It means that the dynamic range is so high that it’s almost impossible to overflow gain, so the recording will not get distorted.
So, you don’t need to be very precise with gain - just record!
Many on zzzee Zzzerman forum use 32bit float devices to make quick recordings of gigs or practice, cause they want to play, not fumble knobs (ok, some like to fumble knobs, but that’s a different story :-))
I’m sure that @howard can explain it much better!?
Review:
If you have the money, get one of the better Zoom devices…
Everything you said is true but 24 bit dynamic range equates to 144dB of dynamic range which trust me you are not using fully if you can still hear at all. Additionally computers use 64 bit IEEE double precision internally anyway so in practice 32 bit interfaces give no real value over 24 bit.
Also, the low order bits get consumed fast so in practice only a few dB of noise will make 32 bit no better than 24 or even 16.
tl;dr, Prefer faster sampling rate and preamp quality over sample width.
The point of 32 bit floating point is not about hearing better quality, it’s about recording audio with (unforeseen) high peaks and low volume better (= gig situation), if I understand it correctly:
The point is, there is no point in recording more than 24 bits of dynamic range.
Each sample of audio is recording an amplitude between silence and the maximum loudness of its dynamic range.
24 bits of dynamic range is the equivalence of the range of sound loudness between silence and immediate irreversible hearing trauma, about like a gunshot directly next to your ear.
Any dynamic range above 24 bits is basically one of two things: unused headroom (best case) or noise (more likely). So, 24 bits is plenty.
Now - is it important to use more than 32 bits internally? Yes absolutely, this is where the floating point math ends up consuming way more bits than 24. Everyone should be on 64 bit computers for music for sure.
But for sample size? Nope, 24 bits is plenty. You’re way better served by increasing audio rate and preamp quality.
You know the Akai S900/S1000 etc samplers every band used in the late '80s and '90s? Those were 12-bit and sounded awesome. The Ensoniq Mirage was 8-bit and was a superb instrument.
In this case I orient myself to the gigging / rehearsing musicians around the corner. They use those devices for the “quick recordings” … of course not for their “professional stuff”, where the FOH guys & gals work their magic….
Look at this another way. Restating my above example in terms of real life values: Look up the dynamic range that your audio interface is capable of recording. Often this is expressed as its noise floor.
For most audio interfaces this will be in the 120 to 130 dB range.
Anything those interfaces capture outside of that range is noise. So the 144dB that 24 bits gives you is much more than enough. The additional 8 bits that 32 bit gives would essentially be (very) subaudible noise.
Incidentally, I held this misconception for a long time too (more bits must be better, right? Number go up!) but the DSP programmers convinced me I was wrong, and then when I started studying DSP programming I understood why. It’s not obvious at all, which is why I said it’s ok to be wrong here - 32 bit samples doesn’t hurt, it just also doesn’t help, and since it is stored as 64 bit internally anyway, the capture size of the samples doesn’t matter at all for things like gain staging, despite that article’s opinion.
If we want to reframe the same values in terms of the loud gigging Germans, remember that 24 bits is enough to accurately represent 144dB of dynamic range. One way to think of dynamic range is the range of values between silence (minimum) and the dynamic range in decibels (maximum). The loudest concert I could find reached an unofficial peak of 143 decibels (which would have caused permanent hearing damage to those in the area where that range was reached).
Leftfield hit 137 at one of their shows and the building began to fall apart (raining dust and plaster on the crowd). 24 bit can get more than twice that loud.
The way dynamic range is used in audio processing goes the other direction (from max signal at 0dB and min at a negative decibel value at the bottom of the dynamic range, with anything quieter than than being noise). You can consider it the accuracy of the signal an interface can sample. 24 bits is enough to accurately represent down to -144dB, which is well below the noise floor of any audio interface I am aware of.
Does it make sense to record live gigs in 32 bit float format?
Answer:
Yes, it makes excellent sense to record live gigs in 32-bit float format, especially when dealing with unpredictable and wide-ranging audio levels as often found in live concert environments.boyamic+1
Advantages for Live Gigs
32-bit float captures the entire dynamic range of both quiet moments and sudden, loud peaks without the risk of digital clipping,evenif input gain is set incorrectly.masteringbox
It removes the need to ride or adjust gain levels during the performance, allowing focus on the event rather than worrying about the recording equipment.zoomcorp+1
In post-production, both overdriven/clipped sections and quiet passages can be balanced without introducing digital artifacts or noise, ensuring a cleaner, more usable recording.zoomcorp+1
Practical Considerations
Larger file sizes (about 33% more than 24-bit files) require sufficient storage for an entire live gig.boyamic
High-quality microphones, proper preamps, and good mic placement are still needed; 32-bit float helps with digital limitations, but not analog ones.masteringbox+1
It is especially valuable in live settings—where uncontrolled volume spikes, crowd noise, and on-stage dynamics can defeat the static gain structure of 24-bit recording.masteringbox
Overall, for live gigs where one chance to capture everything cleanly is critical, 32-bit float is an ideal choice due to its dynamic range, post-production flexibility, and safety net against both clipping and buried signals.zoomcorp+2
In my opinion, this is the core part: “without the risk of digital clipping, even if input gain is set incorrectly”.
With 32 bit float you can be relaxed about setting the (in) correct gain, which is especially good in situation where you have no idea how loud the PA can get??!
This does not contradict what you’re saying. It just adds a buffer that is great when thing turn unexpected (like with every live gig :-))
It definitely doesn’t hurt and eventually all ADC’s will be 32 bit sampling so the distinction will be moot; still, in the end, the mic preamp on the H1 only has 122dB of usable dynamic range (about 20 bits). So, depending on the preamp it might make it easier to record extremely loud shows, but there won’t be any other benefit.
SAIGON KICK - I LOVE YOU https://youtu.be/iy4dzIsZMpI?si=VuPaJnyH_7hvS4Ml Decided it was time to change the strings out. Love the neons, but with all the paint that comes off over time, I decided to go back to the regular strings, but with that being said, new strings have that extra boomy sound to them. Love this song so much, SK had a great sound so it was a no brainer for me. Enjoy and have a great weekend.
I noticed from your videos that the pigment seems to come off with time. Do you know whether this is because the pick wears off the color coating? Does it happen with finger plucking as well?
Thank you! I don’t normally play finger style, but I I can say that the paint was wearing away with my fretting hand as well, so I’m assuming it’s just going to happen over time either way Although it did wear faster where I use a pick with my right hand so. I truly love the colors they offer, but it does make a mess over time. It got all down in the slots around the pickups. I wore these orange ones out more so than the others I tried. This time it just made me change my mind on using them again haha
It’s such a great song!