11. STEP 5: Balance Your Mix: The fifth action we're going to do to improve our sound is making sure our music and sound effects are well balanced in our mix. We're now going to focus on the music to make sure it mixes well with our dialogue. The first step is to level your music so it's not overwhelming the dialogue. A good way to familiarize yourself with a good mix is to play a film that's in the same genre as yours, and just listen. From me, that's Chef's Table on Netflix. What is the loudness difference you here between the spoken dialogue and music? How are the sound effects weaving in? Here is the clip from my documentary, Dottie. I have three layers of sound here, her vocals, the ambient wind rustling through the trees, and the music. Well, life is could be like a dream. You don't get it at all, but it's like a flash of something that is there but gone at the same time. As you can tell, the music is overpowering. First, I'm going to open the audio track mixer and go to the music track. I'm going to use a fader to reduce the loudness of the music. I'm going to mute the wind rustle ambiance for now. [MUSIC] Well, life is could be like a dream. You don't get it all, but it's like a flash of something that is there but gone at the same time. Her voice, although much better than before, still feels overtaken by the music. I could lower the loudness of the music, but then the music gets too soft. Well, life is, could be like a dream. You don't get it all, but is like a flash of something that is there. I'm going to EQ the music, making a cut where Dottie's vocal frequencies lie so it'll give her vocals more room. To do this, I add a parametric EQ effect to the effect rack for the music track. Well, life is could be like a dream. You don't get it all, but it's like a flash of something that is there. Now, there are portions of this clip where there's music but no talking. A cool trick with the essential sound panel in Premiere Pro is auto-ducking. This will automatically generate loudness keyframes for your music so it'll play softer when there's dialogue present. To do this, make sure you have designated your dialogue track as dialogue. Then click on your music track and designate that as music. Check ducking and specify by how many decibels you'd like the music to play softer when there's talking, and how gradual you'd like the ducking to occur. Then hit Generate Keyframes. As you can see, it automatically ducks my music when Dottie is talking. Well, life is, could be like a dream. You don't get all, but it's like a flash of something that is there but gone at the same time. Play with these settings until you are happy with your results. I hope you're feeling excited now as your mix is coming together. Now, we're going to address sound effects. Like music, you don't want this to be too quiet or overpowering. You can also EQ your sound effects to make more room for the vocals. For example, here in DaVinci Resolve, I have a meditative art film I made as part of a 24-hour film challenge. I'm showing this example with DaVinci Resolve to demonstrate how the principles learned in this class can apply to any video editing platform with comprehensive sound editing capabilities. It feels exhilarating like I'm ice-skating on this knife-edge of the unknown. As you can tell, there's also this ice skating sound effect that overpowers my talking. I reduce some of its frequencies that competed with my voice frequencies. Now you can hear my voice coming out much better. It feels exhilarating like I'm ice-skating on this knife-edge of the unknown, where anything could happen. 12. Final Thoughts: I want to congratulate you for finishing this class because that was a lot of material. I hope you found these lessons illuminating, and now feel excited to play with your sounds. I also hope you are leaving with the ability to hear sounds in new, more fascinating ways in your everyday life. Remember to post a before and after of your video clip on the projects page, so we can hear the difference you've made to your sound. Please check out my other classes on my Instructor page if you're looking to further your film-making or creative journey. Thank you so much for taking this class, and know that I am cheering you on. Happy sound editing. 13. Newsletter Sneak Peak: [MUSIC] If you enjoyed this class, I invite you to leave a review and sign up for my newsletter. This is not a ordinary newsletter, but instead a virtual tea-house of wonders where I share curated inspiration, behind the scenes, updates, and more high-value resources on the art of authentic creative living. It is my most intimate space to spoil my readers with delight. Sign up to receive on my course instructor page. [MUSIC]