- Learn the difference between upsampling and downsampling
- Find out when upsampling and downsampling are used
- Are upsampling and downsampling necessary?
- Also, check out our post on the best free sample packs for electronic music!
You might have heard the words ‘upsampling’ and ‘downsampling’ mentioned on audio-related forums or articles but never fully understood what they are or what they do.
Aside from being a handy concept to understand and add to your audio production skillset, the ability to upsample and downsample can also prove an incredibly creative sound design tool for any open-minded music producer.
It’s time to dive into upsampling and downsampling, figure out the differences between the two, and benefit from the best of both!
What Are Upsampling and Downsampling?
Upsampling
When we use our voice or a musical instrument to record into a DAW – via microphone or direct injection, for example – our chosen sound source generates analog waveforms, which get picked up by our chosen recording method.
Before the waveforms hit our DAW, the signal containing their information is neatly converted from analog to digital.
The digital signal is just a computer-processed replica of the analog signal, built out of tiny snapshots at calculated points on the original waveform. These chunks are called samples – also sometimes referred to as bits.
If we could somehow allow the digital signal to carry a higher number of ‘snapshot’ samples based on the original analog signal, then the configured resolution of those snapshots – termed the sample rate – would increase, and the overall sound quality of the audio would appear to have improved.
So, when we fiddle with the preferences in our DAW (or use signal processing plug-ins) to increase the sample rate of our captured audio, we allow for the capacity to increase.
This means our audio can hold more of those tiny digital samples – or snapshots – of the original analog signal:
Related: What Sample Rate Should You Use?
Downsampling
Likewise, if we were to decrease the number of digital samples that our audio can hold, the sample rate will be reduced, and the resolution would appear worse or at least less defined.
When applied skillfully in practice, the ear can identify downsampled audio as a kind of ‘crushed’ sound.
For most audio recordings, downsampling is an unwanted result of low-resolution digital signal processing.
However, most audio streaming platforms will intentionally downsample uploaded audio for easier and quicker online streaming. Some more than others, though.
This is unlike upsampling, which is generally considered a useful tool for clarity and precision – especially good for maintaining time synchronization and quality control when the purpose of the audio is to be attached to high-resolution film clips.
Use Cases
Upsampling
Upsampling is usually used in two main ways:
Firstly, the audio project may belong to the production of a high-resolution film (often a Hollywood film), and the sample rate of the audio will need to exceed a certain threshold to be synchronizable and digitally compatible with the video.
Secondly, the audio project may contain artifacts of a low-resolution digital bounce (audio export), which can be fixed by increasing the sample rate inside your DAW.
Provided your CPU processing power is in good nick, your Logic settings aren’t overloading your computer, and your RAM isn’t capped out, the increase in sample rate hopefully won’t unwantedly affect the pitch of your audio project during playback or exporting.
Annoyingly, that happens when your DAW’s sample processing power becomes outmatched by the fidelity of the sample rate you have selected – like a record spinning faster than a vinyl player’s needle can read. Not good!
Downsampling
In terms of using downsampling in ways typically regarded as pleasant rather than sheer quality loss… there are two main ways to plan your approach!
Decreasing Sample Rate in an Audio Track or Project
Slightly decreasing the sample rate of an audio track or project will usually result in a 90’s Hip-Hop or Lo-Fi kind of sound since the decreased audio fidelity will cause the frequencies to appear ever so slightly lower and a bit ‘hazy’ – almost like the engine running your DAW is running out of juice.
An example of the ‘hazy’ result of a decreased sample rate can be found in most of J Dilla’s work, for example, The Pharcyde’s ‘Runnin’ (which he wrote and produced the instrumental for).
This effect can be very desirable to use in specific genres of music.
Altering the ‘Downsampling’ Function in a Bitcrusher
Another fantastic, easy-to-learn method of downsampling can be actively used for your creative benefit.
This uses the specific Sample Reduction function (sometimes labeled Downsampling) included in the Bitcrusher plug-in: a specially-designed plug-in featured as part of most modern-day DAWs.
Bitcrushers are designed for downsampling individual audio tracks in controlled but often ferocious ways!
By dividing the number of samples held by the audio signal by any amount on a Bitcrusher – usually ranging from x1 to x32 on most plug-ins running on a 32-bit DAW instance – a range of audible ‘bitty,’ ‘robotic’ sound qualities can be heard over the affected audio track.
If you have access to a DAW like
Listen to the blisteringly fuzzy guitars in Queens Of The Stone Age’s ‘Skin On Skin’ and the almost alien-like quality of the pre-drop vocals in Skrillex’s ‘Kill EVERYBODY’ – two excellent examples of creative downsampling.
There are even artists and producers out there – most of which operate on 32-bit or 64-bit processors – who downsample their audio to create retro 8-bit music compositions and arrangements!
Altering the ‘Sample Rate’ Function on a Bitcrusher Plugin
Also, if you listen closely, you might hear a layer of white-noise style effects behind the guitars in the Queens Of The Stone Age track and the vocals in the Skrillex track.
This has been achieved by utilizing the Bitcrusher through its Sample Rate function, which involves bringing the Sample Rate function down from ’24-bit’ to a lesser value.
A lower Sample Rate setting on a Bitcrusher purposefully invites more calculation errors as the digital signal processor inside your DAW attempts to generate a snapshot of the original waveform based on a lower and lower number of samples – and the result can sound delicious!
Oftentimes when using the Sample Rate setting on a Bitcrusher, artists and producers will choose a bit value which is a square multiplication of 2 – e.g., 4-bit, 8-bit, 16-bit, 32-bit – for the sake of an even Nyquist rate.
This causes the quality of the white-noise effect to appear well spread out and pleasant in tonal character.
Read more about understanding the Nyquist rate here!
Are Upsampling and Downsampling Necessary?
Although upsampling and downsampling are intriguing techniques with a lot of troubleshooting potential in audio mixing and large scope for creative originality, they are in no way vital to most audio projects.
At that, they’re hardly ever used as practical tools at all!
Provided the audio mix is strong and the recorded materials are clearly defined as a result of good recording equipment, the project’s sample rate shouldn’t need to be tampered with.
However, here’s a rundown of the above ways we CAN use upsampling and downsampling:
- If you are working on a film project, you might need to up that project to 48hz.
- If you’re fixing errors in the resolution of a finalized audio export, you’re going to need to increase your project’s sample rate (not as far as to 48khz in most cases, though).
- If you’re going for a lo-fi sound, decrease your sample rate by a step or two on the offered drop-down list in the Sample Rate section of your DAW.
- Expect some ‘quick and easy listening’, low-resolution MP3 uploads to SoundCloud and other streaming platforms to sound a little degraded in quality.
- And finally, if you’re feeling creative… get Bitcrushing!
Before you go, check out our guide to Ableton’s Simpler vs Sampler!