What is Dithering? Short and Clear Explanation of it!
What is Dithering
Do you still have issues understanding what is dithering? Dithering or to dither means including noise to the audio signal. Noise is being added on goal to trade a slightly bit of low-level hiss for an excessive amount of distortion.
The distortion is first triggered through the use of a set variety of bits (e.g. 16-bit audio) to symbolize our pattern factors as accurate as possible (within the analog world, there are infinite or steady pattern factors obtainable, whereas they’re finite (discrete) within the digital world).
So, in our DAW, for instance, dithering is done by including the noise of a level lower than the least significant bit earlier than rounding to 16 bits. The added noise has the impact of spreading the numerous short-term errors throughout the audio spectrum as broadband noise.
When using Dithering
You’ll solely want to dither when lowering the variety of bits used to signify a signal. Due to this fact, whenever you cut back a 16-bit audio file to 8 bits, you’ll want dithering. It’s going to create harmonic distortion to exchange truncating or rounding to suit samples into the lowered bit size (also known as lowered “word size”) to unfold out the error over time, creating broadband noise.
Dither will not be needed when using resolutions that are high enough. For instance, using 32-bit floating level, you possibly can rest assured that your dynamic range is so large, that the least significant little bit of your audio is virtually so quiet, it is not perceptible anymore (until your listening to is at the atomic level).
DAWs like Ableton Live and digital audio signal processors often work at this bit resolution. So it is doable and really helpful to use dithering solely when it’s important to deliver your audio file as 16-bit values or lower.
When Not to Use Dithering
If you happen to use dithering when it’s not wanted, it isn’t the worst factor you are able to do to your signal, however this must be prevented.
All you’re doing when using unneeded dithering is including additional and unwarranted noise to your signal. Do that enough times in your signal chain, and the noise will turn out to be loud enough to be perceived by your listeners, which definitely isn’t thing.
If you happen to use dither enough, you’ll enhance your noise floor to a noticeable extent.
Considering most engineers take an excessive amount of effort to keep away from extreme noise of their recordings, it isn’t smart so as to add further noise for no good purpose. That’s why dithering must be used only when wanted.
Once more, in the event you aren’t lowering the bit depth of your recording and creating quantization, distortion or noise because of this, then you definitely shouldn’t use dithering. Dithering ought to only be used to cover quantization distortion. All this is important to understand what is dithering and when to use it.
Are you after a sound engineering job? Here is a HUGE list of opportunities: Sound Engineering
The Different Types of Dithering
You could have noticed that in relation to dithering, you’ve got completely different choices. In Logic Pro you’re offered with the POW-r #1,#2, and #3 choices, whereas different plugins supply noise shaping choices that vary from ‘Moderate’ to ‘Ultra.’
Logic Pro provides a number of dithering types while exporting. Those are very helpful when you understand what is dithering.
Actually, the variations between these completely different settings are minimal, and the impact they’ve on making your mix or grasp sound higher is debatable. With that stated, it’s nonetheless good to have some data on what these phrases really imply, so let’s briefly take a look at these dithering types.
POW-r or Psychoacoustically Optimized Wordlength Reduction
Type 1: Type 1 POW-r dithering is often used for low-dynamic range mixes, reminiscent of extremely compressed pop or rock. No signal shaping is used, which means the noise that’s introduced has not been affected with an EQ curve. In different words, that is noise with a flat frequency response.
POW-r 1 dither is used for music with out a vital dynamic range.
Type 2: Type 2 POW-r dithering is often used for speech. It consists of noise-shaping/equalization that attenuates the noise on the 2kHz space and amplifies it within the 14kHz are and upward. This shaping is used to make sure the noise added continues to be efficient in masking quantization distortion, however not itself perceivable.
POW-r 2 dithering is finest utilized for speech.
Type 3: Type 3 POW-r dithering is used for extremely dynamic recordings, like orchestral music. Like type 2, noise shaping is used, nonetheless, the extent to which certain frequencies are attenuated and amplified is way better. Primarily the EQ curve applied to the noise is larger.
