top of page
Production recommendations
Before we go into the production recommendations, here are the most important terms that will appear in some cases and are closely related to each other. In order for the recommendations to be understandable, you should know them. The question of how much headroom makes sense is also answered here.
If you are not familiar with the unit Bell / Decibel (dB) and the various reference levels in sound engineering, please do some research for further understanding.
Terms
Noise floor:
Any analogue device such as a converter or a microphone has a background noise. This is the lower limit of the technical transmission range. If a level exceeds this limit, the so-called maximum system dynamics in analog technology begins.
Signal to Noise Ratio:
The signal-to-noise ratio is the unit of measurement for the technical quality of a useful signal that is embedded in a noise signal. The distance between the mean value of the background noise and the mean value of the useful signal defines the signal-to-noise ratio.
System dynamics / dynamic range:
As already mentioned, maximum system dynamics begins when the noise floor has been exceeded. The upper limit of the maximum system dynamics is the control limit (+24dbu).
The (effective) system dynamics refers to the area between footroom and headroom, which are included as a safety distance to ensure the best possible quality of the signal from a technical point of view.
Unlike in digital technology, where 0dBFS is the control limit and when it is exceeded, unpleasant non-harmonic distortions (clipping) immediately arise, in analog technology 0dBu is the threshold for headroom. Here the level is often deliberately set higher in order to add (coloring) harmonic distortions. The distortion of the signal is a continuous process in the analogue sector.
During digitization, an analog time/value continuous signal is converted into a digital time/value discrete signal. However, I won't go into the A/D conversion any further.
So after the conversion we get a digital signal with fixed sampling points which records the time and amplitude values.
There are 6dB of dynamics per one bit, so a file with 16 bits has a dynamic range of 96dB and a file with 24 bits already has 144dB. A file with 24 bits that is fully controlled can represent the dynamics in (2²⁴) 16,777,216 values. A 16 bit file in (2 ¹⁶ ) 65,536 values. If the 16 bit file is now 6dB quieter, there are still 15 bits of dynamics remaining and therefore (2¹⁵) 32,768 values to represent the dynamics.
In a 24-bit file you have such a large dynamic range that it is almost impossible to have problems with the resolution or the signal-to-noise ratio. You would have to control it extremely quietly so that this could become problematic in mastering and there would be too little dynamic for further processing. Remember, if a 24 bit file is controlled to -48dBFS (which would be extremely quiet), there is still a dynamic range of a fully controlled 16 bit file.
As a comparison, the dynamic range of the human hearing range is theoretically 120dB-140dB, but from 120 dB(A) there is a very high risk of injury and for most people this is well above the pain threshold. The effective dynamic range is approximately 100dB. From quiet breathing noises of 10dB(A) to a loud chainsaw of 110dB(A), which most people find unpleasantly loud.
So choose a headroom that's right for you, so that you can work well during recording/production and certainly don't overdrive, and you don't have to use limiters, clippers, etc. to prevent this. Unless you want to achieve a specific effect with limiters, clippers, etc. and use them as artistic sound design.
Production recommendations
For music productions, I recommend producing in 44.1 kHz / 24 bit, unless for example. a Hi-Res format or a vinyl production is targeted. As best as possible, production is always carried out in the format of the target medium and the sampling rate and bit depth are never changed. For productions with different target media that have different requirements, I recommend producing at the highest required sampling rate. For CD productions, I recommend producing in 44.1 / 24 bit.
Downsampling and downsampling the bit depth using dithering should always be carried out by the mastering engineer!!!
Leveling: When creating recordings, I recommend leveling as loudly as possible. Guide values are
-12dBFS to -6dBFS and maximum peaks at a maximum of -6dBFS to -3dBFS. This still leaves at least 3dB of headroom in case a level peak goes beyond that. For very dynamic productions, such as when plugins or hardware are looped in and recorded, I recommend aiming for -18dBFS to -12dBFS with a maximum peak of -6dBFS. Even though today's audio devices are very low-noise, the signal-to-noise ratio is as high as possible and the system dynamics are optimally utilized.
When producing with software synthesizers, you should also keep an eye on the output gain. Many of these synthesizers, such as. Serum or Omnisphere already overdrive with the in-house standard presets. This would result in the following plugins in the signal chain being started up too "hot", sounding worse, and their workflow being worsened.
Mixdown : In the mixing process, I recommend doing gain staging first. This has two significant advantages. On the one hand, you are then in a range where the plugins are best controlled, the so-called sweet spot. Depending on the plugin, this is between -20dBFS RMS and -14dBFS RMS. The second big advantage is a much finer fader resolution in the mixer channels, as the dB gradations are much larger in the upper range. This means you don't have the problem of the faders having to be pulled very far down and the smallest changes causing large differences in level.
During mixing, I recommend aiming for a total of around -10dBFS RMS (master channel). Then the peaks should normally reach around -6dBFS, but will certainly not shoot above -3dBFS.
For electronic productions, I often control the kick and bass segment at around -10dBFS RMS. Kick and bass have the most energy in the song. If the other elements are added, you end up with around -6dBFS to -3dBFS.
Depending on the genre, I can also recommend using a limiter on the sum. This already increases the loudness and you get an idea in which direction the end product is going. Depending on the genre, this should not intervene at all, absorb a spike or impose a slight limitation. It is important to remove this or bypass it before exporting the mixdown!
bottom of page