Last Updated 6 hours ago
Filmmaking Dictionary: Number 3 / 32-Bit Float Recording
Definition
32-bit float recording stores audio samples using a floating-point number format. Its very large numerical range can preserve values above the usual 0 dBFS reference that would exceed the range of a conventional fixed-point delivery file.
With a suitable recording system and compatible software, those values can be reduced in level later without the file itself having clipped. The benefit depends on the signal remaining intact before it reaches the file.
The Recorder Matters
A floating-point file cannot recover a microphone capsule that distorted, an overloaded analog input or a clipped wireless transmitter. Some recorders combine multiple converter paths to capture a wide input range, but every physical system still has limits.
Check the recorder’s maximum input level and the microphone’s capabilities. Setting the file format to 32-bit float is not permission to ignore the equipment feeding it.
Advertisement
What Recovery Looks Like
A recorded waveform may appear to run beyond the editor’s normal full-scale display. If the floating-point data remains intact, reducing clip gain in a compatible application can bring it into a usable range. That is different from turning down an already flattened waveform caused by analog overload or fixed-point clipping.
For example, an unexpectedly loud impact may be preserved by a suitable recorder even though its initial waveform display looks excessive. The same file format cannot remove crackling from a damaged cable or restore dialogue buried under clothing noise.
Monitoring and Delivery
Headphone outputs, digital outputs and downstream devices may have different limits from the recorded file. A distorted monitor feed does not always describe the stored data, but it still needs investigation. Confirm the recorder’s routing and review the file rather than assuming every overloaded sound is recoverable.
Before exporting to a fixed-point format, set levels appropriately so the conversion does not clip. Also confirm that the receiving editor or sound application supports the original float files.
Practical Checks
- Match mic-level and line-level connections correctly.
- Respect the recorder’s maximum input specification.
- Monitor for microphone, cable and radio problems.
- Test the complete import and gain-adjustment workflow.
- Distinguish 32-bit float from 32-bit integer.
- Manage final output levels before delivery.
Advertisement
Frequently Asked Questions
Does 32-bit float eliminate the need for a sound recordist?
No. Microphone placement, noise, monitoring, routing and production judgment still determine recording quality.
Can it fix any clipped recording?
No. It can preserve recoverable levels in the file only when the preceding signal path has not already lost the information.
Related Terms
Browse all terms starting with 3 or return to the filmmaking dictionary.