I'm just learning everything about audio and video as I start my channel, and I gotta say, I learned a lot more than just 32 bit float by watching your video! Thank you
I switched over to 32 bit float a couple years ago and never looked back. It’s amazing for solo creators. But it’s true, proper mic placement and capsule protection from wind and plosives are MASSIVELY important.
This issue was expected when actual reviewers were replaced by sponsored hobbyists/influencers. They don't always have the required technical expertise, and focus on simply regurgitating a brand's marketing material. Criticism is also kept at a bare minimum, because otherwise they lose their privilege to future gadgets. This decline of real reviews is not just limited to cameras and audio equipment. It's everywhere and it's unfortunate. I do my best to support content creators who are not sponsored nor limited in their criticism.
Thank you, Wayne. Very well (and politely) stated. I'll be saving this video for my "camera" friends that have a very odd perspective of what 32bit float recording really means. Coming from you, they will probably be able to grasp the idea, whereas, if I tried to explain it, they would simply say, "Tom, you don't know what the Heck you're talking about"!
In finance, we mostly ignore social media commentary from large accounts, ironically it's the accounts with the least followers or activity that know what they're talking about. I also happened to be in audio engineering/computer science a decade ago - you're speaking pure facts here.
I know a few years ago I would have been misinformed by these claims (I have been before on camera stuff) but since I've followed you I've begun to understand sound a lot more. Of course school has helped, but my instructors haven't gone in depth like this or pointed out the scams, of which some of my fellow students fall for. Having a face associated with a brand is even a turn off for me because of how much I've been lied to, or have claims exaggerated upon. Anyway, my own personal ramble over, thanks for being transparent and LOUD about this stuff. You're no doubt saving people from being lied to, falling into the 'buy the next biggest thing' hole. Thank you, from all of us.
Throw the DJI Mic 2 and Rode Wireless Pro into the title. That would help with discovery of your video. Something like "32 bit float war (mic names) are here. (The rest)". Just my 2 cents. But the thumbnail was what caught my eyes. Thanks for the video :)
Thanks for getting out the soapbox. I recently purchased the (now discontinued) Deity Pocket Wireless Lav. Hearing the facts about 32bit has me feeling better about my purchase.
I love recording in 32-bit float but it's definitely not a necessity. In fact, maybe I'll have to make a video on how you don't actually need it... Great time to grab a Deity Pocket Wireless now that they were heavily discounted. The same with the 1st gen DJI MIC which now is probably a better purchase for some creators over the (presumably) much more expensive 2 model.
The number of bits for 24 bits sets the noise floor att about -144 dB, rather than allow different loud sound to be recorded. On can record sounds less than -144 dB and the lower sound gets it fades away into the noise! We can hear sound that i lower than a noise, through the noice! The noice has to be there for get rid of any quantisation distortion and trade the distortion to a set noise level instead! Noise is better choise sinse it sounds like a separate sound rather than trashing the sound like distortion do! Now in practice hardware like mic's or electronics are more noisy than the 24 bit system noise (dither noise) but it is nevertheless how 24 bits work!
My understanding is 32bit Float audio is somewhat like using Log profiles on a camera instead of standard colour where you can get closer to the highlights and shadows in recovery. Correct me if I am wrong but 32bit float will allow getting closer to the quieter and louder??
32bit floating point sampling has "technically" (mathematically) over 1,500 db of dynamic range. However, I have to wonder how much dynamic range even the very best dual-gain analog preamp circuits have today. Maybe 120 db? or 130 db?...maybe or 160 db at very best? Anybody know what the very best analog gain circuits are? (I doubt that Zoom, Tascam or DJI buy anything like that) Also,..most consumer analog mic capsule dynamic range tops out at 120-ish db?? (very best ones) Another thing to tell people is that common 16bit, 24bit and 32bit sampling words use integer math. 32 bit integer is "not" the same as 32 bit floating point math.
I’m hoping they figure that out somehow. Would make my life easier. The DJI Mic falls off all the time when I connected it to the Action 3 and 4. The camera can handle action, but the mic…not so much.
Very well said, i definitely agree with you. Ive seen a 32bit float video just earlier, because I needed a new audio interface and decided against it, as my audio setup (mostly vocal) won't benefit from the larger volume range... Especially if i can adjust the gain ahead of time appropriately
Hi Wayne thks for this, I agree with what you said. I’m started as a photographer n later film. It’s like in photography, you can have 100 megapixel cam but if you don’t light your subject well, you won’t get a good contest n colours even if you shoot raw and push it in post. 4K 8k 12k is not useful if you don’t light your video well, same goes for 32 bit audio. By the way I love the sound of your video here. Would be nice if you could share your setup and mic used . Thks for the great work 🙏
Thank you for the great info here! Can you help me to understand if the 32-bit float .wav files produced by the Zoom F3 are supported in post by LumaFusion?
