And now... the Zoom F6!

[The Elf]

Maybe penny is falling...

That knob is an 'offset' (or multiplier?) to the mantissa, and that offset is stored along with the data?

Precissement!

By jove, I think he's got it! Clever eh?

Ah dun thank I got mah mind right naw bowss!

[Hugh Robjohns]

Thinking about it... it might have been easier if I'd said it was much like your DAW fader.

If you close a fader on a track in one mix pass and save the project, you can always go back and set the fader to some other value in a later mix pass if you want without affecting the actual stored audio.

Same thing in the F6 because it's using the same idea and technology.

H

[jimjazzdad]

Hugh:
Is this dual AD approach used for the F6 any different than the dual AD designs of the Sonosax SX-R4+ or the Zaxcom recorders with "Neverclip"?

[Hugh Robjohns]

Yes, as I said earlier, there are other systems that do very similar things. The R4 is certainly one. I've not examined the Zaxcom design yet.

H

[wireman]

That's how I understand it — it's (a bit) analogous to a digital camera that stores JPG and RAW formats of every photo you take (the JPG here is the file with level data embedded, the RAW is the unaltered wide-dynamic-range version, which you can then render to show data within the required 'window').

(...I think...?)

Cheers!

Chris

More like the standard HDR workflow of taking multiple images at different exposures.

I presume that Zoom are aware of Zaxconn patents (8,878,708)

BRIEF SUMMARY OF THE INVENTION

Briefly stated, in one aspect of the present invention, a method of extending the dynamic range of a system analog-to-digital conversion system is provided. This method includes the steps of: receiving an analog signal; amplifying the analog signal via at least one amplifier; converting the analog signal amplified by a first of the at least one amplifiers into a first digital signal via a first analog-to-digital conversion system; converting the analog signal unamplified, or amplified via a second of the at least one amplifiers, into a second digital signal via a second analog-to-digital conversion system; receiving a first sample from the first digital signal; receiving a second sample from the second digital signal; applying a current calculated gain value to the first sample to create a calculated sample; outputting the calculated sample as an output; automatically switching the output from the calculated sample to the second sample if the calculated sample exceeds a first setpoint; activating a timer to countdown from a maximum value; and automatically switching the output from the second sample to the calculated sample upon the expiration of the timer.

[The Elf]

[Hugh Robjohns]

I presume that Zoom are aware of Zaxconn patents ...

I presume Zaxcom are aware of Neumann's patents...

https://patents.justia.com/patent/6707402

Method for converting an analog signal to a digital signal
Patent number: 6707402
Abstract: With a method for converting an analog input signal (S1) to a digital output signal (S7), the analog input signal (S1) is amplified in a first signal path (2, 3, 5) and is subjected to an analog-to-digital conversion (5). In a second signal path (4, 6), another analog signal (S4) is obtained for transmitting larger signal amplitudes and is subjected to an analog-to-digital conversion (6). The signal (S5) digitized in the first signal path (2, 3, 5) and the signal (S6) that is digitized in the second signal path (4, 6) are fed to a digital signal processor (7), which generates the digital output signal (S7). To avoid an abrupt reduction in the signal resolution and achieve the highest possible dynamic scope, it is suggested that the analog signal fed to the second signal path (4, 5) be distorted non-linear and counter to the amplified analog signal (S2) in the first signal path (2, 3, 5).
Type: Grant
Filed: August 6, 2002
Date of Patent: March 16, 2004
Assignee: Georg Neumann GmbH
Inventor: Otmar Kern

[Mike Stranks]

As you know, the Chinese are aware of everybody's patents... and couldn't care less!

[Mike Stranks]

... back to the F6...

I did have a mental question mark about what's obviously a mock-up in a glass case at NAB, but with more info due in May and product launch in June. I would have thought that at least some pre-production prototypes should have been available for 'show and tell' at something as prestigious as NAB.

I recall the significant 'oops' that Zoom had with the original H2. All the prepub said that it would be a 5.1 recorder, but the execution proved more difficult than the aspiration. In the end, after a few months of delay, it hit the streets in a simplified form. 5.1 recording had to wait a year or so for the Mk2... Let us hope that Zoom have not hit similar issues with the tech in the F6...

[fingers-crossed]

[Hugh Robjohns]

I would have thought that at least some pre-production prototypes should have been available for 'show and tell' at something as prestigious as NAB.

It might have been the marketing/manufacturing target, but R&D can be a fickle thing, and maybe they just couldn't reach a pre-production working prototype In time. It happens.

But NAB is such an important show in that market sector that they would definitely have felt the need to announce the product there, even if it takes another year to bring the thing to market.

I've just reviewed the Radial kL-8 keyboard mixer which was announced and shown as an empty case at NAMM 2018, but didn't fall off the production line as a working it until January 2019 a whole year later!

[awjoe]

"If you're recording in 32-bit, you don't need to set gain at all."

My experience until now has been that Zoom satisfy. Now Zoom amaze.

[wireman]

"If you're recording in 32-bit, you don't need to set gain at all."

