====== six common mistakes in audio editing ====== I see six common mistakes in audio editing. These are mistakes made by beginners and even some experienced editors. Link to slides from the presentation: https://docs.google.com/presentation/d/1dnzIFLhmZWQ6B1xqXxxOYI8g8eGOOK35imY8ha-eJDE/edit?usp=sharing -------------------------------------------------------------------------- ===== 0: Garbage In, Garbage Out ===== I said I was going to introduce 6 common mistakes, but actually there's a potential mistake 0, a mistake that people make before they even start recording that can lead directly to several of the six common mistakes. That is: Garbage in, garbage out. What does it mean? What constitutes 'garbage'? Basically it means that if you start with bad audio, it's very hard to fix it. The other way around, if you start with good audio, you are less likely to run into the six common mistakes. So what can go wrong before you start? - Recording with the levels too high, so that you get clipping and distortion. - Recording with the levels too low, so that you boost noise when you normalize or compress it. - Starting the recording too late, so that you truncate the envelope. - Stopping the recording too early, so that you truncate the envelope. - Bad performance by whoever you are recording, so that you have weird dynamics or momentary clipping. Bad performance can cause all kinds of other problems, and remember that it's hard or impossible for good editing to save bad material. This means that when you are recording, you need to take care to avoid these mistakes. -------------------------------------------------------------------------- ===== 1: Clipping ===== Let's look at the editing mistakes one at a time. The first one can happen in at least two ways: 1. It arrives in the editor that way. 2. You accidentally cause it while editing. As soon as you load your sound in Audacity, notice what the waveform looks like. If it looks like this, with some of the bars reaching all the way to the top of its lane, you've got garbage from which you cannot recover, basically. Unless you're never going to be able to record that again, throw away that audio and record it again. Sometimes while you are editing, effects such as amplification or others can introduce clipping causing the bars to look like this. If you see this as a result, stop right away, undo, and try again without introducing clipping. Even one bar is no good. Don't use that audio. Learn more on getting the levels right when recording here: https://renickbell.net/doku.php?id=getting-the-right-levels -------------------------------------------------------------------------- ===== 2: Forgetting Normalization ===== The second one you can also see. When you open up the audio and it looks like this (too small), if you can re-record it, do so and leave this one aside. If you can't re-record it, you definitely need to normalize it. Even if it looks like this (more reasonable waveform), the first thing you should do is normalize it so that the audio is visible for editing and that you will be more aware of the noise or other details that are contained within it. Here's how your normalize audio... What about this dB thing? What's that? Setting dB on the left side of the waveform You'll repeat your normalization to the target dB at the end of the process as the last step to make sure that your file maxes out at the dB specified by the next person in the process. Learn more about adjusting the volume of a sound file here: https://renickbell.net/doku.php?id=audacity-adjust-volume -------------------------------------------------------------------------- ===== 3: Unnecessary Silence ===== The third one is straightforward: silence at the front or end of a sound. Why do we want to avoid this? When someone else wants to use our sound, there are use cases such as dropping it in video or audio editing software, or even loading it into a sample player. In the case of the editor software, they may automatically move a block to the correct position. If they cannot see the waveform inside well, they may align the block with other material in the editor. When they hit play, they will be surprised that the audio doesn't immediately start but is delayed by the amount of silence at the front. They may also align material to the end of the block, but that will also be wrong if there's a long period of silence at the end of the block. Either way, you are creating extra work for the next person to use your sounds, even if it's you. To avoid this, trim the excess silence. This is how you do it... -------------------------------------------------------------------------- ===== 4: Damaged Envelope ===== When you are trimming the audio, it's very easy to make a mistake and accidentally trim too much. This is our fourth commonly-made mistake. The number one reason for doing this mistake is trying to trim closely but not zooming in enough to see what you are doing accurately. Let's take a look at the zoom controls, which are here... Another way that this can happen is when you have multiple edit points in the sound and do the same thing in the middle of the file. Learn more about the envelope tool here: https://renickbell.net/doku.php?id=audacity-add-envelope -------------------------------------------------------------------------- ===== 5: Forgetting to Remove Noise ===== Our fifth common mistake is forgetting to remove the noise. Let's examine the noise. The origin of this is actually related to the early warning: garbage in, garbage out. However, in practice it's nearly impossible to get a perfectly clean recording, even in what you feel is a silent studio. There are all kinds of things, from computers to ventilation, that add noise to the signal. When you are putting together a complex mix with many tracks (think layers in Photoshop), each bit of noise can add to the others to produce a truly noisy background to your mix that you don't want. This often shows up even in otherwise professional videos. There's some voiceover edited in, and when vocal snippets come in and out, you can hear them preceded by a bit of hiss, and maybe the hiss continues after the speaking is finished. Then the audio cuts out, leaving pure digital silence. When the vocal returns, the hiss often returns first and continues after the next vocal clip is finished. It's super annoying to listen to and makes your work sound unprofessional. To prevent this, we want to remove the noise from our audio. We can use two goals, in the following order, to achieve this: 1. the noise removal effect 2. the noise gate effect You'll have to be very careful, though, because removing noise is destructive, and you could inadvertently damage the audio too much, reducing the quality or making it unusable. Remove as much as you can without creating noticeable damage to the clarity of the audio in the case of noise removal and envelope damage in the case of the noise gate. The noise removal effect is to remove a regular noise, such as the hum from an AC or a computer fan, from a recording. The noise gate effect is to remove the noise between pieces of audio which we want to use. Learn about noise reduction here: https://renickbell.net/doku.php?id=audacity-noise-reduction -------------------------------------------------------------------------- ===== 6: Extreme Dynamics ===== The sixth mistake is one of the harder to avoid and one of the trickier to correct in that the right amount of dynamics is determined both by the use case and the taste of the listener. It's when the dynamics are too extreme, meaning there is too much of a difference between the quiet parts and the loud parts of the recording. What is dynamics? When are the dynamics too extreme? When you are doing vocals and voice overs, generally we don't want a lot of dynamics unless it's very dramatic music like in classical music or in some kinds of movies. It's generally the same for instruments, though we can tolerate dynamics a little more in the case of non-vocal sounds. Why are big dynamics no good? It makes vocals harder to understand, and it can have listeners frequently reaching for the volume controls when it's either too quiet or too loud, sometimes because they've just previously adjusted it for one of the two cases. No one likes to ride the volume. Part of this problem can be controlled during the performance. If the person you are recording can manage to control their volume range, and if you can manage to keep the microphone distance consistent, you're going to be able to reduce the severity of this problem. The second thing you can do is use a compressor. A compressor brings down the volume of the peaks (compresses them) and then is usually followed with a gain adjustment (think normalization) to bring up the level of all of the audio, so that the resulting audio is generally louder then before, with all of the details now accentuated. Learn more about compression and limiting here: https://renickbell.net/doku.php?id=compression-limiting -------------------------------------------------------------------------- Now that your audio is in good shape and you are ready to export it and pass it on, you need to be sure about the format. First, check the spec about sample rate and so on from the person to whom you are providing your audio. Be aware of the usage. If you just need them to check the audio and want them to be able to access it quickly without wasting drive space, use mp3. That's also a good choice for audio podcasts. However, if this is going to be part of a bigger audio or video project, export it as a lossless file, like wav. Finally, avoid spaces and non-alphanumeric characters in the file name other than dash or underscore. These other characters can sometimes cause problems, especially in file transfer, and it would be a tragedy for someone not to be able to download or use otherwise perfect audio that has taken a lot of effort on your part! Learn more about file formats here: https://renickbell.net/doku.php?id=sound-file-formats