Hi how are you? I do photography and video but I recently started digging into audio and I'm learning a lot from your channel, thanks for that! I would like to ask you if it is viable to combine a 24-bit system like the Hollyland Lark Max with the 32-bit Zoom F3 recorder. I would use a TRS to XLR adapter to connect the Lark Max receiver to the Zoom F3. Do I gain the advantage of 32bit float that way? Thanks for your help, greetings from Argentina!
Hi, thanks for your comment! You could definitely connect your Hollyland receiver to an F3 but you wouldn't necessarily get the benefit of 32-bit float as the wireless system is 24-bit. It's like writing on a Post-It and then gluing that Post-It to a sheet of paper. So you're better off using the F3 separately with a different mic connected directly to it.
@@WayneRECs Thanks for your advice, and what do you think about using an xlr levalier with some xrl wireles system? There are many now on the market, and then connecting the receiver to the F3? I wouldn't have a backup recording but everything would go directly to the F3. Can you recommend any accessible xlr transmitter/receiver with phantom power? Sennheiser has a system but it is too expensive. Do you know something about de Depusheng T9 system, is a “no name” brand I think. Thanks for your help!
4:48 it depends where sound source talking to. The man behind you may speak directly to your mic and it will be louder than man in the front of you who talks side way.
I ended up getting a 32-bit float capable tascam X8. I ended up using it with my DJI Mike version 1 to get 32-bit flow point audio and wireless transmission. Seems to be working for the most part but I have a bunch of recordings that sound horrible. It almost sounds as if AI was put on the file and turned all the way up. I'm thinking it might be better to just rerecord it but I wasn't sure
I’m more wanting to know about the Timecode use of this. I chatted with DJI Support and they responded that we can jam Timecode into this (me: Into the Pocket 3 with a USB-C to 3.5 Adapter or straight into the Mic 2? )
I used a Senheisser MKE 600 connected to a DJI MiC 2 (32bit float) to record people speaking on a loudspeaker system. The problem was a very low volume for the people speaking. I had to raise the gain hugely using pro tools. This then meant some sections registered as clipping when imported into Davinci. Not sure what was going on. Be great to see a workflow for dealing with low volume 32but recordings.
You can focus on audio in post processing. You need phase consistent microphone arrays to do it, similar to interferometry. HOAST from IEM has some demonstrations, and Steve Mould made a video on acoustic cameras. This has very little to do with your sample format; a Zoom H2n or H3-VR is the entry level. A more effective and traditional method is to use directional mics and good placement. You can use a parabolic reflector, for instance.
Hi Wayne, great info... my question is, in a noisy environment, what would be the best type of (portable) mic to use if you wanted to pickup more of the voice(s) and cut out ambient sound, and do you need a mic for each voice?
Is there a auto edit that levels or boost and lowers overall file ? I use fcp.. i recorded a karate session and there was a lot of people and spikes in volume. I used nr and it was crap. Yes mic placement is so important😊Thanks for the facts
What about samsung phone function that when you zoom video to record and the sound concentrates for that place that you are filming to greater sound? I tried, i couldnt hear from my place but when took the video to watch it was like standing next to that conversation, but was recorded from bigger distance and near was other noise..
I could be wrong, but my guess it is just picking certain mics to pick up sound from only that direction rejecting sound from other directions and adjusting the gain as needed. For example I know that if you go into pro video mode on my Galaxy S22, you can pick whether you want omnidirectional sound recording, rear, or front. In that case I believe it is just picking the right microphone for the right job.
Looks like a sweet compact multitrack 32-bit float recorder! ZOOM looks like they really tailored it to musicians and has a bunch of cool features for that purpose.
I think it depends on your use case scenario. If you're primary focus is on audio recording, the IMO the F3 would be better. But if it's a convenience/ back-up recording for video, these pocket recorders are easier to use.
Soft box. Really good light. Or if you have a cheap light as I do, I point the light towards the wall, giving me a softer light on my face without the glare.
float recording does for audio what HDR does for photography. Really, there's no reason not to use float32 for everything these days, but it will take a while for the industry to move on from integer formats.
Thanks for this video. I've been confused about 32bit. 16 bit is CD quality. When I got my first mbox2 back in the day i got into recording 24bit. It sounded slightly better to me but probably cause it was more info for me to mix. I never did anything above 48khz tho. Crazy UA-camrs vlogging about them going to Starbucks is 32bit these days and all the music channels are still 24bit. The new Scarlett 4th gen can only do up to 192kHz, 24-bit. What you said makes sense tho. No different that recording the same audio at a lower db on channel two as a backup if the main lav gets too loud. 32 bit can do 4 of those at once you say. Kinda cool.