My experience until now has been that Zoom satisfy. Now Zoom amaze.

I couldn't resist a segue to Fat Larry on Youtuble...

[Hugh Robjohns]

"If you're recording in 32-bit, you don't need to set gain at all."

The whole reason that modern, stock, digital audio interfaces operate at 24 bit was for the exact same reason. A 24 bit digital system has a theoretical dynamic range of 141dB which is more than the human ear. In other words the dither noise floor can be set below the threshold of hearing, with the clipping level still well above the pain threshold. So it should be possible to have replay systems that don't need a volume control, and just reproduce everything at the same SPL as the original source.

In practice this is, of course, completely unworkable in a domestic context, but the idea is still valid for situations where an accurate calibrated replay level is desirable.

For high-end post-production it is necessary to have a wider dynamic range to allow some wriggle room -- in the same way that pro studios in the 70s and 80s used half-inch master tapes while the consumer used cassettes... Hence the attractiveness of a 32-bit float format in next-generation recorders.

[Jadoube]

Good Thread! I learned a few interesting bits. Bravo

[wireman]

Hence the attractiveness of a 32-bit float format in next-generation recorders.

I always assumed (no evidence) that audio software would be using 32-bit types to hold the 24 bit samples in any case so not much of a step to go to a 32-bit float. This should be faster as well.

[IronFilm]

Clever stuff! Really hope this idea drops down to an H4/5 format.

I'd imagine that Zoom will update their H series eventually. As their F series pre amps are MUCH better than in the H series.

[Mike Stranks]

A good first-look review. Worth viewing for it's explanation of gain-staging and the demo of the incredible dynamic range that the F6 can handle with ease.

https://www.youtube.com/watch?v=y4oNd1RgGL0

[Hugh Robjohns]

It's an ugly brick of a thing isn't it? And that power button looks both awkwardly and vulnerably placed -- although that's a complaint that's often levied at manufacturers and usually turns out to be unfounded... so I'll wait til I get my hands on a real one before forming a proper view.

The recording tech is clever, and will make life much easier -- and deliver more usable recordings -- for amateurs who don't understand gain staging. But it does make a lot more more work in post-production of course to wrangle sensible levels out of the 32-bit float file.

So in the pro-world it will be a great thing for assistant producers doing single-camera fly-on-the-wall doco-shoots without a proper crew....and it's another nail in the coffin of professional sound recordists. Just have to hope they point the mic the right way...

Clearly very feature-rich, which is good, but it's tiny... tiny... tiny.... and looks pretty fiddly to use.

[The Elf]

I'm blown away by the A/D technology, but this looks more like a concept design than anything that would make it into production.

Half an hour up on the Yorkshire moors and there's not a set of fingers in the world that will be able to manage those knobs and buttons!

Watching with interest...

[Hugh Robjohns]

Pretty sure that is as close as makes no difference to the full production unit.

Thankfully, the only buttons and knobs you need to fiddle with are the record and stop buttons. Assuming it's been set up sensibly, everything else will take care of itself.... or so they claim!

And clever though the tech is, I doubt the performance would be significantly different in practice from just recording with a generous headroom margin on a current 24 bit machine. The overwhelming element that defines the recorded noise floor will be the acoustic noise into the mic, not the preamp/converter noise -- and by a massive margin.

When recording live action stuff with very wide and unpredictable dynamic ranges and high peak SPLs, I often just set the preamp gain very low so that the ambient noise level is around -60dBFS or so on my Nagra VI (or even lower if the expected peak noise is nearing painfully loud), and then let whatever happens, happen... Seems to get the job done with no fuss, no overloads, and no noise problems in post-production.

Who needs 32-bit float?

H

[miN2]

I'm still impressed by this thing, maybe that's down to my ignorance and lack of experience of hardware mind.

Would it be possible to isolate each converter individually to a seperate audio file? This would allow capturing the quiet sounds in a loud environment?

[Hugh Robjohns]

No.

[ConcertinaChap]

Hope you don't mind me resurrecting this thread, the reason being I'm seriously considering getting an F6. I think I've got how this no gain setting business works in recording and it could be very useful to me. I'm involved in a project at a folk festival next year to record substantial parts of it and my particular corner would likely be sessions. These can have quite large and unpredictable dynamic variations over quite short periods. What I'm wondering is what happens when you try and move your recording to a DAW for further editing. For a 24 bit file where I've got the preamp gain right it's just a simple file import, but for the 32 bit file how do you choose the level for the import? Am I even explaining that clearly??

CC

[Mike Stranks]

Hi CC

Have you looked at the video I referenced above (Post date: 27 May)? Does that cover what you need to know? My impression is that you drag 'n' drop the file(s) from the recorder into your DAW and then use gain increase/normalise to 'see' the file audio level.

[ConcertinaChap]

I'll give it another shufty, it was months ago I viewed it. Other reasons for looking at the F6 include its size. Recording a session you need to be relatively inconspicuous and it occurs to me the combination of an ambisonic mic like my NT-SF1 and the F6 could make a pretty potent recording combination in a small package.

CC