Zoom Product Specialist, Samuel Greene, explains the benefits of using 32-bit float recording technology. We’re Zoom. And We’re For Creators. Zoom #32BitRecording #F3FieldRecorder
32 bit audio recording was already covered by reviewers on RU-vid, but i'm glad zoom actually took their time to explained themselves. And no reviewer touched the fact that you can actually record at 3 different audio levels.
I appreciate that Zoom here has set out to give a quick summary, but it's really dangerous to declare that "clipping is impossible," as if some kind of universal insurance policy. Many if not most users are unaware that clipping frequently happens at the microphone level, no matter what happens further down the chain at the 32-bit float recorder. Here's a great example that hits them close to home: one of Zoom's few other 32-bit float products, the F2, clips pretty easily when the lavalier included with the F2 is at the normal location below one's neck, but, say, someone sings at full volume. The maximum SPL of the lavalier simply can't handle it. The 32-bit float recorder faithfully captures the unusable clipping.
For your roadmap, please bring a 32bit version of the Zoom H3-VR and an H6 version. Since I started recording my samples in 32bit, the processing in RX10 has become much easier and the sound for my customers better and more flexible. So a "ready-to-run" Ambisonic recorder in 32bit would be great!
I noticed many reviewers said the wave magnifier did not effect recording level only headphone volume but it does according to this video. Good to know.
The headphone volume on the bottom of the device does not raise recording volume, but the volume buttons on the front do change the levels as long as you change it before hitting record. The levels are locked once recording.
Ok - Please clarify at 01:15 you used the + to lift the volume in your headphones and increase the gain of the recording. However, others say by lifting the gain in the headphones in no way lifts the input gain, just the monitoring. Please can you clarify. Thanks, Peter.
The headphone volume on the bottom of the device does not raise recording volume, but the volume buttons on the face of it does change the levels as long as you change it before hitting record. Once you hit record, the level is locked.
I think so. Different sample rates and bit depth won't change the result because of how the gates in the converters are designed. You won't record the issue at 48kHz and below but at 96kHz at either 24 or 32 bit there's a chance it will happen, especially with high-frequency transient sounds. Kay Paquin's blog has a very in-depth review of this issue. It happens on Sound Devices Mix Pre as well.. @@LakeSuperiorPhoto
Can the device actually record more than 140 dB of dynamic range? Conventional wisdom says that is impossible. The best converters cap out at 120 dB. Has that changed? A custom 16-bit floating point format could be used as a means of compression to extend the recording time.
That's amazing info - thanks for sharing!! Where did you find the info about converters capping out at 120dB? Cheers! To add onto your comment, I feel like the claim of "recording more than 140 dB of dynamic range" refers to working at 24 bit depth, which has a range of 144dB. 32 bit would offer a larger range, but then the next question would be... what microphones could faithfully capture more than that? And what human beings would be able to notice any amplitude changes outside of the normal threshold of hearing? 😆
"32 bit float had a much larger dynamic range than 24 bit" What is your definition of dynamic range? Using the definition I'm familiar with, something like "the difference in decibels between silence and maximum spl" for a microphone, the bit depth has nothing to do with dynamic range. I guess maybe you are meaning "USABLE dynamic range" because that would make some sense to me, having more usability on the quiet end of the range.
They are using two DA converters in tandem. A high gain and low gain. The unit senses the input to determine the optimal DA converter. I assume the whole thing uses "look ahead" to do so. So basically you are getting the dynamic range of two stacked and slightly over lapping DACs. So yes you would have more dBs between loud and silence.
