AF’s Weblog

February 25, 2011

Much Ado About Dithering

Filed under: Mastering — Tags: , , , , — audiofanzine @ 10:31 am

It’s a dirty job to go from high-res audio to 44/16, but someone’s got to do it.

The ultimate form of digital audio used to have a 16-bit word length and 44.1 kHz sampling rate. Early systems even did their internal processing at 16/44.1, which was a problem — every time you did an operation (such as change levels, or apply EQ), the result was always rounded off to 16 bits. If you did enough operations, these roundoff errors would accumulate, creating a sort of “fuzziness” in the sound.

The next step forward was increasing the internal resolution of digital audio systems. If a mathematical operation created an “overflow” result that required more than 16 bits, no problem: 24, 32, 64, and even 128-bit internal processing became commonplace. As long as the audio stayed within the system, running out of resolution wasn’t a problem.


Nowadays, your hard disk recorder most likely records and plays back at 24, 32, or 64 bits, and the rest of your gear (digital mixer, digital synth, etc.) probably has fairly high internal resolution as well. But currently, although there are some high-resolution audio formats, your mix usually ends up in the world’s most popular delivery medium: a 16-bit, 44.1kHz CD.

What happens to those “extra” bits? Before the advent of dithering, they were simply discarded (just imagine how those poor bits felt, especially after being called the “least significant bits” all their lives). This meant that, for example, decay tails below the 16-bit limit just stopped abruptly. Maybe you’ve heard a “buzzing” sort of sound at the end of a fade out or reverb tail; that’s the sound of extra bits being ruthlessly “downsized.”

Dithering to the Rescue

Dithering is a concept that, in its most basic form, adds noise to the very lower-level signals, thus using the data in those least significant bits to influence the sound of the more significant bits. It’s almost as if, even though the least significant bits are gone, their spirit lives on in the sound of the recording.

Cutting off bits is called truncation, and some proponents of dithering believe that dithering somehow sidesteps the truncation process. But that’s a misconception. Dithered or not, when a 24-bit signal ends up on a 16-bit CD, eight bits are truncated and never heard from again. Nonetheless, there’s a difference between flat-out truncation and truncation with dithering.

Now let’s take a closer look at dithering…

Dithering Rules

The First Law of dithering is don’t dither a signal more than once. Dithering should happen only when converting a high bit-rate source format to its final, 16-bit, mixed-for-CD format (and in the years to come, we’ll probably be dithering our 32 or 64-bit internal processing systems down to 24 bits for whatever high-resolution format finally takes off).

For example, if you are given an already dithered 16-bit file to edit on a high-resolution waveform editor, that 16-bit file already contains dithered data, and the higher-resolution editor should preserve it. When it’s time to mix the edited version back down to 16 bits, simply transfer over the existing file without dithering.

Another possible problem occurs if you give a mastering or duplication facility two dithered 16-bit files that are meant to be crossfaded. Crossfading the dithered sections could lead to artifacts; you’re better off crossfading the two, then dithering the combination.

Also, check any programs you use to see if dithering is enabled by default, or enabled accidentally and saved as a preference. In general, you want to leave dithering off, and enable it only as needed.

Or consider Cubase SX, which has an Apogee-designed UV22 plug-in. Suppose you add this to the final output, then suppose ou add another plug-in, like the Waves L1-Ultramaximizer+. This also includes dithering, which defaults to being enabled when inserted. So, check carefully to make sure you’re not “doubling up” on dithering, and disable dithering in one or the other.

Dithering dans Cubase

If you insert dithering in Cubase SX, it defaults to being enabled. So if you use this, make sure that any other master effects plug-ins you add do not have dithering enabled (in this screen shot, the WAVES dithering has been turned off). Or, disable Cubase’s dithering section and use the other plug-in’s dithering instead.

The best way to experience the benefits of dithering is to crank up some really low-level audio and compare different dithering and noise-shaping algorithms. If your music has any natural dynamics in it, proper dithering can indeed give a sweeter, smoother sound free of digital quantization distortion when you downsize to 16 bits.

To read the full detailed article see:  All About Dithering

February 2, 2011

Mastering: Curve Analysis and Acquision Software

Bob Ludwig, Doug Sax, Bernie Grundman – they’re masters of mastering. They produce hit after hit, with nothing at their disposal other than…well, experience, talent, great ears, the right gear, and superb acoustics.

