How to Make Sound Recordings Not Sound Like Shit – A Beginner’s Guide.

Firstly, a disclaimer:

I’m not all knowing, so don’t take my advice as gospel. Think of it merely as a guide to lead you not in the completely wrong direction.

That being said though, I have been doing Sound Design for animation (and motion picture in general) in some form or another (sound effects editing, mixing, recording, voice over recording, etc.) for the better part of 5 years. And, as I’m yet to get any of my projects on anything aside from the internet, I’d say that internet-based multimedia is what I’m best at.

Side Note:

I do apologise if this guide comes off as a little bit “ranty”. Hopefully my point will still come through though.

Anyway, onto the guide:

1. Microphone!

First things first – a golden rule: The sound quality which you will end up with will only be as good as the lowest-quality device involved in the recording. Now before I go spurting off fancy words like “signal chain” and “signal to noise ratio”, the bottom line is this: If you have an awesome microphone and you record into a crappy sound card, you’ll get a crappy result. Likewise, if you have a crappy microphone and an awesome sound card, you’ll get a crappy result.

The only way to overcome this problem is to do something which everyone who does stuff for free hates – spend money. This is hardware we’re talking about. No matter how many torrent sites you visit, you can’t download a microphone.

For web based animations, your end result isn’t going to be some highly-dynamic ultra awesome cinema mix, so you don’t need professional quality gear to have fair results. Generally a Dynamic microphone will do the trick nicely. You can pick up a fairly decent Cardioid Dynamic Microphone from most electronic stores for around $20 - if “Cardiod Dynamic Microphone” is too long and difficult for you to remember, just ask for a karaoke microphone instead. They’re the same thing.

I can’t stress enough how important it is to get a fairly decent microphone for your voice recordings. A headset mic, a gooseneck mic or any other small-diaphragm microphone like that will not suffice. They are designed to be used for voice chat, and they give off about the same quality as a telephone call. You don’t want your epic Mario vs. Sonic monologue battle to sound like they’re on the phone to each other, do you?

Now this is going to sound harsh, but this is basically the bottom line: If you aren’t prepared to buy a microphone other than a headset mic, then give up all hope of ever getting any sound quality that doesn’t sound like a hat full of assholes.

Side Note:

If you’re looking to spend a bit more money, go and buy a Shure SM57. They’ll set you back around $150, but their quality (and versatility) is unmatched for the price. If you’re recording into a motherboard soundcard, or a gaming soundcard, you’ll also need to buy an XLR to 3.5mm minijack cable.

Ok, so you’ve gotten your microphone. Now, you need a few other things whilst you’re still at the shops. Go buy a pair of stockings and a wire coat hanger. Why? Not for funny business. Trust me on this one. You’ll need them later.
On to the preparation stage.

2. Acoustics

A microphone is a device that will record all sound – it won’t just record your voice. So any background noise that you hear within your recording area, there is a very high chance that your microphone will pick it up. And of course you don’t want to hear someone mowing their lawn in the background of your epic Sonic vs. Mario showdown, now do you?

To stop this, you need to prevent as much exterior noise getting into your recording area as possible. Ideally, you’d use a soundproof booth, but we’re not all rolling in money, so I doubt this is possible. Generally it is best to close all windows within the area, turn off any washing machines you have going, shoot your neighbours that are mowing the lawn, etc. etc. If there is too much noise within the area, wait until there is less. I generally do most of my voice recordings at 2:00am, because that’s the only time all the damn wildlife shuts up.

Now if you have a noisy computer, it is often a good idea to put some blankets around it, or other acoustically dampening material. I DO NOT recommend doing this for an extended period of time however, as it will probably make your computer overheat and, quite possibly, explode. I know it goes without saying, but remove all of the blankets and whatnot as soon as you’ve finished recording.

3. Recording

Firstly, you’ll need to set up your microphone. If you have a microphone stand, then use that. Otherwise, you’ll need to find some way of keeping the microphone in a steady position. I don’t advise having the voice actor hold the microphone, as it is easy for them to become distracted and move the microphone around. This leads to variances in volume level, and handling noise of the microphone.

If you have a friend who can kindly hold the microphone for the voice actor, and they have a steady hand, then use them. This will eliminate volume variances, and hopefully most handling noise.