POW-r 3 dither is claimed to be finest for extremely dynamic music.
The Terms ‘None,’ ‘Moderate’ and ‘Ultra’ below the shaping tab of most industrial limiter plugins refers back to the quantity of noise shaping that happens. Because the names may suggest average moderately shapes the dither, whereas ultra extra significantly alters the signal.
Usually, this shaping attenuates the noise within the 2kHz are and amplifies the upper frequencies of the noise. This happens in a similar way to the way it was detailed above, when discussing type 2 and type 3 POW-r dithering.
Rounding and truncating
Think about we have to change 24-bit information into 16-bit ones. To make them match, all the knowledge from the unique 24-bit audio information has to enter the 16-bit spaces.
This means one thing has to get left behind. However, how will you do the least damage?
The primary choice is to easily reduce the numbers off to make them match. That is known as truncation and it’s the least correct approach of fixing the issue.
As an alternative of simply reducing the numbers off, we are able to try rounding them up or right down to get nearer to representing the 24-bit audio in 16-bits.
However what occurs when the quantity you need to round is true in the middle? You may attempt a rule of thumb—say you resolve to round up each time you encounter a number proper in the center.
This way at the very least you’ll be right some of the time. The trade-off is that you already know you’ll be mistaken simply as often! That is known as a rounding error. This is a common error when you are not full aware what is dithering.
The noise (or random variation) that dither provides basically randomizes the choice to round up or down.
After dithering, the distortion that might be created by truncation or rounding has a lot less detectable impact.
Eliminate quantization distortion
Bit depth defines the variety of values out there to symbolize the amplitude of the samples which make up a digital audio file. While you convert an audio file to a decrease bit depth, you cut back the variety of values that are out there to symbolize the amplitude. As such, the bit value of some samples will likely be ‘quantized’.
Which means that they’re rounded to the closest accessible value. These rounding errors trigger one thing known as quantization distortion. Quantization distortion can introduce harmonics and different distortion varieties that are correlated to the unique signal.
Dither noise is generated by switching the last bit within the data on and off at random. This creates a clean, low-level white noise/hiss. Switching this bit on and off at random randomizes the quantized error. As such, it decorrelates the errors from the unique signal, in turn eliminating distortion.
At the kind of bit depths that we’re usually working with, i.e. 16-bit and 24-bit audio, this distortion is extraordinarily quiet. It turns into far more obvious at decrease bit depths. However, even at 16-bit audio resolution, it’s possible to listen to quantization error.
This often occurs at instances when the music that would otherwise mask it is extremely quiet.
For example, you would possibly hear quantization distortion throughout a fade-out or on the end of a reverb tail. Actually, even in a case, you don’t hear the distortion itself, its presence could make your audio sound harsh. As such, it’s vital to make use of dithering to remove quantization error distortion.
All this business with rounding and error could seem fully educational, but right dithering could make a distinction in your final product.
On the very least, including dither on the right level in your project is the closest factor you will get to bettering your mix with one click on, even when it’s only a small enhancement.
As an engineer, it’s best to attempt for the absolute best sound you possibly can—and never miss any alternative to make it better.
Now that you already know your approach around dither, get back to your DAW and bounce some tracks.
Sample Rate and Bit Depth
When audio is digitized by an analog to digital converter (ADC)—say, whenever you record into your interface—and it will get sent to your DAW, the analog audio signal out of your microphone or DI is remodeled into a series of ones and zeros that digitally symbolize your audio.
To perform that, the ADC captures snapshots (aka “samples”) of your audio at a specified rate and size, which are known as the sampling rate and the bit depth.
The sampling rate governs what number of samples your converter captures in a second to digitize the audio, and is vital by way of the frequency response.
The final rule, primarily based on the Nyquist Theorem, is that you need to have a sampling rate that’s better than or equal to twice the very best frequency of the audio. In any other case, the distortion is known as “aliasing” can mar the sound.
