Topic: My adventures with a Zoom F3, Zoom H1 and other gear [tests and analysis]

[Rairun]

For most of my taping 'career', I've used a little Zoom H1. I bought it back in 2011, after I started recording video of a song or two at shows as a souvenir. My camera's microphones couldn't handle the high SPL, so the H1 felt like a huge upgrade then.

But after a few years and a couple of attempts at taping complete shows, I grew tired of the internal mics. In 2015, I bought a pair of CA-11 cards and an old STC-9000 preamp. Several people told me to upgrade to the Sony PCM M10, but I didn't have a lot of cash, so I figured the H1 would do. And it did serve me very well! I learned how to best compensate for its shortcomings and made a lot of good recordings.

Essentially, to get good recordings with the Zoom H1 (and afaik the H1n, which uses the same preamps), you need to think of it as only having two gain settings instead of a 0-100 scale:

0-15: Digital attenuation. Never use.
16: 0 dB gain. This is one of the usable gain settings
16-36: Steps of +0.4 dB gain, which increase the noise floor by 0.4 dB each. Essentially, it appears to be level 16 with digital gain added on top. Never use.
37: +13.5 dB gain. This is the second usable gain setting. This is where a cleaner high-gain preamp is engaged, and there is a noticeable drop in hiss from levels 36 to 37.
38-100: Steps of +0.4 dB gain, which like the previous 'ladder' appears to be level 37 with digital gain added on top. Never use.

I was happy with the Zoom H1 for a long time, but I found myself recording shows with a very high dynamic range, and the quiet parts began to suffer. In 2022, I bought a used Roland R-05 (which by all accounts is similar to the mighty M10), but my tests showed that it actually had marginally more self-noise than the Zoom H1 at levels 16 and 37.

At zero gain (level 16), the Zoom H1 has less self-noise (-99.8 dBFS[A-weighted]) than the Roland R-05 (around -96 dBFS[A]). At +13.5 gain, the Zoom H1 has a self-noise of -97.5 dBFS(A) — still better than the R-05 before any gain at all is added! It is only when you need more gain from the recorder itself that the R-05 shines. It can provide much higher levels of relatively clean gain than the H1, but I didn't need that because I used the STC-9000.

So I reverted back to using my Zoom H1, but I kept itching for a solution to the noise problem in high dynamic range situations.

This year, I finally bit the bullet and bought a Zoom F3. I'm very happy with it — both in terms of quality and footprint (I don't need to use an external preamp or battery box). But after recording a couple of fairly quiet shows, I found the gear still had higher self-noise than the room's own noise floor. This is why I decided to run these tests: to figure out what exactly is going on.

My modus operandi with the Zoom H1 was that I would pick either 0 or +13.5 dB gain in the recorder, and then use the STC-9000 to add however much more gain was necessary. My STC-9000 unit only has two gain settings (+16 dB and +36.3 dB), plus an attenuation pot. For most shows, I'd set the Zoom H1 to +13.5 dB, the STC-9000 to +16 dB, and then adjust the attenuation pot as necessary.

Don't get me wrong, -97.5 dBFS(A) of self-noise is not bad at all. It's better or on par with much more well-regarded handheld devices. But because I often recorded shows with unpredictable dynamics, I needed to be very conservative with my gain staging in order not to clip; but then it soon became clear that being too conservative would definitely expose the recorder's own noise floor. If the loudest parts peak at -15 dBFS and the quiet parts at -45 dBFS, normalising to 0 dBFS will bring the recorder's self-noise up to -82.5 dBFS. Add some compression, and suddenly that self-noise starts cropping up during the quieter parts.

Even if all things were equal, the Zoom F3's perfect gain staging with dual ADCs and 32-bit float would allow me to reduce noise by at least -15 dB (compared to my conservative staging). I thought this would be enough to eliminate all audible noise from my recordings, but this didn't turn out to be the case. So yesterday I decided to run my gear through some tests.

THE TESTS

In a quiet and empty house, with the living room's door closed, I placed my CA-11s on a table. About 150cm away, I set my laptop up with volume at 8%. For each gear configuration, I would press record, wait 5 seconds, play the Windows notification sound, wait 20 seconds, play the notification sound a second time, wait another 20 seconds, and then press Stop.

When analysing the resulting files, I took the peak data from the notification sounds (I always recorded two to make sure they were consistent). I took the noise data from the stretches of silence between notifications.

All figures below are calculated with the assumption that the Zoom H1's 16 gain setting indeed represents 0 dB. I'm fairly sure this is correct.