Next, you’ll need that pair of stockings and coat hanger you bought earlier. Stretch the pair of stockings over the coat hanger, and position it between the voice actor and the microphone (this generally works best if you have a microphone stand). This is more or less a cheaper version of a “pop filter” – a device which eliminates the loud “KAFOOMP” sounds you hear on poor recordings when people use plosives (“p” sounds, “s” sounds and other words with sharp attacks). If you’re willing to buy a proper pop filter, it’ll set you back around $30 from any professional audio store.

Now you’ll need to plug the microphone into your computer. I can’t help you here. You’ll probably need to plug it into the back of the tower, so get out a flashlight and go hunting for that mic input. Chances are that it’ll look exactly like a headphones port, but will be a different colour. Consult your computer/sound card manual for where the input is located.

Now you’ll need some software to use for recording. I personally use Pro Tools, but that isn’t cheap, and you can’t acquire it through non-legal means as it needs an actual external hardware device for the program to even boot up.

So, I’d suggest using Audacity. It’s a freeware audio editing program. Download it from audacity.sourceforge.net/


Firstly, within Audacity, you’ll need to edit some Preferences. Go to Edit-> Preferences to do this. Then, click on the “Quality” tab. Change the “Default Sample Rate” to 48000Hz and the “Default Sample Format” to 16-bit. This brings your recordings in line with international audio standards for video projects. Feel educated. You’ll probably need to restart the program for the changes to take effect.

Next, in the top left toolbar on the screen, you’ll see a picture of a microphone with a volume slider next to it. Just to the left of that again, you’ll see a drop down menu. From that menu, choose “microphone”.

Now, you’ll need to set an appropriate recording level. The middlemost toolbar at the top of the program has two sets of volume meters on it. One is for playback (indicated by a speaker), the other is for recording (indicated by a microphone). Click on the down arrow next to the microphone, and click on “Monitor input”. Providing that you’ve plugged your microphone into the right port on your computer, you should see a little red bar filling the volume meter. This indicates the current voice recording volume. Notice that when you speak into the microphone, it “bounces” up and down? That’s the level of volume coming into the computer from the microphone.

You’ll now need to perform something called a “gain check”. It is used to optimise the microphone’s volume level for recording. Get your voice actor to speak into the microphone at the loudest they will throughout the recording. Ideally you want the peak level of volume coming into the computer to be as close to the right hand side as possible, without it hitting 0.

When the volume meter hits 0, it causes “clipping”. This creates very bad distortion that makes sound recordings sound like shit. Seeing how this guide is called “How to Make Sound Recordings Not Sound Like Shit”, it seems like something you wouldn’t want to do. Therefore change the microphone’s volume settings to stop clipping this by moving the volume slider next to the picture of the microphone.

Side Note:

Generally, it is good to do any particularly “loud” lines (such as Mario and Sonic screaming at eachother, Dragon Ball Z-style. Because I just *know* that that’s going to appear in your animation) in separate takes, so that you can set the volume level accordingly for them.

Now, here comes the big moment. Time to record something.

When your voice actor is ready, hit the button with the red circle in it. This is (obviously) the “record” button. You’ll notice a line moving towards the right of the screen, drawing in a squiggly line behind it as it goes. This is a visual representation of the audio that it is recording. If the audio “peak” reaches the very top or bottom of the track, then it has clipped. If this happens, start again, and here’s a virtual slap in the face for making a sound recording that sounds like shit. *slap*.

4. Clean-up

Now say that even though you’ve followed this guide to perfection, you still manage to get horrible quality sound recordings. Audio clean-up isn’t really a delicate art like image clean-up. Any changes which you make will be highly brutal changes that will drastically alter the overall properties of the voice recording. The most common way to reduce noise is to use an Equalizer, or an EQ.

In Audacity, to use the EQ, select a segment of the audio, and then Click “Effect -> equalization. You’ll get a window pop up that has a sort-of line graph at the top. The pitch of the sound (frequency) is mapped horizontally, and the volume is mapped vertically. You can alter the volume of certain frequencies of the sound by drawing points and moving them within the EQ graph.

