DO YOU ACTUALLY NEED 32 BIT FLOAT?

[@johnnymiller3684]

"Specific use cases" is the perfect answer to the 32-bit float conundrum. Professionals, and even semi-professionals, might never find the need to use 32-bit float. On the other hand, there is a huge, worldwide market for this type of tech when it comes to amateur content creators (albeit not for live streamers). For instance, my friend and her daughter do a small podcast, and they often clip the audio, even after setting the levels correctly (they both laugh extremely loud). So, I let them borrow my Zoom F3 to record one of their podcasts, and it eliminated the clipping problem. They immediately bought a Zoom F3. Lol

[@borislavatatchev]

I recently got Zoom F3 Field Recorder. It is an incredible device. It saves my audio every time. Extremely high quality interface with extremely low noise floor. Simply the best.

[@DevonBlack]

To bounce off of what you're saying, there are very few things in audio that are NEEDS. Outside of one's specific use cases, most things are just nice to have. You need a mic... it'd be nice to have one that sounds great. You need a preamp... it'd be nice to have one with low self noise. You need cables... it'd be nice to have ones that are durable and keep out radio interference. You get the picture. In my use case, I record a lot of self-directed character VO for various projects where dynamics range wildly. I spent many years keeping an eye on my levels - making sure to turn the mic down for louder lines and back up for softer parts and so forth- and it was fine I guess but It took a lot of mental juggling going in and out of character so much to ensure I didn't clip. What I like about having a Zoom F3 is I can just walk into my booth with my micro SD card and just act. No additional need for set up or watching my levels like a hawk. It's nice. For the rare times I have a session where I need to be directed live, I have a UA Volt that I run from my booth to my laptop and a splitter so I have a back up recording on the F3 in case anything goes wrong. It's a very specific use case and I can't really see it being as helpful for something like promo or commercial VO, but it's perfect for me and that's what matters.

[@julesc8054]

To be real some people really really should be using it. Particularly some of my clients who are photographers first and haven't quite grasped the concept if audio. For them this is a Godsend. It also allows competent cameramen to focus on the shots and not worry about "is the audio level ok? " Often the last question is only answered in post, when its too late.

[@DeKoepi]

I do have the Zoom UAC-232 and like it a lot. For usual meetings in Teams/Zoom, I used cheap 24bit Interfaces (and now changed to a Vocaster 1 because the preamps are much less noisy and really offer enough gain for the Podmic). So I know about the struggles getting the levels right. But the UAC-232 is for capturing my synthesizers. For mixing and layering the music audio, 32 bits float (that is important, now cheap chinese audio interfaces seem to offer 32 bits int which is pure BS) are very welcome because I don't need to think about being over the limits. The export task then calculates everything to be on the non-clipping side. For live streaming I see no advantage as you need to set some proper volume anyways to be audible.

[@thephotoelite]

I would love a full breakdown from you. You do this in a way that just tickles my brain in a good way!

[@DarkCornerStudios]

@thephotoelite awww...thanks for the kind words!

[@tomwatson283]

Excellent video again. The old stuff still works for me. I have the 32bit float gadgets, lot's of 'em, just in case. It's always good to try and keep up with tech before it passes us by in the age race. Just invested in the Rodecaster video (still on pre order, not got it yet). Despite listening to all the so-called users 4k "negatives". One has to eventually give what's on offer a go at the "normal" users price cap. Can't wait to take the leap into the unknown world of today, instead of tomorrow. That never comes, as - by the time it does come it's today again.

[@32bitSounds]

As a field recordist I use 32bit on every recording, it's handy for unpredictable increases in sound such as birds tweeting close to the mics, I can see how it might not be essential in a studio, but for outside recordings it's very handy.

[@An0n_Ym0us_79]

Great explanation. Thanks!

[@QueMusiQ]