Here is a list of all the gear I used:

Church Audio CA-11 cardioid microphones
Micboosters Clippy EM272M omni microphones (-6 dB less sensitive than the EM272Z1 capsules, same 80 dB SNR, more resistant to RF interference)
Battery Box BB2
Uši phantom adapter
MEIRIYFA 1/8 Female to Dual XLR Male Stereo Cable

When connecting the microphones directly into the Zoom F3, I used the Uši phantom adapter and MIC +24V phantom power. When using a preamp or battery box, I turned phantom power off (still MIC, not LINE) and used the MEIRIFYA adapter instead.

CA-11s > Zoom H1 (level 16)

Peak: -64.6 dBFS
Noise (A-wtd): -99.8 dBFS
Gain: 0 dB
EIN: -99.8 dB

CA-11s > Zoom H1 (level 37)

Peak: -51.1 dBFS
Noise (A-wtd): -97.5 dBFS
Gain: +13.5 dB
EIN: -111 dB

CA-11s > STC-9000 (low gain, no attenuation) > Zoom H1 (level 37)

Peak: -35.1 dBFS
Noise (A-wtd): -87.7 dBFS
Gain: +29.5 dB (16 dB + 13.5 dB)
EIN: -117.2 dB

CA-11s > STC-9000 (high gain, no attenuation) > Zoom H1 (level 37)

Peak: -14.8 dBFS
Noise (A-wtd): -68.8 dBFS
Gain: 49.8 dB (36.3 dB + 13.5 dB)
EIN: -118.6 dB

CA-11s > Uši phantom adapter > Zoom F3 (x1)

Peak: -54.3 dBFS
Noise (A-wtd): -107.2 dBFS
Gain: +10.3 dB (16.8 dB - 6.5 dB)
EIN: -118.1 dB

CA-11s > STC-9000 (low gain, no attenuation) > Zoom F3 (x1)

Peak: -55.1 dBFS
Noise (A-wtd): -107.8 dBFS
Gain: +9.5 dB (16 dB - 6.5 dB)
EIN: -117.3 dB

CA-11s > STC-9000 (high gain, no attenuation) > Zoom F3 (x1)

Peak: -34.8 dBFS
Noise (A-wtd): -87.5 dBFS
Gain: +29.8 dB (36.3 dB -6.5 dB)
EIN: -117.3 dB

Clippys > Uši phantom adapter > Zoom F3 (x1)

Peak: -35.4 dBFS
Noise (A-wtd): -94.2 dBFS
Mic sensitivity: +18.9 dB (compared to CA-11s)
Gain: +10.3 dB (16.8 dB - 6.5 dB)
EIN: -123.4 dB

Clippys > STC-9000 (low gain, no attenuation) > Zoom F3 (x1)

Peak: -35.7 dBFS
Noise (A-wtd):  -94.3 dBFS
Mic sensitivity: + 19.4 dB (compared to CA-11s)
Gain: +9.5 dB (16 dB - 6.5 dB)
EIN: -123.2 dB

Clippys > STC-9000 (high gain, no attenuation) > Zoom F3 (x1)

Peak: -15.2 dBFS
Noise (A-wtd): -74.2 dBFS
Mic sensitivity: +19.6 dB (compared to CA-11s)
Gain: +29.8 dB (36.3 dB - 6.5 dB)
EIN: -123.3 dB

Clippys > Battery Box BB2 > Zoom F3 (x1)

Peak: -48.6 dbFS
Noise (A-wtd): -106.4 dBFS
Estimated mic sensitivity: +19.4 dB (compared to CA-11s)
Gain: -3.4 dB (+3.1 dB - 6.5 dB)
EIN: -122.4 dB

OBSERVATIONS

- When I run the CA-11s straight into the Zoom H1 (with no gain) and the Zoom F3 (with no waveform magnifcation) with the Uši phantom adapter (+16.8 dB gain), the Zoom F3 is louder by 10.3 dB, which indicates that its LINE mode records the signal -6.5 dB quieter than expected (I don't know if this attenuation occurs digitally or at the analogue stage). The noise floor recorded was -107.2 dBFS, so if we digitally attenuate the signal in post to match the Zoom H1, the noise is really -118.1 dBFS (vs the Zoom H1's -99.8 dBFS). The Zoom H1 bridges this gap somewhat at level 37, going from -99.8 dBFS to -111 dBFS — still an 8 dB difference there.

