This FAQ aims to demystify what 32-bit conversion technology implies, resolve the most common misconceptions and provide some insight as to why we use 32-bit converters in SSL products such as PURE DRIVE QUAD, PURE DRIVE OCTO and SSL 12.
“Do I still need to worry about setting the input gain when using my SSL device that has 32-bit converters? I’ve heard 32-bit devices can’t be clipped...”
In short - yes, you do still need to set an appropriate gain to make sure your SSL device does not run the risk of clipping.
Why?
We use 32-bit fixed point (also referred to as ‘integer’) converters. The confusion comes from a small subset of audio devices on the market that happen to use a 32-bit floating-point recording system. These devices usually involve complex multi-converter systems per recording channel (most often two converters, covering different ranges) as a way to achieve an extended dynamic range at the A-D stage. Also, the digital stream from the A-D stage is presented in a 32-bit float format, which numerically allows for a dynamic range beyond that of the actual converter stage. Such products are designed to suit specific industry needs like field-recording, where astonishingly high dynamic range material needs to be recorded or where it is not possible to know beforehand exactly how loud your source is going to be. Imagine been tasked with recording a space shuttle lift-off and having to ask for another take because your recorder clipped...
Our friends at Sound Devices do offer products aimed at field-recording applications that deploy 32-bit floating point recording formats, if you’re curious to find out more about the technology.
Bottom line is, your PURE DRIVE QUAD/OCTO or SSL 12 (with a correctly set preamp gain) will do you just fine to record your drummer’s sloppy tom fills... and here is why:
First, the distinction should be made between recording/playback formats and mixing formats. Recording format refers to what type of digital encoding is used for the digital audio stream generated by the ADC (analogue to digital converter). Equally, playback format refers to what type of digital encoding is used for the digital audio stream received by the DAC (digital to analogue converter). Mixing format is what is internally used by your DAW to process the audio streams. Usually it’s a floating-point format: without getting into the maths behind this statement, we just need to know this choice guarantees practically infinite dynamic range for any practical processing within the DAW. Care must be taken whenever fixed-point format is involved, where dynamic range is practically limited, and it is possible to experience clipping and distortion if a correct gain staging is not observed.
In a recording scenario, with a signal source (microphone or instrument) into a preamplifier followed by an Analogue-to-Digital converter, we need to make sure that:
- The input signal is within the dynamic range of the preamplifier. Failure to do so will result in distorting the signal within the preamplifier. This distortion cannot be undone at a later stage.
- The output signal of the preamplifier is within the dynamic range of the Analogue-to-Digital converter. Failure to do so will result in signals which would be represented above 0 dBFS to be clipped at exactly 0 dBFS, which causes distortion.
In a playback scenario, audio generated by the DAW should be correctly gain-staged to not exceed 0 dBFS, which is the loudest or highest level representable by the Digital-to-Analogue converter. A perfect example is a mix-down in your DAW of multiple sources. The sum of these signals on the master bus could result in a signal louder than 0 dBFS. While it’s possible to lower the level digitally within a DAW (under the assumption a floating-point architecture is used) without causing distortion, the same cannot be done once we reach the DAC. Any signal exceeding 0 dBFS at the DAC will be represented as 0 dBFS i.e. clipped, resulting in distortion.
“So why does it matter to have 32-bit converters?” (Advantages of 32-bit converters vs 24-bit)
There are some distinct advantages in choosing an SSL interface that uses 32-bit converters.
The latest product offerings from major AD/DA converter manufacturers nowadays starts at 32-bit of resolution. This means you have the latest technology being used in your device. AD/DA Conversion is not all about the resolution: other figures like dynamic range, maximum sampling frequency and distortion (often referred as THD+N) play a role in defining recording/playback audio quality. The THD+N figures for PURE DRIVE are notably class leading. Significant improvements have occurred in conversion technology over the years since it first became accessible in the late ‘90s / early ‘00s. Knowing your product uses a 32-bit converter reassures you your device is at least up to speed with modern conversion quality levels.
A reason here at SSL we favour 32-bit DACs in our latest products, is that they allow us to offer digital volume control of the outputs without incurring loss of quality – this is the case for SSL 12. The same cannot be said of digital attenuation on a 24-bit system, where truncation error results in random noise loud enough that when summed to your signal effectively increases the noise floor. This artifact can be avoided by use of dither but at the expense of additional computing power in the device’s DSP. A direct consequence of implementing output level digitally is perfect stereo matching, or at least as good as the highest tolerance component in the analogue chain. On SSL DAC to output paths, that would be metal-film resistors which are 0.5% or better (better than +/- 0.08 dB stereo mismatch).
“Does a 32-bit converter give me higher dynamic range than a 24-bit converter?”
Short answer: it depends on the converter used and on the circuit implementation, but not inherently due to bit resolution.
- 16-bit conversion gives a theoretical dynamic range of 96 dB.
- 24-bit conversion gives a theoretical dynamic range of 144 dB.
- 32-bit conversion gives a theoretical dynamic range of 192 dB.
Noise inherent to the analogue electronics (within the converter and around) is actually the bottleneck here and usually lower noise devices mean higher cost and increased power consumption. This means that you will find some examples of older 24-bit high spec converters which offer higher dynamic range than some newer 32-bit converters.
“Can I still record and playback 24-bit audio with a 32-bit interface?”
Yes. This is a setting in your DAW, and you can choose whichever resolution you might prefer there.
Figure 1: Audio format bit depth selection within Reaper DAW
Common terms
- Dynamic Range: Difference between highest (loudest) undistorted representable signal and lowest (i.e. noisefloor). This quantity is normally expressed in dB.
- dBFS: Decibel Full Scale. Used to express a level referenced to the maximum (i.e. full scale) representable quantity by a fixed-point system. Maximum undistorted representable level is 0 dBFS. Negative values are undistorted, safely representable values. Positive values are distorted, not correctly representable values.
- ADC: Analogue-to-Digital Converter
- DAC: Digital-to-Analogue Converter
Useful links
https://www.mixinglessons.com/bit-depth/
https://www.youtube.com/watch?v=4YRp-FIsNDA&ab_channel=JulianKrause