So maybe you’re missing one or more of those elements, and wish that what came out of your studio sounded as good as what comes out of theirs. So, why not just analyze the spectral response curves of well-mastered recordings, and apply those responses to you own tunes?

Why not, indeed – but can you really steal someone’s distinctive spectral balance and get that magic sound?

The answer is no…and yes. No, because it’s highly unlikely that EQ decisions made for one piece of music are going to work with another. So even if you do steal the response, it’s not necessarily going to have the same effect. But the other answer is yes, because curve-stealing processors can really help you understand the way songs are mixed and mastered, and point the way toward improving the quality of your own tunes.

As to the tools that do this sort of thing, we’ll look at Steinberg’s FreeFilter (which was discontinued, but still appears in stores sometimes), Voxengo CurveEQ, and Har-Bal Harmonic Balancer. They’re very similar, yet also, very different.

How They Work

FreeFilter and Voxengo split the spectrum into multiple frequency bands in order to analyze a signal. These create a spectral response, as from a spectrum analyzer, while a song plays back. During playback, the program builds a curve that shows the average amount of energy at various frequencies. You can apply this analysis (reference) curve to a target file so that the target will have the same spectral response as the analyzed file, as well as edit and save the reference file.

Har-Bal isn’t curve-stealing software per se. While optionally observing the response of a reference signal, you can open another file, and see its curve superimposed upon the reference. You can edit the opened file’s curve so it matches the reference signal more closely, but this is a manual, not automatic, process.

Fig. 1: The black line is the spectral response for Madonna’s Ray of Light; the red line represents a Fatboy Slim mix. Fatboy’s has a lot more treble, while Ray of Light has a serious low-end peak.

The manual vs. automatic aspect is in some ways a workflow issue. FreeFilter and Voxengo start by creating the reference curve, but give you the tools to adjust this manually because you’ll probably want to make some changes. Har-Bal takes the reverse route: You start out manually, and if you want to, use the tools to create something that resembles the visual reference curve, which was generated automatically when you opened the file. Also remember that curve-stealing is only a part of these programs’ talents; they’re really sophisticated EQs.

So what do some typical curves look like? Check out Fig. 1. The black line is the spectral response for Madonna’s “Ray of Light,” while the red line represents a Fatboy Slim mix. Past about 1 kHz, Fatboy’s curve shows enough high frequency energy to shatter glass. “Ray of Light” has a higher response below about 400 Hz, due mostly to a prominent kick. It has a more thud-heavy, disco kind of vibe, whereas Fatboy Slim leans more toward a techno style of mastering. Apply these curves to your own music, and they’ll take on the characteristics of the reference tunes – but the results may not be what you expect, as we’ll see.

Now let’s take a look at the individual software…

So What Does Work?

Using your ears to compare your work to a well-mastered recording is a tried-and-true technique, but it shortens the learning process when you can actually compare curves visually and see what frequencies exhibit the greatest differences.

I’ve found a few reference comparison curves for Har-Bal that work well for certain types of music: Fatboy Slim for when dance mixes are too dull, “Ray of Light” for a house music-type low-end boost, Cirque de Soleil’s “Alegria” for rock music, and Gloria Estefan’s “Mi Tierra” for acoustic projects. On very rare occasions I use their curves, but when I do, they’re more like “presets” because they end up getting tweaked a lot. Automatic curve-stealing just doesn’t do it for me, but “save me 10 minutes by putting me in the ballpark” does.

But my main use for curve-analyzing software is for stealing from myself. After mastering a music project for a soundtrack, one tune sounded a little better than the others – everything fell together just right. So, as an experiment, I subtly applied its response to some of the other tunes. The entire collection ended up sounding more consistent, but the differences between tunes remained intact – just as I’d hoped.

Another good use was when German musician Dr. Walker remixed one of my tunes for a compilation CD, but used a loop for which he couldn’t get legal clearance. Rather than give up, I created a similar loop that wasn’t a copy, but had a similar “vibe.” Yet it didn’t really do the job – until I applied the illegal loop’s response curve to my copy. Bingo! The timbral match was actually more important than the particular notes I played in terms of making the loop work with the rest of the tune.

To read the full detailed article please see:  Curves of Steal

This does produce a weird paradox, though: I used a piece of curve-stealing software to avoid stealing a piece of copyrighted material. I guess it’s all part of the living in the 21st century.

Blog at