does 32-bit float recording glitch for other people?

[DJ Saint-Hubert]

I'm forced to record as 32-bit int because if I record in 32-bit float, I get bursts of loud sound or silence like a sample long, but with multiple tracks it messes up the whole timing of the track if I edit them out. The bursts are like the maximal volume for 32 bit float I think, so it's easy to spot them. There's not that many bursts but it's enough to be annoying. But I heard internal 32 bit float to 32 bit int conversion is no better than 24 bit int, so I'd rather output straight to 32-bit float

[IXix]

Yeah, I had similar trouble with float32, tiny random spikes and drops peppered across the output file. That was several years and a different laptop ago so I've not tried recently but 32bit int is a ridiculous amount of dynamic range, I don't think it's much of a compromise.