4:30 yes.that but also micing when you’re in nature in general. You can record all the things across the dynamic range - a lion’s footfall AND roar is the common example. You have a wider berth for your gain stage. But as primarily a music guy? Higher bit depth is useful if like me, you have no formal training and lack the proper discernment to not use every fancy new plugin that can otherwise bring up your noise floor. Across multiple tracks that adds up and can effect the overall sonics especially once you consider the “self-mastering” stage. (a beat I’m mixing can have as few as 12 tracks or as many as 100+ depending on how ambitious the project.) So even when going for a 12 or 16 bit emulated sound or classic era samplers, it’s best to “emulate” it…at as high a bit depth and sample rate as your rig and signal chain will allow. For example, grab some excellent quality vintage emulation processors a la decimort or rx-20 or the sp emulation and resample it at 32bit (or just set the project to 32bit float) and as high a sample rate as possible without your system crashing. Why: Because that mathy float point stuff? When you then load it into Studio One or ProTools or some other primarily “mixing” system? You can gain stage without worrying about noise floor or too little dynamic depth. I can “record” at a lower level without being too concerned with clipping. (This is more important in some genres than other like metal, alternative, etc where it can jump from loud to quiet quickly, and with emcees who don’t really stand tf still in relation to the mic because of the nature of what we do, and the cadence and delivery requirements of spitting.) you can get a great take with as high a bit depth as your gear will allow, which to my knowledge can top out at 24bit with most rigs. But then it’s all the plugins. If you use well thought out vintage emulation a la Slate Digital or UAD’s offerings, you MUST - I repeat - MUST gain stage as though you are in an analogue environment for those VMR or VTM type vst’s to respond optimally. I aim for -12db peak across the board. The closer you get to that in the take (even when dealing with virtual instruments and drum samples out of a modern system from MPC or Maschine), the better. But if you’re starting out with 32bit sounds in the daw and bounce it at a higher rate in compliance with Nyquist (if your end master is gonna be CD quality of 44.1, set your project to double at 88.2; and if broadcast standards of 48khz, go 96). But if you come into the mix session at double the sample rate, and you import into a more robust mixing DAW, set that to the same so the internal math doesn’t get wonky via conversion. Then your effects and what not will behave as designed and are less destructive. Mathematically, you can only alter the sound by subtraction- that is, REMOVING bits. Anything you do to alter the sound is subtractive in the digital space, when it’s additive in analog like with 2” tape machines. The float point helps, and the noise floor is so low that it is not a problem no matter how much you use the multiple compressor mix philosophy. As a novice who didn’t know that, I can attest that this was ALWAYS A PROBLEM. Yes, even entirely ITB.

[@donbullock-techihound]

That’s the best explanation I’ve heard so far. Thanks.

[@DarkCornerStudios]

@@donbullock-techihound cheers! Glad I could break through for some people!!

[@HasanNiz]

I think for beginner KZbinrs with a studio and a fixed mic placement the 32bit workflow might get in the way, if the zoom levels are not set correctly they’ll need to use a DAW to set the levels in post, or else they’ll end up clipping if their video editor doesn’t support 32bit float audio and their recording level zoom (is that the correct term?) was set too high

[@DJPLAYNICE]

Thanks for the vid! 🤘🏼😎RAWK!

[@fakshen1973]

32 but float is storing higher levels of precision when calculating the result of each step of DSP. A plug in will have multiple steps of DSP. You may have several plug-ins in the path of just one audio track by the time it leaves the master bus. If at every step of DSP you're lopping of details, you lose accuracy. The float is there to preserve as much as possible and then only introduce the big lop-off at the very end when we reduce to our session bit depth. 33 but float isn't for recording. It's for processing.

[@steveschnetzler5471]

Which technology actually has the better dynamic range: condensor, dynamic or ribbon? Which of these might be able to take advantage of 32bit float ( when you set your gain too high or low)? Thanks

[@DarkCornerStudios]

@@steveschnetzler5471 easily condensers. Specifically for the really quiet signals

[@RobKristjansson]

🤔 Pro VO artist here. In general ribbon mics have the lowest sensitivity, with dynamic mics next, and condenser microphones having the highest. So a Neumann TLM103 is sensitive to incredibly low inputs. On the high end, well, that's a little more complicated. There is a reason that guitar cabinets are mic'd with Shure SM57s - they can handle the volume! But so can a lot of condenser mics, too. For example, for some video game and fictional drama projects I need to SCREAM into the mic, and my Earthworks ETHOS has more than enough dynamic range (145 dB SPL) to handle it. Fun fact, I can't scream louder than the dynamic range of my interface's preamps (113 dB) Ribbons are super sensitive to high dB SPL (sound pressure level) inputs, so you just want to be careful with them. Bottom line for mics - know what you're recording and check the microphone's dynamic range to be sure you will be able to hear the full range of your sound source(s). But even if you don't, the mic can probably handle it. But I think we've got to think about this differently. Are you NOW recording sources that will require two different gain settings because they are both incredibly quiet AND incredibly loud, and you can't change that gain between sounds? Then 32 bit float recording is your solution. Otherwise, put the effort into getting your gain right, or accept you're going to be spending more time in post tediously fixing it

[@BrentLeVasseur]

32bit float is like HDR in video. It’s a float depth which enables high dynamic range, which means you will never accidentally clip your audio and can normalize it in post without any signal clipping or losses. For someone who doesn’t care about leveling while recording it’s great. That’s why for these portable Zoom recorders, it makes it basically idiot proof as you don’t have to worry about your mic levels.