To capture frequencies as much as 20kHz, which is the high range of human hearing, the bottom sampling rate generally used is 44.1kHz (the speed in the CD spec), however, as of late, rates of 48kHz or greater are also frequent. The higher the sampling rate, the better the precision for reproducing the frequency range.
How do I add dither?
There are a couple of alternative ways to use dither. Most mastering plugins provide the potential to use dither. The bounce dialogue in some DAW’s additionally provides you the option to add dither. Some limiter plugins have a dither possibility. Or alternatively, you should use a devoted plugin that exists solely for dithering.
There are a few guidelines to follow whenever you apply dither to your session. Firstly, in case you’re including dither through a plugin, be sure to add it as the final plugin in your master fader.
There needs to be no different processing after it. Additionally, you only want to use dither once to every session. So, add dither as the final plugin, once per session, whenever you bounce to a set level bit depth.
Lots of the choices that help you apply dither will provide you with the choice to use ‘noise shaping’. Noise shaping focuses the frequency content material of the dither into areas of the frequency spectrum that are tougher for us to listen to. You should utilize this as a method of creating dither noise extra transparent.
When to Dither in Audio Mastering Process
Basically, it’s essential to use dithering whenever you’re mastering the file at a lower bit rate than you worked at through the production stage. Say you recorded in 24-bit and you’re mastering the audio file in 16-bit; it’s essential to use dithering.
However, in case you had been recording and producing in 16-bit and also you bounce it at 16-bit audio, you don’t have to dither. Anyway, you will need to check these options before you send the track for the mastering process.
In case you’re sending a track to a mastering engineer, it is best to by no means export it at a lower bitrate than the one you had been working at, as it will imply the audio is already damaged before the engineer even begins working on it.
How Audible Is Dither?
With our audio interface’s output cranked all the best way up, at a bit depth of 16, dither amount set to “Strong”, and noise shaping set to “Off”, We are simply able to hear the noise that dither introduces. That is primarily a worst-case scenario because it is fairly unusual to bounce music lower bit depth from 16.
Dithering Mixing Tips
- All the time dither in case you’re creating 16-bit files for a CD from a 24- or 32-bit mix.
- You mustn’t dither when going from 32-bit floating-point to 24-bit (because the 32-bit floating-point doesn’t have a higher bit depth), however, you do from 32-bit fixed point to any decrease bit depth.
- In case you’re making use of 24-bit or 32-bit information for a mastering engineer, don’t dither. Let your mastering engineer, who’s an expert, deal with it.
- Be sure that dithering is the final step in your processing chain. By no means insert a processor after it in your DAW or audio editor.
- Attempt to avoid dithering twice, because the impact is to add extra noise to your audio. That said, in case you go down in bit depth twice throughout a project (which is considered unlikely except you begin at a 32-bit fixed-point, it is best to dither throughout each of those conversions.
- In case you’re confused in regards to the numerous dither choices offered, go with the essential default provided by your dithering software program. In case you’re feeling extra ambitious, examine the outcomes with totally different dither and noise-shaping choices selected to see which works best for a selected track.
At first, dither can appear fairly confusing. Not least of all because it includes deliberately adding noise to our audio, which is often one thing that we’d look to keep away from. Our suggestion is to be sure what you do before you do it, that’s why we wrote this text.
Hopefully, this article has given you an understanding as to why including dither noise is definitely very useful. While you are reducing the bit depth of an audio file, you introduce quantization distortion.
You additionally lose low-level details. By the usage of dither, you can’t only remove the distortion, but additionally, protect the low-level details too. As such, an audio file with dither utilized ought to sound identical at a lower bit depth because it did at a higher bit depth, however with the addition of very low-level dither noise.
Are you positive to at all times apply dither whenever you cut back the bit depth of your audio files? At what factors do you usually want to scale back the bit depth of your audio files whenever you produce music? Leave your ideas within the comment field beneath.