How to use Sampling Frequencies and Bit Rates to Your Advantage

I had a conversation recently with a student from Garnish School of Sound who described himself as a “glitch-dubstep-lo-fi composer” He went on to describe a couple of his tracks which turned out to be pretty much what you might expect from such a description. When I asked him about his production he said to me: “I make all my tracks with a low sampling-rate and bit-rate for the whole project”. At this point I was a bit confused and so asked him to elaborate. Apparently all his samples and synthesizers were imported / played automatically at 22,050Khz sampling rate with 12bit bit-rate. The entire project was also set at those values. Now utterly dumbfounded I asked his reasoning. “Because I want everything to sound really grungy and lo-fi” was the response. I queried if his computing power was any issue and he replied that his computer was top notch and that he could set a project at top rates if he wanted to. What he did complain of however was a problem in getting any kind of depth or warmth in his tracks.

After returning home after this perplexing chat I got onto the internet and checked out his music on SoundCloud. The result was what I had expected: a flat and poor quality sounding collection of tracks. The actual composition was damn good including some fantastic rhythm lines with great syncopation and groove. The harmony and melody aspect was all well presented, the bass-lines all evolving and suitably Dub-steppily wobbly and wonky with some interesting twists. The mixes were balanced, clear and well compressed. Yet all the sounds were lifeless, lacking in depth and had a serious hole where a big dollop of Dub-step phatness should be.

The reason I felt such confusion whilst chatting to this guy was because the idea of intentionally downgrading the quality of the sound before doing anything else to a piece of music is a complete mystery to me. Distortion, bit-crushing and over-driving sounds is a great way to manipulate sound and for getting new ones, but the great thing about them in this instance, and with all effects really, is that you have the choice to put them on. With a clean recording, well edited samples and as high a quality environment as possible to compose or produce in, you have all the options in the world for putting sounds through the mill as far as you like. However, if you start a track with viciously over-compressed drum hits, hugely overdriven synthesizer samples and gnarly super limited bass, that’s all you have to work with. You can never get back lost dynamics, a clean signal of an distorted instrument, or the quality stripped away by bit crushing.

Simply enough, the guy I had chatted to had an idea about what he wanted to achieve and then set about getting it in completely the wrong way. It’s an easy mistake to make thinking that because a really heavy distorted sound in a track you like sounds fantastic that applying a massive amount of distortion will produce the same or similar sound; sadly, it doesn’t. Whilst it’s a bit of a blunt saying, put rubbish in, get rubbish out. Fill in the word rubbish with whatever you like here. To illustrate this, you could take say a sustained power chord from a guitar and make two copes in your DAW on separate tracks, both recorded at 24bit, 96kHz quality. Take one and convert it to say, 128kbps mp3. Then place identical guitar-amp plug-ins such as Logics amp simulations and see which one can be made to sound beefier. You can replicate this with a high quality synthesizer sound or perhaps a vocal take. Listen back carefully and notice the difference in texture, depth and richness that the better one offers, especially after being put through effects. 96Khz, 24bit sound compared to mp3 is quite extreme but it simply points out the difference quality can make. I’m not saying you need pristine audio to make great records, just that if you have the quality, why throw it away?

Scratch just a tiny way into the surface of music technology and you will be sure to come across the continual discussion between the benefits of digital and analogue equipment. Just a few decades ago, digital technology was in many ways the holy grail of music production. No more wow and flutter from tape and synthesizers that didn’t need tuning and settings to save rather than having to jot down parameter values. Also, the possibilities of digital sampling and not needing to have to hire musicians made many producers weak at the knees. However, it was not to be. As soon as musicians discovered that samples could nowhere near play like musicians, and that memory capacity and computing power was only capable of using inferior sounds, they breathed a sigh of resignation and got back on the phone to their favorite instrumentalists. The good thing here is that many new styles of music grew out of these restrictions, but that’s another story. Even worse though than not having the facilities to achieve what digital audio promised, it was soon seen that things were a very long way off in terms of audio quality. Skip to the present and we have the ability to accurately sample instruments with velocity layers, multiple voices, anti-machine gun facilities and probably more computing power than all the first batch of music computers put together. Yet still a trained ear can tell the difference between a sampled instrument and a real one and it has to be a pretty good impression to fool anyone. I slightly digress, but it all goes to say that quality is good and should be maintained.

Another consideration on the quality of sound when making music on computer is the soft synth. Once again digital audio doesn’t quite manage to make the grade of analogue. Don’t get me wrong, I think digital sounds fantastic, it’s all I use and I’ve never owned an analogue synthesizer in my life, but I do know how great they sound. One of the major reasons why they sound better is harmonics. Because the electrical circuits in analogue synthesizers is imperfect, it creates slightly imperfect signals. This is how the warmth and richness of analogue shines through. In comparison digital is too perfect, simply a collection of ones and zeros re-created into an imitation of the result of an electrical circuit. It’s this warmth, or lack of it which helps explains another example. To imitate analogue warmth there are plug-ins to imitate it, or other methods such as putting a sound through a tube driven guitar amp and re-recording it to give it some bite. Simply enough, the more pristine detail that a digital sound has, i.e. the higher its sampling rate, the more harmonics it then has to warm or ‘excite’ to give it that all important tone.

Just to point you in the right direction, I suggest to all digital musicians to make the following precautions. Always work with the highest bit-rate and sampling rate that your system will allow. If this means occasionally having to freeze a track or even bounce a number of tracks to re-import as audio, so be it. Always record at the highest rates you can and don’t downgrade that signal until the very last step, i.e. dithering and conversion in the mastering stage. Make sure that no equipment or software you use downgrades audio quality. Don’t sacrifice audio quality for hard disk space with audio libraries. After all, hard disk space is as cheap as chips these days, take advantage. If you have to convert, make sure you do it the longest but best quality setting.

Most importantly in all this, don’t let poor quality limit your choices with music making. Always make sure you can take your pick for as long as you can before losing that possibly all vital quality. After all It’s better to have the option of measuring twice and cutting once. If you come to the end of your track and still want to filth it up worse than the most distorted gabba ever made, please do, and you’ll be glad you have the option. Also, spare a thought for the poor sods in the early eighties shaving milliseconds off the audio tail of a sample to save precious kilobytes so that they can have six instruments in one track rather that five.