You make a great point! Recording in 24-bit with a safety track will produce similar results as 32-bit float albeit just as 2 separate files instead of one. While I personally love 32-bit float, it's basically another redundancy tool to capture clean recordings.
Thanks for that, it makes sense to me to some degree now 👍 every day is a school day learning more, I love these helpful snippets of info and when you do your how to type things, as a audio beginner clueless learner it's appreciated 😀
As they say, the proof is in the pudding ....The DJI Mic 2, the first of anything I've had that was 32 bit float. Hold it a few inches or so from your mouth, makes you sound like you're doing a professional podcast. I've had some nice audio recorders, none of them came close to that. Most recorders make people sound like they're in a video, on tv, or radio ...like the TV in the other room I hear right now, my wife watching Survivor. Jeff sounds like he's on tv, not in the room. My wife and I did a DJI Mic 2 recording together earlier, played it back on my computer and listened to it from outside my closed office door. It sounded perfectly like there was actual people in my office talking.
The DJI Mic 2 says their mics are recording 32bit float. If I use this mic with my Osmo Action 4, does this mean that the audio file that is recorded with the video on the camera is recorded in 32bit float or is it only the audio file that is recorded on the transmitter?
32-bit float really only pertains to audio files, meaning the wav file recorded on the Mic 2 TX unit itself (provided 32bf is turned on in the settings). The Action 4 audio baked into the video file is recorded in 16-bit 48kHz. So to utilize the benefits of 32-bit float, it means recording back up audio on the TX, copying the WAV file onto your computer and the syncing it with your video file in post.
@@WayneRECs Can I ask, as an audio engineer do you recommend we use the noise cancelling built into the transmitter or would you recommend applying noise cancelling in post? The reason I ask is that I tried the noise cancelling feature and while it did do a good job, the audio sounded a bit distorted to me.
So….I’ve been thinking about this a lot and been experimenting with this concept…(which we can discuss next time we talk 😅) potentially it could be better quality because of all the extra range of sounds you otherwise may not have picked up… (i.e. asmr content)… I think when software technology gets better to separate all the different layers of sound that 32-bit float can capture…like a centrifuge…lol… then we can probably realize even more benefits than the limitations we have now. Ok now I’m rambling… 😂. Talk soon, bud. ✌️👍😁
It's exactly those layers it won't help you with, because the mantissa, the actual sample portion of the IEEE binary32 floating point value, is 24 bits (including the implied 1). And the data being put into those bits comes from some ADC with some gain input stage, just as we're used to. It won't have more precise data to put in unless you have analog stages with that precision. What it can do is substitute lower precision data from a different ADC when one is overwhelmed. A good float recorder is essentially two: One low gain backup track, so you have something when the high gain is clipping, and one higher gain that could be doing automatic gain. The advantage is that they should be matched up on a common scale and take 25% extra space instead of 100%. In reality, not all devices are ideal.
Say their names call them out ! Lol I'm just kidding. It's true though. Some people get products they aren't sure about simply because of their number of subs . Then they don't give the product the right amount of research and right amount of coverage because it's not something they believe in or care to really know about.
A very simple to answer in 30sec, Mic's range, audio IC, output file format, are they all matching to 32bit, it is not a real 32bit recorder if any part is not in 32bit.
Sadly: influencers can not be trusted DUE to the inducements they receive. You do have to research so much to get the fair answer and also hunt for the people who have experience, skill AND NO sponsorship.
32-bit recording is overrated. For decades, we didn't need it. Technician is capable to foresee 99% of situations. I don't say it is not needed, but there is just too much fuss over it.
You're sort of contradicting yourself. How do the manufacturers do it? They take the 24-bit output of the ADC and write it into 32-bit float. For most microphone audio interfaces don't be surprised if the ADC is working at 12-bits! Or even 10! Speaking bluntly, 32-bit float is a marketing scam ;) It's beyond trivial to take a 16 or 24-bit recording and convert it to 32-bits. I feel you needed to take a harder stand on this issue.
I obviously can’t check, but my Rode NT-1 I’m told includes 4 ADC’s, each covering a different window into the 32-bit float space. No idea if that’s completely accurate, nor what the effective bit width of each ADC is, but it is difficult to deny that the results of 32-bit recording for a complete amateur like me is exactly what I’d expect, i.e. it’s more difficult for me to screw things up in a way where I can’t recover good audio.