[@Bloor005]

Headroom or float, or there is another way - two mics with different gain levels. This is how Tina Turner was recorded in the 80s when she had her massive resurgence. I have heard of some voice actors also doing this so they can scream mid scene/take. Never used float myself. I find 24/96 perfect for anything I've personally ever done, and tbh I think my ageing laptop would crumble with 32bf.

[@DarkCornerStudios]

Yeah...there are definitely niche uses that make it unbelievably cool and necessary... Doubtful any of them are recording on to a sub $500 interface...lolz

[@titaniumben9923]

That's a good point. I have two Rode M5s and did this to record thunderstorms as the closest thunder clipped like hell. So used the 2nd mic on the RCP2 at far lower gain.

[@Bloor005]

@@titaniumben9923 Yeah a great reason to buy a matched pair imo.

[@fatherwoolley]

I want to buy a USB mic for livestreaming a country western band, with drums, keyboard, acoustic guitar, pedal steel bass and of course vocals, for under $200. I thought the zoom 32 bit float mics would be good, but I just heard you say that the 32 bit float doesn't help for streaming. I'd also like to just set it and forget it. I use ecamm live to livestream through Facebook live. Do you have any recommendations?

[@ww1www]

Is there any newly released device in 2024 that is not 32bit float? Maybe some USB microphones and smartphones. But if manufacturers don't want to shoot themselves in the foot, they need to add 32 bit to anything now. I still have some field recorders with 24bit, but I prefer the 32bit workfkow for my solo-shot videos. I still set gain (magnification level or digital trim) but 32bit float allows me to focus on the picture (cameras, lights) and my acting in front of the camera.

[@augiearvizo]

Why didn’t the Rodecaster Pro 2 & Duo get 32 Bit float? I guess if you are doing sound checks and monitoring maybe it’s just like camera folks preaching it’s better to get your exposure right before you hit the shutter/record? Same applies to audio?? 🎤🎙️

[@DarkCornerStudios]

Really the whole point to the unit is to monitor the signal. You have all the meters...and all the faders. 32bit float would be a bit unnecessary... Or at least it wouldnt be much of a marketable feature

[@DannyBrownPodcasts]

I'm still waiting for the company that releases an option to remove mouth clicks at recording stage.

[@mycosys]

There is not a single true 32float ADC in existence. '32Float' ADCs are actually a pair of 24bit ADCs at 2 different gain levels that change a gain mantissa as they switch. There is no analog gear with more than 23 bits of dynamic range, so 32 float is pointless anyway - a 32float DAC is simply an alternative to setting your gain properly, identical to using 2 channels with different gains.

[@eDrumsInANutshell]

I just learned that 32bit float has a dynamic range of 1500dB , that it weird. But there is also 32bit integer (192dB) and that will be enough I guess for the most semi pro guys. And after that fancy recording, folks just squeeze the audio through a tiny mono bluetooth speaker.

[@DarkCornerStudios]

@@eDrumsInANutshell essentially. Mix it down to 256kb/s too

[@tricogustrico]

Math is fun and not hard to understanad

[@DarkCornerStudios]

Wholly depends on the person. I am good to a point...but there is a point...lol

[@graysonpeddie]

I don't need 32-bit recording if I'm doing Teams/Zoom meetings.

[@johnrohdejensen1218]

I do video from a lot of small concerts and for me 32float is a lifesaver. I have to operate the camera(s) and don't have the time to fiddle with sound levels. Especially not for 4 mics and 2 line inputs. I still take some care not to overpower the recorder but that is it. My F6 takes care of everything else with 32float. Sweet.

[@RandumbTech]

Personally, I don't care much about the why. I am much more concerned about the does it actually work. And in my case, as a solo content creator, the answer is an overwhelming YESSS!!! I'll never buy another piece of gear without 32-bit float (or at least with a built-in digital limiter like the Sony hotshoe mics)

[@seanwebb605]

Blasphemy on this day when we should be giving thanks.

[@DarkCornerStudios]

@@seanwebb605 give thanks that you probably don't need 32bit float!!

[@seanwebb605]

@@DarkCornerStudios I will take every bit that is offered and be grateful for each of them.