Dithering Explained: What it is, When to Use It, and Why it’s Important
Although dithering is an important step of the mastering process, knowing how to use it, when to use it, and why you should use it will improve your recordings and final mastered audio files.
What Is Dither?
Dither is low volume noise, introduced into digital audio when converting from a higher bit-resolution to a lower bit-resolution.
The process of reducing bit-resolution causes quantization errors, also known as truncation distortion, which if not prevented, can sound very unpleasant.
To understand this better, we must understand bit-depth.
What Is Bit Depth?
In the digital audio domain, bit-depth is what defines the number of measurement values to describe the amplitude of a single audio sample. Each bit effectively represents 6db of dynamic range.
For example, a recording made at 24bit-resolution would have a potential range of 144db. (See figures 1-2 below).
As one reduces the bit-depth, such as from a 24bit-resolution sample to a 16bit-resolution sample, you are reducing the number of values available to measure the amplitude of any given sample.
In other words you now have less values available to describe the dynamic range of your audio. (see figures 3-4 below).
As a result, certain values that are no longer there will be forcibly rounded off to the next closest value.
This truncation results in the loss of very low signal levels, and the creation of audible distortion where the values have been rounded, squaring off the waveform. (see figures 5-6 below).
How Does Dither Prevent This?
If you apply dither to a silent audio file, and turn the volume way up, you can hear the sound of dither alone.
You might even see this visually if you add an EQ to the end of your production chain and see the noise moving around even though there is no audio coming from the speakers.
Introducing this subtle noise to an audio file prior to reducing the bit-depth eliminates the truncation distortion. You are in effect trading the distortion for noise.
Given this information, one could determine that the use of dithering on a 24bit sample, and exporting in 24bit or higher resolution bit-depth would be ineffective, as there is nothing being replaced with noise, and is only necessary when down-converting to a lower bit-depth.
The Computer Screen Analogy
To help better understand dithering, I like to use the hand over your computer monitor analogy. How it works is you start by holding your hand over your computer monitor.
Notice that you can see your computer monitor perfectly with the exception of the block where your hand is.
Now, if you wave your hand rapidly back and forth from left to right across the screen (applying dither), it allows you to see the entire screen as apposed to blocks of the screen.
Now that you have a better understanding of what dithering is, you might be asking yourself, “why dither?” Especially if you can just keep your 24bit-resolution file and avoid dithering altogether.
The answer is simple; all finished, mastered audio files are 16bit. Although 24bit is a higher quality sound with more audio detail, and eliminates truncation distortion altogether, the reality is that 90% of all playback devices are 44100/16bit.
Which means if you try and play a 24bit audio file through one of these 16bit playback devices, it will sound like shit.
In this regard, you should keep the consistency of bit-depth throughout your production process from beginning to end. If you are producing in 24bit and your playback is set to 16bit, then you should be using a dithering tool in your production chain.
If you are recording and producing in 16bit, and your playback is in 16bit, then there is no need to dither. If you are producing in 16bit, and your playback settings are 24bit, there is no need for dithering.
What are the current settings of your project? What are the current bit-depth of your samples? What is your playback bit-depth settings set at? These are all things you should know when producing your track.
Note: If you intend to have your song mastered, it is best to export at the same bit-depth or higher as your project settings are set to. For example, if you are producing in 16bit, be sure to export in 16bit or higher.
If your project settings are in 24bit, and you export in 16bit without dithering, your audio file is damaged before it even goes to the mastering engineer.
Types of Dither Algorithms and Shaping Options
Many dithering options offer noise shaping. Noise shaping allows you to add an eq curve to the dither noise, helping move the energy of the noise to less audible regions within the frequency spectrum for an even better result.
Here are a few popular types. (I will be using Ableton’s dithering options, though is very similar options in all programs).
- Triangular – By default, Triangular is selected, which is the safest mode to use if there is any possibility of doing additional processing on your file.
- Rectangular – Rectangular mode introduces an even smaller amount of dither noise, but at the expense of additional quantization error.
- The Three Pow-r Modes – The three Pow-r modes offer successively higher amounts of dithering, but with the noise pushed above the audible range.
The Images Analogy
Image dithering works the exact same way and is no different than audio dithering. Below are four images.
