Despite CD becoming a niche format, most will be aware of the need to dither. With the all-conquering convenience of streaming, where does that leave dither and do we still need it? We investigate.

In the analogue era, considerations around dither didn’t exist. However, in the forty years or so since CDs have been on the market, it is something that has entered into the vocabulary of both manufacturer and engineer alike. In the era where CD is now a minority format, is dither still needed? When it comes to dither, some engineers know what it is, and will use it appropriately, while others might use it when needed without necessarily knowing what it’s doing. While this guide cannot cover the finer points of digital audio, if you want to know more about why you need dither read on…

Why Use Dither?

For the uninitiated, dither exists to solve a distinctly digital problem. That is, how to resolve very low level sounds that happen beneath the lower limit of a system’s working range. In a 24 bit system, things are easier, as its ‘footroom’ extends far below the quietest things that are happening before the converter. In a 16 bit system, however, very low level sounds will not be encoded, causing audible artefacts on the audio. This is especially noticeable on things like reverb tails at the point where they ‘go under’ the lowest encodable level. Dither fixes this problem.

How Does Dither Do It?

If you think that all digital audio has a perfectly ‘black’ noise floor look away now. Dither is noise (think of it as hiss) that is intentionally added to 16 bit audio around the lowest encodable audio level. The system is able to encode the noise, and while doing so lets through the quieter low level audio that lives within the noise as well.

This is possible because in a digital system, the first bit of data (representing the lowest encodable level) can have its state rapidly toggled on or off by the noise. Cleverly, if this bit can resolve the noise, so too can it resolve the wanted audio within. Put another way, the wanted audio offsets (or ‘biasses’) the noise, making the former audible as opposed to being truncated.

Hear for yourself how low level sounds are affected by 16 bit audio in Eli Krantzberg’s video below. He then adds different types of dither to restore the audibility of the wanted signal, even when it is below the lowest encodable level!

Dither For Streaming

Whether the underlying details of dithering are understood or not, most engineers understand that audio for CD production must be dithered owing to its 16 bit format. Some streaming services also use 16 bit codecs. The question of how these services transcode masters is another another subject but the takeaway is that where required, 16 bit files still need to be dithered. This is because when a 24 bit encode is used at the service’s end, this will render any low level 16 bit truncation as well. Many services support 16 bit files at the end user (such as Apple’s ALAC, or TIDAL HiFi) but this won’t tell the engineer what dithering, if any, is employed by that service. If in doubt, check before upload for the current requirements.

Studio One 5.5 And Dither

Users of PreSonus Studio One have recently received expanded dithering features in the 5.5 update. Its new dithering algorithm ensures that transcoded audio won’t suffer from audible noise, artefacts, and distortion that can be caused by lesser algorithms. This is currently a feature exclusive to Studio One.

Use Dithering for Playback and Audio File Export is engaged by default, and means that PreSonus’ triangular dithering is applied when the audio signal’s bit depth is reduced from a higher bit depth by a device or during file export. Users can of course still use a third-party dithering solution of their choice, such as a limiter Insert effect on the Main Output. As always, dithering must be used post-fade at the end of the signal chain regardless of DAW.

Describing their “exclusive cutting-edge dithering algorithm” PreSonus add:

Studio One 5.5 features a new, proprietary, custom-designed dithering algorithm that minimizes conversion artifacts and preserves low-level signals. It sounds so good that we’ve replaced Studio One’s previous dithering algorithm with it, so if you’ve got Dithering turned on (in the Advanced Preferences/Audio menu) you’re already using it!

Quality Counts

While the usual discussions in digital audio continue to revolve around things like sample rates, or loudness, the subject of dithering seems to be less on people’s radar. Perhaps by way of its prosaic nature, to overlook it is to forget another piece of the puzzle regarding streaming quality. OK, so no-one is pretending that dither is a subject to get excited about, but proper attention can be paid to getting it right simply by using tools that take care of it for the engineer. Whether or not to pay extra to stream uncompressed 24 bit music over a lossy bluetooth connection will, however, always be one for the listener to decide…