- One thing that threw me off while looking at these numbers was how the Zoom F3 barely recorded any level differences when I added the Church Audio preamp to the chain! To make sure I wasn't overlooking anything, I ran the test again with a battery box, and the results were consistent. I'm skeptical that my XLR/phantom adaptors are the culprit here, so this must mean the Zoom F3's MIC setting is 16.8 dB more sensitive when it is outputting +24V phantom power! This appears to make little difference in terms of EIN, but I wonder if this affects the analogue clipping levels of the device. It'd be cool if other Zoom F3 owners investigated this! It turns out the Uši phantom adapter provides 16.8 dB of gain! Thanks, commongrounder. I have updated this post to reflect this.

- I actually have two pairs of CA-11 cards. All this data was gathered with the newer pair I received in 2023. Preliminary testing showed that my older pair had a strange noise profile — the right channel displayed the same amount of self-noise as either channel of the newer pair, but the left channel produced +10 dB of self-noise.

CONCLUSIONS

- When using the CA-11 cardioid microphones with a Church Audio preamp, you're not going to get any less noise by upgrading from a Zoom H1 to a Zoom F3 or any such recorder, IF you don't need to use the CA preamp's attenuation pots to set conservative levels. The CA-11's self-noise only allows you to get a maximum EIN -118 dB, no matter how much more clean gain you throw at it.

- The measurements with the Clippy microphones show that they have a better SNR than the CA-11s, and the EIN plateaus at -123.3 dB or so even when you add a lot more clean gain. This was only achieved with a Zoom F3. A previous, less strict set of tests showed me that the Clippy mics, when paired with a Zoom H1 (level 37) and the CA STC-9000 (low gain) displayed an EIN of around -120 dB. Overall, at those levels it starts getting really hard to make precise measurements because of my living room's own noise floor. With the the CA-11s, the recordings didn't have any discernable ambient noise over the the gear's self-noise. With the Clippy, you start being able to discern the sound of traffic far away outside.

TLDR:

(1) I found some weird behaviour in the Zoom F3 when toggling phantom power on and off in MIC mode. If any of you can get to the bottom of this, I'd be interested! commongrounder cleared this up for me.

(2) If you use CA-11 cardioid mics:

- A Zoom F3 is definitely better than a Zoom H1 when it comes to recording quiet sounds.

- A Zoom H1 + a CA preamp combination is equal to a Zoom F3 IF AND ONLY IF you are able to set your levels perfectly. The bottleneck is the microphone's self-noise and also the unpredictable levels tapers record.

[Rairun]

I meant to post this thread under "Recording Gear". If any mods could move it for me, I'd really appreciate it!

[commongrounder]

Quote from Rairun:- One thing that threw me off while looking at these numbers was how the Zoom F3 barely recorded any level differences when I added the Church Audio preamp to the chain! To make sure I wasn't overlooking anything, I ran the test again with a battery box, and the results were consistent. I'm skeptical that my XLR/phantom adaptors are the culprit here, so this must mean the Zoom F3's MIC setting is 16.8 dB more sensitive when it is outputting +24V phantom power! This appears to make little difference in terms of EIN, but I wonder if this affects the analogue clipping levels of the device. It'd be cool if other Zoom F3 owners investigated this!

It might be helpful for you to know the LOM Uši Phantom Adapters are not just voltage adapters, but also contain active electronics in them with gain. The circuit also provides a low impedance truly balanced output. I have several of these adapters that I’ve used and modified for various mic capsules, and they are fantastic. The Zoom F3 definitely does not alter its preamp gain when phantom voltages are turned on or off. I can add that I have been getting outstanding results using Primo EM272 capsule/LOM phantom adapter based mics into my F3, both with music, and with ambient nature soundscape recording.

[Rairun]

It might be helpful for you to know the LOM Uši Phantom Adapters are not just voltage adapters, but also contain active electronics in them with gain. The circuit also provides a low impedance truly balanced output. I have several of these adapters that I’ve used and modified for various mic capsules, and they are fantastic. The Zoom F3 definitely does not alter its preamp gain when phantom voltages are turned on or off. I can add that I have been getting outstanding results using Primo EM272 capsule/LOM phantom adapter based mics into my F3, both with music, and with ambient nature soundscape recording.

Wow, I didn't expect that! I wouldn't be surprised by a difference of a couple of dB, but +17 dB gain is a lot. I'd have thought they'd advertise this or at least include it in the spec sheet.

