Audio dithering during export and mastering
What is audio dithering?
Audio ditheringis a very quiet noise signal that occurs when reducing theBit DepthIt makes quantization errors less noticeable. It is primarily used when a 24-bit project is exported as a 16-bit file at the end of mastering, for example for CD masters or certain distribution formats.
Without dithering, rounding errors in very quiet signals can become audible as harsh distortion or grainy noise. Dithering distributes these errors more favorably from a psychoacoustic perspective, so that the transition to the lower bit depth sounds more natural.
It's important to note that dithering isn't a sound enhancer for every export. If the bit depth remains the same, further dithering is usually unnecessary. It's primarily useful at the end of the signal chain when the final master is intentionally converted to a lower bit depth.
What is quantization noise?
Quantization NoiseQuantization noise is noise that arises when converting an analog signal to a digital signal. When an analog signal is converted to a digital signal, it is divided into a series of discrete values called quantization levels. Because the analog signal is continuous, it cannot be perfectly divided into these discrete quantization levels. This introduces an error known as quantization error, which can lead to signal distortion.
Quantization noise occurs when the quantization error is combined with random noise. Noise can be minimized through dithering by adding artificial noise to the signal. This can help spread the quantization error over a wider frequency distribution, thereby reducing quantization noise.
Quantization noise can be particularly audible at low resolutions and bit depths. However, at higher resolutions and bit depths, quantization noise is typically less audible and can be further reduced by using noise reduction algorithms.
Audio dithering from the pros! We make all your signals audible
What are quantization errors?
Quantization errors are errors that occur when digitalrecording sessionor during the processing of analog signals. They arise when an analog signal is converted into discrete values (sample values) in order to store or process it digitally.
Quantization involves breaking down the analog signal into discrete values that are stored in a digital format. The resolution of the digital format, i.e. the number of discrete values available, determines the accuracy with which the analog signal can be reproduced. If the analog signal is between two discrete values, it is rounded to the nearest discrete value. This rounding process results in an error known asquantization errors or rounding errorsreferred to as.
The size of the quantization error depends on the resolution of the digital format and the rounding accuracy. The higher the resolution of the format, the smaller the quantization error.
Why is high resolution good for music production?
Higher resolutions in digital audio typically refer to a higher sample rate and/or a higher bit depth.
TheSample RateIndicates how often per second an analog signal is sampled to generate a digital signal.
Dithering smooths out the bit depth reduction at the end of mastering — where this fits into the overall workflow is determined by theMastering Guide.
A higher sampling rate means that the analog signal is sampled more frequently, resulting in greater accuracy. Typical sample rates for audio CDs are 44,1 kHz, while high-resolution audio formats such as FLAC or MQA support higher sample rates of up to 192 kHz or even higher.
Bit depth refers to the number of bits used to represent a sample. A higher bit depth means that more bits are available to represent the signal level. With a higher bit depth, the number of stages is larger, meaning there are more discrete values that the signal can take on. A higher bit depth therefore means that theDynamic Rangeof the signal can be larger, allowing for greater accuracy and clarity of the audio signal.
Overall, higher sample rates and larger bit depths typically result in better sound quality and a more natural sound image, especially for music that contains a lot of detail and nuance. However, higher resolutions also require more storage capacity and higher audio signal processing requirements, both during recording and playback.
Analog sound for your tracks?
Which dithering should be used when?
Choosing the right dithering method depends on the target medium, the bit depth, and the export format. In practice, the crucial factor is whether the bit depth is actually reduced. A typical case is exporting from 24 bits to 16 bits; for 24-bit or 32-bit float exports, additional dithering is usually not required.
Some common types of dithering are:
- Rectangular dithering:A simple form with consistent noise. It can work at low resolutions, but is rarely the first choice for high-quality mastering.
- Triangular dithering:A common, robust variant with less distortion than rectangular dithering. It is well-suited for many standard exports.
- Noise-Shaping Dithering:It distributes the dither noise more evenly across frequency ranges where it is less noticeable. This can be useful with 16-bit masters, but it must be appropriate for the source material.
- Dynamic dithering:It better matches the noise to the signal and can be useful when the signal strength varies.
In a professionalAudio MasteringDithering will only be decided upon at the final export stage.DDP masteringFor CD playbacks or certain 16-bit deliverables, it is often relevant; for streaming masters with higher bit depth, it does not necessarily have to be applied automatically.
Frequently Asked Questions about Audio Dithering
When should you dither – and when not?
Dithering is only necessary if you reduce the bit depth – typically when exporting from an internal 32-bit float or 24-bit file to the 16 bits of a CD. If you stay at 24-bit or deliver a 32-bit float file, dithering is usually unnecessary because there is hardly any audible quantization. Rule of thumb: only dither to a lower bit depth in the final step.
Should you dither multiple times?
No. Dither should only be applied once at the end of the signal chain, during the final reduction of the bit depth. Any further dithering only adds extra noise without providing any benefit – so don't activate it in every plugin and again during export.
What is noise shaping in dithering?
Noise shaping shifts the already very quiet dither noise into frequency ranges where the human ear is less sensitive (usually the high frequencies). This makes the noise seem even less noticeable. It's an option in many dither algorithms and is particularly useful when moving to 16-bit resolution.
Can you hear dithering in the finished track?
At normal listening levels, dither is practically imperceptible – it lies as a very quiet, consistent hiss far below the music signal. Conversely, it prevents the significantly more disruptive quantization distortion, which would be particularly audible in quiet decays. When used correctly, dither thus avoids audible artifacts and keeps quiet decays cleaner.