@@StefanHolmes The 4 ADC's "theory" of how they do it is fake science if you think about it. Can you take a 4-way splitter and input your mic into four audio interfaces, each with a different gain, and then combine them? No. Conservation of energy. Each voltage would be 1/4 strength, or 1/4 weaker. Okay, so you say, you pre-amp it first and then send to 4 ADCs. Okay. But what proof to you have that 4 ADC will resolve the signal better than 1? I've not seen that. IF it were really true you'd see it in the specs, rigtht? When you shop audio interfaces do you see a section for number of ADCs? No. IF you notice a difference between 32-bit float and 24-bit fixed it's because they've added a limiter or automatic gain control in the signal path. A 100 year old technology. And it compresses the signal so degrades the fidelity. 32-bit float is a ploy to get amateurs to spend money on this new gear that has it. My latest video explains some of the business and consumer realities why this myth perpetuates. 32-bit float isn't worse. No harm if you use it. But I wouldn't because it's just a waste of compute because you're resolve to fixed at the end anyway. The "float" just means more scale, which is exponential. 32-bit float is the SAME as 24-bit fixed in precision. You can look that up on wiki.
@@maxcanthelpit Do I think it's electrically possible that there can be multiple ADC's separated with a crossover? Yes I do. We do the same in reverse for speakers. To be fair, I'll reserve judgement until someone does a teardown / reverse-engineer of such a mic. (maybe someone did, I haven't looked) I'm aware of the mechanics of how 32-bit float versus 24-bit fixed works. I've been poking around in and understanding file formats for 30-ish years, including the formal IEEE definitions of floating point formats. 32-bits is a ridiculous amount of headroom. Simply unnecessary. Still a great format for transporting audio data, because any format that can far exceed requirements whilst still being "cheap" to use is the perfect format. What someone might not like about it is the fact they can't hide behind it. If you screwed up your audio, you're going to have to own up to it and you can't blame your gain settings or your audio interface (especially in cases where your "interface" is simply USBC straight into a laptop). As to fidelity being degraded because of tricks that make 32-bit "work" in a mic... I'm sure that's true, but I'm not sure that's universally true. I'm not going to evangelise Rode as being perfect, I have literally no experience of their brand except for this recent (weeks, not months or years ago) purchase of an NT-1. I think it would be risky to their reputation to make completely BS claims about the tech in this mic. I'm not one to take marketing at face value, but if Rode say "hey, the noise floor is X, we have Y number of ADC's and you can't clip audio that sits within the rated max SPL (132dB!) of the mic itself", I'm going to give them some benefit of some doubt. Maybe the noise floor is really X+4dB and the max usable SPL is less than 130dB (I'd expect it to be, honestly), but that still leaves me with a mic that, to my ears, does a good job whatever mode I put it in. That being said, mine gets used mainly as a standard mic via XLR into an inexpensive MOTU M4 for realitime audio and I probably have the gain set a bit too high and I should definitely put a compressor into the path somewhere, but I'm not a pro and I'm just incrementally figuring out what's what.
Thanks for the good explanation. Audio is clearly something you can. The big channels or camera channels have no idea what you are explaining here.
I'm just learning everything about audio and video as I start my channel, and I gotta say, I learned a lot more than just 32 bit float by watching your video! Thank you
I switched over to 32 bit float a couple years ago and never looked back. It’s amazing for solo creators. But it’s true, proper mic placement and capsule protection from wind and plosives are MASSIVELY important.
I know you have a tentacle sync, did you ever pick up an Instamic? Glad to hear that you're enjoying the Action 4!
All well said, Wayne. 👍
@@curtisjudd Thanks Curtis! I appreciate you stopping by!
This issue was expected when actual reviewers were replaced by sponsored hobbyists/influencers. They don't always have the required technical expertise, and focus on simply regurgitating a brand's marketing material. Criticism is also kept at a bare minimum, because otherwise they lose their privilege to future gadgets. This decline of real reviews is not just limited to cameras and audio equipment. It's everywhere and it's unfortunate. I do my best to support content creators who are not sponsored nor limited in their criticism.
Thank you, Wayne. Very well (and politely) stated. I'll be saving this video for my "camera" friends that have a very odd perspective of what 32bit float recording really means. Coming from you, they will probably be able to grasp the idea, whereas, if I tried to explain it, they would simply say, "Tom, you don't know what the Heck you're talking about"!
Lol, thanks Tom! Sounds like me trying to understand color grading!
Yoda: Begun, the 32-bit float wars have
Thanks! Great explanation of 32-bit Float.
In finance, we mostly ignore social media commentary from large accounts, ironically it's the accounts with the least followers or activity that know what they're talking about. I also happened to be in audio engineering/computer science a decade ago - you're speaking pure facts here.
So does it mean 32bit have more dynamics? And that is just one part of audio quality.