I found the Primo EM272 capsules too sensitive for loud shows (they overloaded my Zoom F3 when I tried). I guess I should try them again at some point using only a battery box!

[admkrk]

I only skimmed through this, but why are you concerned with the recorded levels from an F3? With 32 bit float it does not really matter. Just normalize and all is good, for the most part.

[commongrounder]

It might be helpful for you to know the LOM Uši Phantom Adapters are not just voltage adapters, but also contain active electronics in them with gain. The circuit also provides a low impedance truly balanced output. I have several of these adapters that I’ve used and modified for various mic capsules, and they are fantastic. The Zoom F3 definitely does not alter its preamp gain when phantom voltages are turned on or off. I can add that I have been getting outstanding results using Primo EM272 capsule/LOM phantom adapter based mics into my F3, both with music, and with ambient nature soundscape recording.

Wow, I didn't expect that! I wouldn't be surprised by a difference of a couple of dB, but +17 dB gain is a lot. I'd have thought they'd advertise this or at least include it in the spec sheet.

I found the Primo EM272 capsules too sensitive for loud shows (they overloaded my Zoom F3 when I tried). I guess I should try them again at some point using only a battery box!

The F3 has a really handy function that I appreciate with loud sound sources: It will allow phantom operation in line-in mode. This means a high output condenser can be used without overloading the preamp input. The line-in is a -20dB pad. This means, of course, that noise performance is degraded by 20dB, but this would normally not be audible in the circumstances under which this mode is used.

The Primo EM272 capsules themselves overload at around 120dbSPL, so they might not be ideal for close micing of shows. You could also try using a “Simple P48” based XLR phantom adapter. This circuit adapts the 48-volt phantom voltage down to the needed capsule voltage, but adds no gain as it passes the capsule audio directly through using a clever circuit design. Essentially like a battery box that runs off of phantom.

[Rairun]

I only skimmed through this, but why are you concerned with the recorded levels from an F3? With 32 bit float it does not really matter. Just normalize and all is good, for the most part.

You're right that it doesn't matter where the audio is placed in the digital container (it matters with 16-bit and 24-bit, but 32-bit float doesn't have bits allocated to a specific dB range), but that's not what I set out to measure here. What I measured was this: given a sound of fixed volume recorded by the device, how loud is the noise floor of my entire rig in comparison? What parts of the rig are most responsible for the noise floor? And how do other devices fare against this?

I didn't think I was going to focus on the recorded levels of the F3, but I encountered an anomalous result, which according to commongrounder is due to the XLR>Minijack phantom power adapter I was using. This matters inasmuch as it reflects the strength of the signal that goes into the F3, because it can be overloaded like any other device (I've overloaded it before with the Clippy mics). Again, when I started running the tests, I wasn't expecting to focus on this - and I don't think that's the overall point of my post, just an interesting side note.

[Rairun]

It might be helpful for you to know the LOM Uši Phantom Adapters are not just voltage adapters, but also contain active electronics in them with gain. The circuit also provides a low impedance truly balanced output. I have several of these adapters that I’ve used and modified for various mic capsules, and they are fantastic. The Zoom F3 definitely does not alter its preamp gain when phantom voltages are turned on or off. I can add that I have been getting outstanding results using Primo EM272 capsule/LOM phantom adapter based mics into my F3, both with music, and with ambient nature soundscape recording.

Wow, I didn't expect that! I wouldn't be surprised by a difference of a couple of dB, but +17 dB gain is a lot. I'd have thought they'd advertise this or at least include it in the spec sheet.

I found the Primo EM272 capsules too sensitive for loud shows (they overloaded my Zoom F3 when I tried). I guess I should try them again at some point using only a battery box!

The F3 has a really handy function that I appreciate with loud sound sources: It will allow phantom operation in line-in mode. This means a high output condenser can be used without overloading the preamp input. The line-in is a -20dB pad. This means, of course, that noise performance is degraded by 20dB, but this would normally not be audible in the circumstances under which this mode is used.

The Primo EM272 capsules themselves overload at around 120dbSPL, so they might not be ideal for close micing of shows. You could also try using a “Simple P48” based XLR phantom adapter. This circuit adapts the 48-volt phantom voltage down to the needed capsule voltage, but adds no gain as it passes the capsule audio directly through using a clever circuit design. Essentially like a battery box that runs off of phantom.

Yeah, when they overloaded the the F3, I was using MIC. I did switch it to LINE as a test, and while it didn't overload that evening, it did come dangerously close! Too close for comfort. It wasn't even the loudest of shows. I'll look into a simple p48 adapter, thanks!

