Mastering with iZotope Ozone 5: Dithering (Part 10)

Transcript:

Hey guys, Eric Tarr here. I’ll be discussing how to use dithering as the last stage of mastering. I’ve been using the Ozone 5 plugin to master my song.

[song being mastered with iZotope Ozone 5]

Dithering is one of the most complicated concepts in mastering because the result of dithering is to actually add noise to your signal to make it sound better. This might be counterintuitive because normally noise is bad with audio. The thing to understand is the addition of noise is inevitable when you print a track whether you use dithering or not. With dithering, you at least have control over noise to make it almost impossible to hear. Without dithering you run the risk of the noise being audible to listeners.

Why does noise come in when you print a track? The signal you’re printing is going from higher bit-depth to lower bit-depth and the resolution or detail of a signal is reduced. Your computer rounds up or approximates the amplitude of the signal to fit the lower bit-depth. The resolution of a CD WAV is 16-bit — plenty of resolution for playback purposes of audio. This is usually the bit-depth you want to print to when bouncing a song from your DAW. Although 16-bit is sufficient for playback purposes, there are advantages to recording at higher bit-depths, such as 24-bit for processing purposes.

Whether you record at 16 or 24 bit, you still need dithering. Your digital audio workstation is going to process the audio signals internally at 32-bit floating point regardless of whether you record signals in your DAW at 16 or 24 bit. Your DAW is automatically giving you a higher resolution for internal processing.

There’s always a need to have this final conversion from 32-bit to 16-bit. Then the round off of the resolution. This approximation can add noise to your signal because sometimes the amplitude gets rounded up and sometimes rounded down. Without dithering, these rounding errors are correlated with your audio signal. Assuming your signal contains a lot of energy in a frequency range that’s easy to hear, the rounding errors could also be within a frequency range that’s easy to hear. With dithering, you intentionally add noise that’s difficult for the listener to hear. This is because dither noise can be shaped, it can be in a frequency in a range that is difficult for humans to perceive.

Let’s look at dithering in Ozone 5. I’m working inside of the maximizer module of Ozone 5. Here you’ll find the dithering capabilities of the plugin.

I’ll turn dithering on. I’m adding in noise now that’s uncorrelated with my song and more difficult to hear. I’m using the MBIT algorithm but there are several other algorithms. Inside each algorithm you actually have several parameters for shaping noise to make it even less audible. To demonstrate these I’m gonna switch to the dither plot, so you can see the frequency spectrum of the noise I’m adding. I have noise shaping off, so this is noise that has a flat frequency response. I can turn on noise shaping. I’m scooping out some energy of the noise that’s in a frequency range that’s most audible for human hearing. There’s more energy of the noise in the low frequencies and really high frequencies. As I turn on noise shaping even higher, more high frequency noise gets added in. This is gonna be difficult for humans to hear. So I can go all the way up to ultra.

You can also change bit depth. If you’re gonna print to 16-bit audio, leave it on that. But if you’re using DVD audio then you might want to use 24 bit. You can also use, if I turn 8-bit audio, that all of a sudden the noise becomes more audible. I’m gonna switch it back to 16-bit. And you can change the dither amount, whether you want it to be extra dither, none or low. I’m leaving it on normal. Show you some of these other algorithms. Again, you can have noise shaping. Each algorithm has its own type of noise shaping. For the most part, they’re all gonna be better than working without dither. The more noise shaping you use, it’s gonna put the energy in these high frequencies and more difficult to hear. You can change the number of bits, the amplitude of the noise that you’re adding in from 1 to 2 bit. You can see as you get into more complicated algorithms and noise shaping, primarily focusing the energy in higher frequencies.

Last thing is DC Filter. This makes sure there’s no offset going to my speakers so my speakers are at a steady point of no energy, rather than having them pushed forward and oscillating or pushed back. That’s DC Offset you want to filter out. That’s all there really is to the dithering module of Ozone 5. That concludes the mastering with Ozone 5 series.

Eric Tarr

Eric Tarr

Eric Tarr is a musician, audio engineer, and producer based in Columbus, Ohio. Currently a Professor of Audio Engineering Technology at Belmont University in Nashville, TN.
Smiley face
Recommended