Yep, a few things always seem to be missed. 🙂
I know a few years ago I would have been misinformed by these claims (I have been before on camera stuff) but since I've followed you I've begun to understand sound a lot more. Of course school has helped, but my instructors haven't gone in depth like this or pointed out the scams, of which some of my fellow students fall for. Having a face associated with a brand is even a turn off for me because of how much I've been lied to, or have claims exaggerated upon. Anyway, my own personal ramble over, thanks for being transparent and LOUD about this stuff. You're no doubt saving people from being lied to, falling into the 'buy the next biggest thing' hole. Thank you, from all of us.
Just in case noone told you yet: THANK YOU!
Very good explanation, ive already watch creators promote it as the best thing for sound recording ever.
Thank you 🙏
You got a new subscriber from Norway
Throw the DJI Mic 2 and Rode Wireless Pro into the title. That would help with discovery of your video. Something like "32 bit float war (mic names) are here. (The rest)". Just my 2 cents. But the thumbnail was what caught my eyes. Thanks for the video :)
Thanks for getting out the soapbox. I recently purchased the (now discontinued) Deity Pocket Wireless Lav. Hearing the facts about 32bit has me feeling better about my purchase.
I love recording in 32-bit float but it's definitely not a necessity. In fact, maybe I'll have to make a video on how you don't actually need it... Great time to grab a Deity Pocket Wireless now that they were heavily discounted. The same with the 1st gen DJI MIC which now is probably a better purchase for some creators over the (presumably) much more expensive 2 model.
Hi what’s the difference between rode wireless pro 32bit recoded and zoom f3?thx
Great info, thank you. I couldn't quite understand how the hardware side would work with 32bit float. Now I am a little closer to getting the picture
Float wars? Makes me conjure up naval battles 🤷♂️ 🤣 ✌️🤙
C2... You sunk my Battleship! Lol
@@WayneRECs hahaha 👍 touché 😆
The number of bits for 24 bits sets the noise floor att about -144 dB, rather than allow different loud sound to be recorded. On can record sounds less than -144 dB and the lower sound gets it fades away into the noise! We can hear sound that i lower than a noise, through the noice! The noice has to be there for get rid of any quantisation distortion and trade the distortion to a set noise level instead! Noise is better choise sinse it sounds like a separate sound rather than trashing the sound like distortion do! Now in practice hardware like mic's or electronics are more noisy than the 24 bit system noise (dither noise) but it is nevertheless how 24 bits work!
Gasp! Content creators disseminating missinformation! Seriously, thanks for this info, it's helpful. :)
Excellent. Thank you for sharing your expertise. Its nice to find a channel that’s authentic and isn’t influencer content.
Thank you so much for your clarification. I now have a better understanding of the concept.
My understanding is 32bit Float audio is somewhat like using Log profiles on a camera instead of standard colour where you can get closer to the highlights and shadows in recovery. Correct me if I am wrong but 32bit float will allow getting closer to the quieter and louder??
who knew this was such a spicy topic! thanks for keeping it real 🌶️
32bit floating point sampling has "technically" (mathematically) over 1,500 db of dynamic range. However, I have to wonder how much dynamic range even the very best dual-gain analog preamp circuits have today. Maybe 120 db? or 130 db?...maybe or 160 db at very best? Anybody know what the very best analog gain circuits are? (I doubt that Zoom, Tascam or DJI buy anything like that) Also,..most consumer analog mic capsule dynamic range tops out at 120-ish db?? (very best ones) Another thing to tell people is that common 16bit, 24bit and 32bit sampling words use integer math. 32 bit integer is "not" the same as 32 bit floating point math.
Thank you sir!! Best explanation and practical advice on 32 bit float I've heard!
Thanks. Il seems thé New limit is thé analog stage especially the preamp.
Why can't I find the DJI Mic 2 for sale?
I'm hoping that the Action 4 has a secret hidden inside that will allow it to work with the new DJI mic 2. No receiver or wires would be great.
That would be amazing!
I’m hoping they figure that out somehow. Would make my life easier. The DJI Mic falls off all the time when I connected it to the Action 3 and 4. The camera can handle action, but the mic…not so much.
Very well said, i definitely agree with you. Ive seen a 32bit float video just earlier, because I needed a new audio interface and decided against it, as my audio setup (mostly vocal) won't benefit from the larger volume range... Especially if i can adjust the gain ahead of time appropriately
i like your videos, honest and real.
Hi Wayne thks for this, I agree with what you said. I’m started as a photographer n later film. It’s like in photography, you can have 100 megapixel cam but if you don’t light your subject well, you won’t get a good contest n colours even if you shoot raw and push it in post. 4K 8k 12k is not useful if you don’t light your video well, same goes for 32 bit audio. By the way I love the sound of your video here. Would be nice if you could share your setup and mic used . Thks for the great work 🙏
Thank you for the great info here! Can you help me to understand if the 32-bit float .wav files produced by the Zoom F3 are supported in post by LumaFusion?