From left to right, the first image is an 8 bit image at full resolution, next is the same image reduced to 1 bit with no dithering, 3rd is the same greatly reduced image with dithering added, and lastly in image 4 is the reduced bit image with added dithering, plus noise shaping option added. (see figure 8 below).
Last Notes and Conclusion
Note that dithering is a procedure that should only be applied once to any given audio sample. If you plan to do further processing on your rendered audio sample, it’s best to render to 32-bit to avoid the need for dithering at this stage.
Lastly, you only want to dither your rendered audio if it’s final. If you’re sending it to someone else for mastering, or it’s just not yet the master, then don’t dither.
Regarding which mode is best, it’s really best to use your ears and spend some time with the results.
Images provided by izotopeinc – http://www.youtube.com/user/izotopeinc?feature=watch
*Helpful Video Demonstrations:
Thanks mate for this really helpful post. IT IS the first time I REALLY understand what dithering is! I’ve read so much, was watching so much vids, but yours is the best so far. Especially I thought is was funny with the comparison with the waving hand in front of your screen ;o)) keep on
Hi Chief, really glad you got some use out of the article. Cheers!
Excellent explanation of the dithering process. Very straightforward and accurate.
Oddly, some recommend dithering even when going from 16-bit to 16-bit. Haven’t figured out why one would do that.
Thanks for this – I’ll pass it along to clients who are curious.
Hi Don, thanks for commenting. I’m glad to hear you got some use out of the article. The reason why someone would dither from 16bit to 16bit might be just an extra precaution but I don’t think it would really have any effect on the audio signal.
UPDATE: When dithering from 16bit to 16bit audio files, in my experience, can actually harm the audio file more. That is because if the 16bit file you are dithering from has already been dithered, you are essentially just adding extra noise on top which can be more audible than you might expect.
If you are dithering from a 16bit file that was never dithered from a larger bit-rate audio file in the first place, then that file is already damaged. Adding some noise to it “might” cover-up the poor track, but really you should only dither with lowering the bit depth.
Thanks for this great article. I’ve been using Dithering For Years And could always hear a difference but now I actually understand what the difference is!
Eddie, I’m glad you enjoyed the article. Cheers!
I didn’t know anything about dithering in till I read this- I have seen the word but couldn’t figure what it did lol Thanks 🙂
Glad you liked the article Calvin. Cheers.
Super helpful explanation, thanks!
very helpful.you’re great.
Thanks chef George!
brill really helped me understand and important information big up
Re: 16 Bit to 16 Bit Dithering…Many people do it because the internal processing of their chosen software is possibly at 24, 32 or even 64 bit thus dithering is still required at the output stage.
What if you take a 24-96 file and resample to 24-48. Is dithering not needed?
A second question. Suppose I get an lp a friend transferred at 32-192 and he did not dither. the lp is now at 24-96. Will dithering improve quality at all?
Dithering is just adding subtle noise to the signal when truncating bits, so dithering only makes sense when going to 16 bits. The program’s internal processing algorithm will very often take care of the down-converting from a higher sample rate.
You should always use dither when down-converting bit-rate, and it should only be done one time at the last stage. If 24bit is your final goal, then you should dither from 32bit to 24bit. However, humans have a hard time distinguishing 24bit from higher bit-rates. Mathematically, 32bit is higher quality, but 24bit carries all the data you will hear, and you won’t be able to hear a difference. so all is not lost if you did not dither from 32bit to 24bit.
You should ALWAYS dither from a higher bit-rate to 16bit, however.
I’ve read your very insightful blog, Thank you!
I’m bouncing a finished track from Reason at 16 Bit without Dither enabled. some of the Samples (kick drum, snare etc.) are 24 Bit files, would you recommend I bounce the overall project to a 24 Bit with Dither or leave it at 16 Bit without Dither On?
I would like for the track to be compatible and played correctly with any speaker system.
I’ve read your very insightful blog, Thank you!
I’m bouncing a finished track from Reason at 16 Bit without Dither enabled. some of the Samples (kick drum, snare etc.) are 24 Bit files, would you recommend I bounce the overall project to 24 Bit with Dither or leave it at 16 Bit without Dither On?
I would like for the track to be compatible and played correctly with any speaker system.
