Taperssection.com
Gear / Technical Help => Ask The Tapers => Topic started by: ts on August 14, 2022, 11:39:53 AM
-
So it's useful when levels are recorded to low or to high, but what about other situations where levels can clip unexpectedly. Useful there to? Like steady strong winds.
-
'Too high' is tricky, any of these devices can be overloaded at the input before conversion, and there's no helping that. Wind with it's strong low frequency content is a prime contender for permanent clipping.
-
I wouldn't think so... 24-bit PCM already has 144dB of range, and few, if any ADCs take full advantage of that.
Background noise in most quiet rooms is above 20dB, even, and likely much louder in venues with ventilation and other people breathing.
32-bit float is convenient for post-processing, since most software uses 32-bit float internally and will convert wav/flac/whatever to float for calculations before reconverting it back to wav/flac/whatever.
I wouldn't rely on float as a storage/archival format, though, since it lacks metadata and built-in checksumming.
Larger sizes also means more wear on storage devices and higher likelyhood of bitflips and data corruption.
My recorders only support 24-bit PCM wav (and nominally so, ISTR the M10 only has about 80-something dB of dynamic range).
I always FLAC everything ASAP to multiple HDDs (using btrfs on Linux) and do all post-processing on machines with ECC memory to avoid bitflips.
-
You can clip it at +770dbfs lol
-
You can clip it at +770dbfs lol
That's in the digital domain, but it's important to note what EmRR said above - you can overload the input stage on the preamp, and that's not fixable.
-
I wouldn't rely on float as a storage/archival format, though, since it lacks metadata and built-in checksumming.
Larger sizes also means more wear on storage devices and higher likelyhood of bitflips and data corruption.
32-bit float audio files do not support metadata? Mine do.
The storage size difference is negligible has never been enough to concern me. I record in 32FP/48kHz, and files are only slightly larger than 24/48. I convert my originals to FLAC for archiving, but keep them floating-point.
There really is no higher risk of data corruption compared to other formats. Everyone should be running regular backups anyway, preferably to optical storage in addition to HDD/SDD.
-
Personally, I wouldn't characterize a 33% increase in file size as "negligible" (irrespective of anyone's opinion with respect to added value or cost)...
-
Personally, I wouldn't characterize a 33% increase in file size as "negligible" (irrespective of anyone's opinion with respect to added value or cost)...
Fair enough; poor choice of words. I clarified my post. I never keep the original WAV files anyway; I make FLACs of the untouched originals and delete the WAVs after post work is done.