Hi how are you? I do photography and video but I recently started digging into audio and I'm learning a lot from your channel, thanks for that! I would like to ask you if it is viable to combine a 24-bit system like the Hollyland Lark Max with the 32-bit Zoom F3 recorder. I would use a TRS to XLR adapter to connect the Lark Max receiver to the Zoom F3. Do I gain the advantage of 32bit float that way? Thanks for your help, greetings from Argentina!
Hi, thanks for your comment! You could definitely connect your Hollyland receiver to an F3 but you wouldn't necessarily get the benefit of 32-bit float as the wireless system is 24-bit. It's like writing on a Post-It and then gluing that Post-It to a sheet of paper. So you're better off using the F3 separately with a different mic connected directly to it.
@@WayneRECs Thanks for your advice, and what do you think about using an xlr levalier with some xrl wireles system? There are many now on the market, and then connecting the receiver to the F3? I wouldn't have a backup recording but everything would go directly to the F3. Can you recommend any accessible xlr transmitter/receiver with phantom power? Sennheiser has a system but it is too expensive. Do you know something about de Depusheng T9 system, is a “no name” brand I think. Thanks for your help!
What setup would you recommend for cycle vlogging? My budget is under $500. It can be wired or wireless, and my biggest issue will be wind noise
4:48 it depends where sound source talking to. The man behind you may speak directly to your mic and it will be louder than man in the front of you who talks side way.
I ended up getting a 32-bit float capable tascam X8. I ended up using it with my DJI Mike version 1 to get 32-bit flow point audio and wireless transmission. Seems to be working for the most part but I have a bunch of recordings that sound horrible. It almost sounds as if AI was put on the file and turned all the way up. I'm thinking it might be better to just rerecord it but I wasn't sure
I’m more wanting to know about the Timecode use of this.
I chatted with DJI Support and they responded that we can jam Timecode into this (me: Into the Pocket 3 with a USB-C to 3.5 Adapter or straight into the Mic 2? )
What's your recommendation...
Dji mic 2 or hollyland lark max?
I used a Senheisser MKE 600 connected to a DJI MiC 2 (32bit float) to record people speaking on a loudspeaker system. The problem was a very low volume for the people speaking. I had to raise the gain hugely using pro tools. This then meant some sections registered as clipping when imported into Davinci. Not sure what was going on. Be great to see a workflow for dealing with low volume 32but recordings.
Hello my friend. Thanks for the video. The link to your gears doesn’t work on my side. Can you help me ?
Great story. Thnx. 🙏🏿❤️
Do you have an opinion about DJI mic 2 vs Røde wireless pro? 😅
You can focus on audio in post processing. You need phase consistent microphone arrays to do it, similar to interferometry. HOAST from IEM has some demonstrations, and Steve Mould made a video on acoustic cameras. This has very little to do with your sample format; a Zoom H2n or H3-VR is the entry level.
A more effective and traditional method is to use directional mics and good placement. You can use a parabolic reflector, for instance.
What gain level do you prefer?
I use the Spinal Tap wireless mic... 33-bit float. It goes to 33...
Hi Wayne, great info... my question is, in a noisy environment, what would be the best type of (portable) mic to use if you wanted to pickup more of the voice(s) and cut out ambient sound, and do you need a mic for each voice?
Does that 32 bit float audio need a more powerful computer too edit or process the file?
Is there a auto edit that levels or boost and lowers overall file ? I use fcp.. i recorded a karate session and there was a lot of people and spikes in volume. I used nr and it was crap. Yes mic placement is so important😊Thanks for the facts
Can you still use the 32-bit-float option in Osmo pocket 3 with rode wireless mic?
Rode (Pro) records 32 float internally in the transmitter
What about samsung phone function that when you zoom video to record and the sound concentrates for that place that you are filming to greater sound? I tried, i couldnt hear from my place but when took the video to watch it was like standing next to that conversation, but was recorded from bigger distance and near was other noise..
can't zoom in with audio... bogus marketing
I could be wrong, but my guess it is just picking certain mics to pick up sound from only that direction rejecting sound from other directions and adjusting the gain as needed. For example I know that if you go into pro video mode on my Galaxy S22, you can pick whether you want omnidirectional sound recording, rear, or front. In that case I believe it is just picking the right microphone for the right job.
dude what do you think about the Zoom R4? insane little piece of 32 bit float
Looks like a sweet compact multitrack 32-bit float recorder! ZOOM looks like they really tailored it to musicians and has a bunch of cool features for that purpose.
I know it’s a different item but would you say that the zoom f3 is a better choice (if you have mics etc)
I think it depends on your use case scenario. If you're primary focus is on audio recording, the IMO the F3 would be better. But if it's a convenience/ back-up recording for video, these pocket recorders are easier to use.
