Dithering, Is It Really Needed?

Dithering. I remember learning about this in college. I remember it being a subject that was easy to understand within the confines of a software processing environment. It was simply a way of taking a large binary word and turning it into a smaller one to save hard drive space without compromising the quality of the file. It’s used in graphics, video, and audio. But how does this apply to my home studio? It can’t be that important, right? I mean, I never see anyone giving instruction on dithering when they’re printing their final mix. To be honest, I have no idea why they aren’t talking about dithering.

Depending on your recording format, dithering can be an essential step to make sure that your final mix is as good a quality as your initial recording. Here’s how it works.

audio_bit_reduction_from_24_bit_to_8_bit_with_and_without_ditheringThe gold standard for audio fidelity is still the compact disk (CD). It’s been that way for many years and I don’t think it’s likely to change anytime soon. CDs offer superior audio clarity and quality in a file size that can be processed by even the most rudimentary processors by modern standards. Remember, CDs are still a purely digital format. That means that unlike vinyl or tape, they use “0”s and “1”s in binary to store audio information and convey it to the CD player. In doing so, they require a fixed bit depth and sample rate. The bit depth is the size of the binary word that’s read by the processor. The sample rate is how fast the processor can read those words. CDs use a word length of 16 bits and a sample rate of 44,100 Hz. This means that when the music is being converted from analog to a digital format the computer is taking 44,100 samples per second and each one of those samples contains 16 bits of information. The CD player uses the same depth and rate to read the CD and play the audio. The problem is that most studios (commercial or otherwise) record at a much higher fidelity than 16/44.1.

I typically record 32-bit/float at 48,000 Hz. I find this to be a good balance of audio quality without being too taxing on my CPU. Especially when I have 30 tracks, most of them with one or more plugins on them. But in order to make the file smaller and compatible with every other device on the planet, I need to reduce the size of my final mix from 32f/48 to 16/44.1. This is where dithering comes in.

On a side note, 32-bit float is a misleading label. The word size is actually 24-bits, but it contains an extra 8 bits tacked onto the end to accommodate decimals during processing. Every time you do something to a digital audio file – be it something as simple as changing the volume – the bit word gets recalculated buy the CPU. In a normal 16-bit or 24-bit word this can lead to degradation of the file quality due to the quantization error of calculations that don’t equal whole numbers. In other words, they have decimals. But a 16-bit or 24-bit word can’t use this decimal, so it’s truncated or dropped. This leaves you with an inaccurate calculation. Enough of these inaccurate calculations over the course of a mixing session and the file becomes an imprecise representation of the original audio it was derived from. 32-bit float allows for these decimals to be included in the calculation. But, they are still dropped when you print the final mix. This process gives you a much more accurate representation of your audio.

That being said, if we were to convert that final 24-bit word to 16-bits without dithering, we would simply lop off the last 8 bits. This would mean that we’d lose a third of audio information. We’d just be chopping it right off. But with dithering, we can use randomized noise to quantize that information into the remaining 16 bits. This random noise fools your brain into thinking it’s hearing more information than it really is. The end result is that when we do finally cut off those last 8 bits, a large part of our missing audio information appears to be retained.

Luckily, most modern DAWs have some sort of dithering routine included in them. Some are better than others, but they’re usually a standard feature. There are also several plugin options to choose from. But when should we use them? That depends on whether or not your project is being sent off for mastering.

Dithering, as well as sample rate conversion, is something that is usually handled at the mastering studio.  We didn’t even touch sample rate, so I’ll cover that in another article. When you send in your mixdowns they’ll usually ask for them in their original format. This way they can make sure that the dithering is done to the highest possible standards. If you’re planning on mastering your own project, or not mastering at all, then dithering should be the very last step in the mixdown before you print your final stereo mix for distribution.

If you have any other questions about dithering or mastering, please feel free to contact me at K14Studios.NC@gmail.com.



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s