The most common usage of this EQ filter is to remove the bottom end of a voice recording. This is a more barbaric way to remove those “pop” sounds that you’d hear if you weren’t able to make a pop filter. Generally if you set the graph to the lowest point possible at around 150Hz, it should remove all of those horrible sounds.

Whilst it will get rid of those popping sounds, it will also take with it any sound below that range. So if for some reason, you want Sonic or Mario to sound like Barry White, this will pretty much ruin the voice take.

There are more advanced EQ techniques for finer audio editing, but they are too difficult to explain here. There are tutorials available online if you want to pursue this further.

Side Note:

There are also “Noise Reduction” plug-ins available for some audio editing programs. Whilst some of them are very good at removing microphone noise and other artefacts, they can also make your recordings sound like they’re coming from a fish tank. Use them with caution.

5. Compiling

Now that you’ve gotten all of Sonic and Mario’s lines recorded for your epic faceoff, you need to compile them into a “scratch track”. Now I’m not going to go into detail here, as I’m not all that familiar with the workings of Audacity. There are plenty of free tutorials on how to use the program available on the intertubes, however.

Basically, within the scratch track stage though, you’re aiming to time out the voice actor’s lines appropriately. It takes a bit of practice to get it sounding like a normal, flowing conversation, but essentially the best way to achieve this is to listen to the lines both individually and as part of the entire scene for pacing purposes.

6. Post Production

So now that you’ve made your awesome showdown between Sonic and Mario, you’ll need to add in sound effects and whatnot. I don’t think that Audacity has the ability to import a video file, so you’d need to look to a program such as Sony Acid. You’ll also need to convert your animation into an Acid-friendly video format. Swf is not one of these. Generally I’ve found the easiest way is to export the animation from flash as a PNG sequence, and then create a video from that PNG sequence in a video editing program. I achieve that through Sony Vegas – but once again, you’re on your own for working that one out.

I’m not going to go into detail on how to use Sony Acid. Rest assured though, if you can work out how to use Flash and Audacity, you should be able to work out Acid. What I will give tips on though is the overall construction of sound for a scene.

Firstly, you’ll need to create or source an atmosphere sound for the scene. Seeing as I’m sure that Sonic and Mario will be flying in mid air with no obvious explanation as to why they *can* fly, you’ll need some wind blowing, or other general ambience.

Next, you’ll need some key sound effects, such as them punching each other and whatnot. You can source a whole heap of sound effects from the internet. www.findsounds.com is the best for this, but don’t expect every sound effect from that site to be a quality result.

Then obviously you’ll need music, but I’m sure some form of a Sonic or Mario remix will be in order. There are a lot of free music tracks available on websites such as www.newgrounds.com , and I’m sure you’ll be able to acquire any other track you need through various means.

Finally, you’ll need to mix the damn thing. There are no “real” rules to the mixing of a soundtrack, EXCEPT that for an animation, dialogue ALWAYS comes first. If you can’t get the dialogue intelligible, no one will be able to understand Mario’s inner turmoil against… hell, I don’t know. Just make sure that the dialogue is intelligible.

Once again, make sure that the entire final mix doesn’t clip. This will be indicated by a small red light on the peak meter in acid. If it’s clipping, you’ll need to adjust sound levels to stop this. You can cheat and use a dynamic compressor, but once again they’re tools that should only be used by people who know exactly what they’re doing. And I’m not going to explain that as it’ll take far too long.

Side Note:

As a general tip throughout the sound recording, editing and mixing process, keep everything in lossless formats – don’t save sounds as MP3’s. Save them as .wav files. To draw a comparison to the visual side of things, you wouldn’t repeatedly save a photoshop picture as a JPEG after every time you’ve worked on it, would you?

As a final tip, get other people to listen to the mix before it is released – preferably people who have never heard it before. That way they’ll be listening without previous viewings and no bias. As such they’ll be able to give you a rough guide as to where there are faults within the mix, and what needs to be adjusted, etc.

Conclusion

Hopefully this guide has helped you to have better sound quality in your animations. I find that in about 90% of the web-based animations I’ve seen (ESPECIALLY on newgrounds), the soundtrack as a whole is absolute balls. So I hope that through reading this, there’s less of a chance of your animation becoming a number within that percentage.

I’m sure Mario and Sonic will thank you for it. : )