You can clip a lav mic
Everyone needs to remember…
Marketing departments are the proverbially tail wagging the dog.
I am always intrigued by how you are able to light your videos without glare on your glasses???
Soft box. Really good light. Or if you have a cheap light as I do, I point the light towards the wall, giving me a softer light on my face without the glare.
It’s the positioning of the light source that reduces glare
float recording does for audio what HDR does for photography. Really, there's no reason not to use float32 for everything these days, but it will take a while for the industry to move on from integer formats.
Thanks for this video. I've been confused about 32bit. 16 bit is CD quality. When I got my first mbox2 back in the day i got into recording 24bit. It sounded slightly better to me but probably cause it was more info for me to mix. I never did anything above 48khz tho. Crazy UA-camrs vlogging about them going to Starbucks is 32bit these days and all the music channels are still 24bit. The new Scarlett 4th gen can only do up to 192kHz, 24-bit. What you said makes sense tho. No different that recording the same audio at a lower db on channel two as a backup if the main lav gets too loud. 32 bit can do 4 of those at once you say. Kinda cool.
You make a great point! Recording in 24-bit with a safety track will produce similar results as 32-bit float albeit just as 2 separate files instead of one. While I personally love 32-bit float, it's basically another redundancy tool to capture clean recordings.
Thanks for that, it makes sense to me to some degree now 👍 every day is a school day learning more, I love these helpful snippets of info and when you do your how to type things, as a audio beginner clueless learner it's appreciated 😀
I still record 24bit. So large s/n floor ratio.
Please DJI give us time code syncing!
As they say, the proof is in the pudding ....The DJI Mic 2, the first of anything I've had that was 32 bit float. Hold it a few inches or so from your mouth, makes you sound like
you're doing a professional podcast. I've had some nice audio recorders, none of them came close to that. Most recorders make people sound like they're in a video,
on tv, or radio ...like the TV in the other room I hear right now, my wife watching Survivor. Jeff sounds like he's on tv, not in the room. My wife and I did a DJI Mic 2 recording
together earlier, played it back on my computer and listened to it from outside my closed office door. It sounded perfectly like there was actual people in my office talking.
24 bit can do 320 audio format
The DJI Mic 2 says their mics are recording 32bit float. If I use this mic with my Osmo Action 4, does this mean that the audio file that is recorded with the video on the camera is recorded in 32bit float or is it only the audio file that is recorded on the transmitter?
32-bit float really only pertains to audio files, meaning the wav file recorded on the Mic 2 TX unit itself (provided 32bf is turned on in the settings). The Action 4 audio baked into the video file is recorded in 16-bit 48kHz. So to utilize the benefits of 32-bit float, it means recording back up audio on the TX, copying the WAV file onto your computer and the syncing it with your video file in post.
@@WayneRECs Awesome, thanks for the info.
@@WayneRECs Can I ask, as an audio engineer do you recommend we use the noise cancelling built into the transmitter or would you recommend applying noise cancelling in post? The reason I ask is that I tried the noise cancelling feature and while it did do a good job, the audio sounded a bit distorted to me.
Someone find me a butterfly!
So….I’ve been thinking about this a lot and been experimenting with this concept…(which we can discuss next time we talk 😅) potentially it could be better quality because of all the extra range of sounds you otherwise may not have picked up… (i.e. asmr content)… I think when software technology gets better to separate all the different layers of sound that 32-bit float can capture…like a centrifuge…lol… then we can probably realize even more benefits than the limitations we have now.
Ok now I’m rambling… 😂. Talk soon, bud. ✌️👍😁
It's exactly those layers it won't help you with, because the mantissa, the actual sample portion of the IEEE binary32 floating point value, is 24 bits (including the implied 1). And the data being put into those bits comes from some ADC with some gain input stage, just as we're used to. It won't have more precise data to put in unless you have analog stages with that precision. What it can do is substitute lower precision data from a different ADC when one is overwhelmed.
A good float recorder is essentially two: One low gain backup track, so you have something when the high gain is clipping, and one higher gain that could be doing automatic gain. The advantage is that they should be matched up on a common scale and take 25% extra space instead of 100%. In reality, not all devices are ideal.
Loud 24 bit load 32bit not so loud so why not 32 bit audio system as 40 bit float
"i like sound"
- My ears
thanks
Well done
If they only know about SPL and Proximity Effect. And, I’m a hybrid shooter. 😅
Say their names call them out ! Lol I'm just kidding. It's true though. Some people get products they aren't sure about simply because of their number of subs . Then they don't give the product the right amount of research and right amount of coverage because it's not something they believe in or care to really know about.
Dude, you’re doing what you accuse others of doing.