[Rairun]

It might be helpful for you to know the LOM Uši Phantom Adapters are not just voltage adapters, but also contain active electronics in them with gain. The circuit also provides a low impedance truly balanced output. I have several of these adapters that I’ve used and modified for various mic capsules, and they are fantastic. The Zoom F3 definitely does not alter its preamp gain when phantom voltages are turned on or off. I can add that I have been getting outstanding results using Primo EM272 capsule/LOM phantom adapter based mics into my F3, both with music, and with ambient nature soundscape recording.

By the way, sorry for sounding a little skeptical at first, but you were totally right. I've just tested the adapter with an audio interface, and the result is exactly the same. I really had no clue that the LOM Uši Phantom Adapter was essentially a little preamp fed on phantom power and providing PiP. I'll have to edit the post with the correct calculations.

[unidentified]

I would be curious to know what the impact is on the AA batteries in the Zoom F3 when using this device. Would they be drained as quickly as they would be if they were still providing 48 volts of power to the mics?

[Rairun]

I would be curious to know what the impact is on the AA batteries in the Zoom F3 when using this device. Would they be drained as quickly as they would be if they were still providing 48 volts of power to the mics?

The Uši adapter converts 24-48V phantom into 8V PiP, so I've only ever used it with 24V (which should save you battery). When I first got the Zoom F3, I did 3 test runs with a new a pair of Amazon Basics AA rechargeable batteries with this rig: Clippy mics (Primo EM272M capsules) > Uši adapter > Zoom F3 (24V). I consistently got between 6 and 7 hours of battery life. A few months later, I noticed that the battery meter was going down faster than I expected when actually taping shows (with CA-11 cards instead of Clippys). I never ran out of battery during a show, but I've now left it running three times after a show, and I got 2h15min, 3h10min and 3h30min before the F3 shut down.

Not sure if my batteries are going bad, if the CA-11s are more power hungry, or even if other variables cause more battery drain (e.g. I got almost 7h of battery life with the Clippys in an extremely quiet bedroom, while the CA-11s were assaulted by high SPLs - does this make a difference? I'm not knowledgeable about this).

[commongrounder]

The specs from LOM on the Uši Phantom Adapter indicate the device draws ~3ma at 48v. This is right in the common range of many condenser microphones, so I wouldn’t expect there to be much of difference on battery life compared to other mics.

[Rairun]

It might be helpful for you to know the LOM Uši Phantom Adapters are not just voltage adapters, but also contain active electronics in them with gain. The circuit also provides a low impedance truly balanced output. I have several of these adapters that I’ve used and modified for various mic capsules, and they are fantastic. The Zoom F3 definitely does not alter its preamp gain when phantom voltages are turned on or off. I can add that I have been getting outstanding results using Primo EM272 capsule/LOM phantom adapter based mics into my F3, both with music, and with ambient nature soundscape recording.

I tried to contact Jonas from LOM to know more details, but he hasn't replied. I did find an old FB post where someone asks why their mics were so sensitive with the adapter, and Jonas seemed to find the suggestion that there was any gain strange. I'm happy with adapter! I just don't understand why LOM is so cagey about it.

[commongrounder]

Interesting about LOM and info sharing. I’ve never communicated with Jonas about anything, but I have had the LOM Uši Phantom Adapters apart as I build my own mic configurations. The board is a (beautifully made) surface mount design, and looks like their version of a modified classic Schoeps preamp circuit. This basic circuit has been adapted and “improved” (“Alice” variants) for use with various externally polarized and electret capsules (adding an appropriate capsule bias voltage supply). I have built a version myself utilizing through-hole components to upgrade a base-model 3Dio binaural mic I have to balanced phantom. These all have some gain compared to the electret capsule output in a “simple P48” configuration (capsule output direct to preamp input). The LOM adapter is really quite a bargain for what you get (two for $35us), which is why I’ve taken to buying them now instead of building my own, and repurposing them for my mic building projects.

[Rairun]

I would be curious to know what the impact is on the AA batteries in the Zoom F3 when using this device. Would they be drained as quickly as they would be if they were still providing 48 volts of power to the mics?

So it turns out my batteries weren't dead, but my charger was detecting them as full when they were nowhere near fully charged. I got a new charger, and now the CA-11s > Usi Adapter > Zoom F3 (24V) rig runs for around 7h40min. I wouldn't expect using 48V to lower this by more than 20 min, tops.