How is a 32-bit float file recorded?

The MixPre II Series are the industry’s first and only USB audio interfaces to offer 32-bit floating point audio for MacOS Catalina and Windows 10.¹ Audio inputs on the MixPre appear as 32-bit floating point audio sources to a DAW for recording and mixing. DAW software can output 32-bit floating point audio back to the MixPre. This is a significant advance because the gain setting on the USB interface and in the DAW are no longer critical to achieving excellent sound. With 32-bit float, audio gain decisions can be made after recording. This may sound provocative to an experienced sound engineer. In a 24-bit world, gain needs to be set correctly–high enough to overcome noise yet not too high that headroom is affected. 32-bit float changes that. 

So how is a 32-bit float file recorded? Four distinct subsystems are required for 32-bit float operation. These include:

  1. A very high-dynamic-range analog microphone preamplifier
  2. Multi-stage A/D converters
  3. 32-bit floating point processing and signal paths, including USB interconnection
  4. 32-bit float WAV recording file format

32-bit floating point data is required throughout the entire signal chain, or the system doesn’t work.

32-Bit Float Stages

Analog Microphone Preamplifier

The analog microphone preamplifiers in the MixPre II recorders have >142 dB of dynamic range. This allows the input to capture the entire dynamic range of any connected microphone. The custom-designed topology, with large-geometry, discrete transistors connected in a class-A configuration, result in the lowest noise and distortion possible. The preamplifier is so quiet that connected microphones will generate more noise than the preamplifier. Regardless of MixPre’s gain setting, the microphone’s signal and its noise is heard, not the preamplifier. It is the definitive “transparent” microphone preamplifier.

Multi-Stage A/D Converter

The analog-to-digital converter is one of the most important stages in a digital recorder. The best audio A/D converters presently on the market yield roughly 130 dB of dynamic range. This on its own is quite excellent, but not good enough to capture the entire dynamic range of all microphones. Sound Devices patented a method of combining multiple fixed point 32-bit A/D converters. This unique combination results in an A/D conversion stage which yields over 142 dB of dynamic range. The output of this stage of the hardware is 32-bit floating point audio data.

32-Bit Float Processing

It is worth reviewing the technical article on the math behind 32-bit floating point. The important point to remember is that a 32-bit floating point representation of an audio signal exceeds 1500 dB of dynamic range, so virtually any amount of gain can be added to or subtracted from the signal. Said another way, it is impossible to clip the signal or bury the signal in the noise floor. All processing, including gain, EQ, routing, etc., is performed using 32-bit floating point math.

When USB audio interfaces are connected to computers and DAW software, the USB audio interconnection becomes part of the signal chain. Typical USB audio uses 24-bit fixed point signals. The MixPre II Series, when set for 32-bit float, can send a 32-bit floating point signal over the USB to MacOS Catalina and Windows 10 with the ASIO driver. The DAW can then perform all of its processing using 32-bit floating point math resulting in a signal that cannot be clipped or buried in noise.

32-Bit Float WAV Files

The WAV file is a universal container for audio data, capable of storing material in a variety of bit depths, including 24-bit fixed point, 32-bit fixed point, and 32-bit floating point. In order to maintain the benefits of 32-bit floating point audio, audio data must be stored as 32-bit floating point. The audio application which is receiving the 32-bit float WAV files needs to support this format. While many audio applications do support 32-bit float, support is not yet universal.

How Is 32-Bit Float Different From 32-Bit Integer?

32-bit float signals should not be confused with 32-bit fixed point. A 32-bit fixed point signal can accommodate considerable dynamic range, roughly 192 dB. While 32-bit integer offers a much larger dynamic range than 24-bit audio’s theoretical 144 dB, it pales when compared to the 1500 dB available to 32-bit float signals. If, for instance, processing adds 90 dB of gain to the 142 dB of dynamic range coming in, then 232 dB would be needed to capture this signal with no loss in headroom or noise. 32-bit float can do this, and 32-bit integer cannot.

It’s the Hardware

The MixPre II Series hardware described above is what enables these units to offer 32-bit floating point audio. Having the massive latitude of 32-bit float audio available for audio processing and storage yields significant benefits. Knowing that the signal from a microphone can never be over-driven or recorded too quietly brings peace of mind and ease of recording, allowing the recordist to focus on their craft rather than dialing in an exact gain setting. Whether recording to a DAW via USB, or recording to an SD Card, the MixPre II Series offers the next-generation benefits of 32-bit float today.

¹ As of March 2020, no other USB interface offers 32-bit floating point I/O.
² Based on commercially available analog to digital semiconductors as of March 2020.