Can you please expand? I'm curious.
i love your vids,
But for 99.5% of us.. 32bit? who cares???
A very simple to answer in 30sec, Mic's range, audio IC, output file format, are they all matching to 32bit, it is not a real 32bit recorder if any part is not in 32bit.
Mobile not that good sound as loudness only if in frequency won't be ropey
Sadly: influencers can not be trusted DUE to the inducements they receive. You do have to research so much to get the fair answer and also hunt for the people who have experience, skill AND NO sponsorship.
You need to watch a colour grading tutorial because your entire video is sinking in green.
rather strange point of view, i'll disagree with most of it, i think you need to get some experience with 32bit float recording first.
32-bit recording is overrated. For decades, we didn't need it. Technician is capable to foresee 99% of situations. I don't say it is not needed, but there is just too much fuss over it.
64 bit float is better lol
The only 64 I like starts with an N lol
@@WayneRECsC
You're sort of contradicting yourself. How do the manufacturers do it? They take the 24-bit output of the ADC and write it into 32-bit float. For most microphone audio interfaces don't be surprised if the ADC is working at 12-bits! Or even 10! Speaking bluntly, 32-bit float is a marketing scam ;) It's beyond trivial to take a 16 or 24-bit recording and convert it to 32-bits. I feel you needed to take a harder stand on this issue.
I obviously can’t check, but my Rode NT-1 I’m told includes 4 ADC’s, each covering a different window into the 32-bit float space.
No idea if that’s completely accurate, nor what the effective bit width of each ADC is, but it is difficult to deny that the results of 32-bit recording for a complete amateur like me is exactly what I’d expect, i.e. it’s more difficult for me to screw things up in a way where I can’t recover good audio.
@@StefanHolmes The 4 ADC's "theory" of how they do it is fake science if you think about it. Can you take a 4-way splitter and input your mic into four audio interfaces, each with a different gain, and then combine them? No. Conservation of energy. Each voltage would be 1/4 strength, or 1/4 weaker. Okay, so you say, you pre-amp it first and then send to 4 ADCs. Okay. But what proof to you have that 4 ADC will resolve the signal better than 1? I've not seen that. IF it were really true you'd see it in the specs, rigtht? When you shop audio interfaces do you see a section for number of ADCs? No.
IF you notice a difference between 32-bit float and 24-bit fixed it's because they've added a limiter or automatic gain control in the signal path. A 100 year old technology. And it compresses the signal so degrades the fidelity. 32-bit float is a ploy to get amateurs to spend money on this new gear that has it. My latest video explains some of the business and consumer realities why this myth perpetuates. 32-bit float isn't worse. No harm if you use it. But I wouldn't because it's just a waste of compute because you're resolve to fixed at the end anyway. The "float" just means more scale, which is exponential. 32-bit float is the SAME as 24-bit fixed in precision. You can look that up on wiki.
@@maxcanthelpit Do I think it's electrically possible that there can be multiple ADC's separated with a crossover? Yes I do. We do the same in reverse for speakers. To be fair, I'll reserve judgement until someone does a teardown / reverse-engineer of such a mic. (maybe someone did, I haven't looked)
I'm aware of the mechanics of how 32-bit float versus 24-bit fixed works. I've been poking around in and understanding file formats for 30-ish years, including the formal IEEE definitions of floating point formats. 32-bits is a ridiculous amount of headroom. Simply unnecessary. Still a great format for transporting audio data, because any format that can far exceed requirements whilst still being "cheap" to use is the perfect format.
What someone might not like about it is the fact they can't hide behind it. If you screwed up your audio, you're going to have to own up to it and you can't blame your gain settings or your audio interface (especially in cases where your "interface" is simply USBC straight into a laptop).
As to fidelity being degraded because of tricks that make 32-bit "work" in a mic... I'm sure that's true, but I'm not sure that's universally true. I'm not going to evangelise Rode as being perfect, I have literally no experience of their brand except for this recent (weeks, not months or years ago) purchase of an NT-1. I think it would be risky to their reputation to make completely BS claims about the tech in this mic. I'm not one to take marketing at face value, but if Rode say "hey, the noise floor is X, we have Y number of ADC's and you can't clip audio that sits within the rated max SPL (132dB!) of the mic itself", I'm going to give them some benefit of some doubt. Maybe the noise floor is really X+4dB and the max usable SPL is less than 130dB (I'd expect it to be, honestly), but that still leaves me with a mic that, to my ears, does a good job whatever mode I put it in.
That being said, mine gets used mainly as a standard mic via XLR into an inexpensive MOTU M4 for realitime audio and I probably have the gain set a bit too high and I should definitely put a compressor into the path somewhere, but I'm not a pro and I'm just incrementally figuring out what's what.
just mumbling, he has no idea.