Svengali, I would recommend exporting the track out at 24bit without dithering if you plan on getting it mastered. But if you plan to release the track as is to share and distribute I would recommend exporting to 16bit with dither on to be compatible with most players. Dithering should only be done as the final step before distribution and it should only be done once!
Thank you very much sir! I’ll probably be back 🙂
BTW it doesn’t let me Enable Dither when Exporting to 24 Bit :/ only allowed on 16 Bit?
Usually you only want to dither when down-converting from a higher bit-rate to 16bit.
So, it would make sense that dither option is only offered at 16bit.
It’s best to just export at the highest bit-rate possible with no dither added if planning to send off for mastering.
Hi Svengali, in the case that you do not want it mastered and intend to distribute your track after exporting from Reason, I recommend exporting to 16bit with dither enabled.
If you plan to have it mastered I recommend exporting to 24bit with dither disabled.
Its really helpful link to clear up the fundamentals about dithering.Thanks again
Cheers Amaresh. Thank you.
Thank you very much for the great article. Well explained compared to other articles on the internet. Keep the good work up!
I know what dithering is, and why it’s important. But I only have one question that I can’t find any answer to, despite how simple the question is. If I could get a simple answer, that would be GREAT. I am about to master an album. My files are 24 bit/48k and will bounce to 16/44.1. Should I set my dither rate at 16 bit, or 24 bit? I don’t really need an explanation, just an answer. Thanks.
Hi, really great article that clear some things up. I have a question about mixing and mastering. I’m using fl-studio with a sample rate of 44100hz and export my mixdown as a 24-bit wave-file. then I open this in a new project to do the mastering. After the mastering is done I export it to 16-bit wave-file. At this stage, should I enable dithering at the end (in my DAW) or should I, for example, use the dithering function in the L2 limiter from waves? Also, should i use dithering when rendering from a 24-bit wave-file to a 320kbps mp3 file? Thank you in advance.
Hi Mike, good question. Dithering should only be applied once at the final step before exporting the mastering. Therefore, any ONE of the options you mentioned is good. I prefer to use the VST for the dithering, then export from the DAW without dither at 16bit. If the VST has too many options for dither as I know some do, just use the safe method of dither and do it from the DAW using the Triangular dithering option.
Also, for MP3 don’t worry about the dither. After export of the final 16bit master WAV file just make sure to keep the import/export settings as consistent with the WAV file. I.E. Import your MP3 using 16bit, Export in 16bit. Should be good.
Thank you for your answer. So if I understand correctly: I make my track and do the mixdown in the same project and export this at a 24-bit wave file without using dither. After this I import the finished mixdown in a new project to do the mastering and when exporting to 16-bit I use dither (within my DAW or in a VST-plugin). When exporting for example to 24 bit, I don’t use dither right?, is this correct? Thanks in advance.
You got it!
Same goes for 32bit-24 bit. For example if you are recording in 24bit, but the final resolution is going to be 24bit, then dither is needed. Always when you are down-converting bit rate only. I.E. 24bit to 24bit no dither needed.
Do I need to dither going from 16 bit to 24 bit. Recordings were made at 16 44.1 and I’ve been advised to convert to 24 bit for mastering (DIY)?
It is impossible to up-convert bit rates. Therefore, it is not necessary to dither to 24bit from 16bit. You will essentially just have a 24bit recording of a 16bit recording, meaning that it is unnecessary. You should only dither when down-converting bit rates.
Thank you for your article on dithering explained.
I have one question which has me baffled about quantisation errors. For example, truncating 24-bit to 16 bit without dither. Are quantization errors only audible in the lower limit of the 16-bit dynamic range? Will it be audible in music which has, say, only a 60dB dynamic range?
The main reason I ask, I have a 16 bit R2R DAC which accepts and processes 24-bit via USB all the way up until the DACs R2R stage, only the chips themselves are 16 bit. In this case, will quantization errors be audible with 24-bit files with a dynamic range of 60dB?
Thank you for your help.
Excellent artical on dither, I now have a good understanding of it so I just want to make sure I understand. I record my project in 44100 24 bit, bounce down to 24 bit wav, I then bring that file into T-Racks for mastering then save it as 44100 16bit with dithering triangle applied?
That is correct. I like to use a limiter that has a dithering option so I can measure the peak level before exporting, rather than use the dither option that comes with the DAW during export.