Topping and Tailing Ripped Beats - Truncating and Normalising
How to top and tail drum sounds extracted from a beat. Topping and Tailing refers to truncating and normalising audio along with fade ins and outs.
Before you start to layer drum sounds or slice and dice drum beats there a few steps you need to take to ensure all the samples are at the same volume and have all dead spaces removed before and after the sample. This is an Old School technique called ‘topping and tailing’. Topping refers to normalising the audio to a specified ceiling value (usually at just below 0 dB) and tailing refers to the removal of dead space and the application of fade outs if required.
As most samples are lifted (recorded) off vinyl their start and end times will vary as will their gains and we need to make sure that all samples are peaked to 0 dB and truncated to remove dead space. This ensures all samples start at the same time and their relative gains are equal. This makes the whole layering process easier to execute and manage.
Drum beats are the same in that start and end times have to be defined and dead spaces removed. All the peaks in the drum beat need to be managed so as to provide a constant volume across the whole drum beat. Achieving the exact cycle start and end times is important if you want a drum beat to loop perfectly.
Normalisation is a digital signal processing function that’s available in a lot of digital audio editing software. It scans through the programme material for the highest peak level, and if that level doesn’t reach the maximum available dynamic range, the software boosts the overall signal so that the peak hits the highest level possible. This is referred to as Peak Normalisation. If the audio displays the highest peak value as -3 dB then we need to use peak normalisation to raise that level to 0. This in turns raises all peak transients by the same value.
RMS (Root Mean Square) Normalisation works differently. Instead of using the highest peak value as the reference it now takes all peak values and divides them by the number of peaks. Once this average has been established normalisation will use the value from this to normalise to 0.
There are a couple of things to be wary of: because Normalisation normalises the whole audio signal it will also raise the noise floor in the process.
If you need to normalize then think carefully about whether to use use Peak or RMS (average level) as each will output a different response.
I tend to find that RMS (Root Mean Square) works best on long audio files that have varying peaks and troughs. Peak tends to work well on single shot samples, much like drum hits. BUT try both and judge for yourself.
Normalisation still has a part to play in this industry and in particular sound design whereby samples are often truncated and peak normalised to 0. This allows the samples to have a constant value when processing.
These prep steps are important if you want to layer samples. Having all the sample start times start at 0 crossings and with gains being matched ensures you can layer without any timing and level problems. Additionally, by making sure samples start at 0 we are guaranteed that triggering will be the same for all the samples. The worst scenario is having samples that start at different times and at different volumes. With drum beats we tend to chop/slice beats into smaller drum element components which we can then use in our own drum beat creations. So, the same principles apply as topping and tailing individual samples.
In the video I use Timbaland’s Ice Box and rip a section for processing. I explain how to truncate the beat and normalise it. I explain each process and guide you step by step making sure to cover important areas like o point crossings, peak versus RMS normalisation, cycle start/end etc…
Plugin used in this video:
Topics covered in this video are:
- What is Topping and Tailing
- Normalisation – Peak/RMS
- Cycles – Start and End
- Sample Cycles
- Sample Count
- Resolution versus Size
- Sound Design Prep