Topic: Zoom H2essential. It seems to be a 6 channel device! - Version 2.01 F/W update

[Ozpeter]

Here's another video from a different channel, where the H2essential was used at the front of a chamber performance stage (you can see it there) and an H2n was used as a spot mic on the piano.  Again, I find the results to be surprisingly good and could be tweaked a little to add to the bottom end according to taste.  Apparently the XY configuration was used in the recorders rather than MS.

https://youtu.be/P-TevXJWXOc?si=bH0QnefKSQtuAxEJ

There are some moments of silence between movements where I can't hear any significant system noise. 

I have heard from the uploader of the video involving larger scale performances that indeed the H2e was used without external mics.  The levels were somewhat processed afterwards for the purpose of the video, including the use of a limiter effect, which means that the original dynamic range was greater than that in the video.

After all my years of running cables trip-free through concert halls I am quite jealous to see recordings being made without any cables at all!

[Ozpeter]

I just came across a video where someone who runs a substantial business recording classical singers in London has reviewed the H2essential from the point of view of classical singers making their own demos.  Clearly he has enough business not to worry if some potential customers adopt a DIY approach. 

Anyway, he has done a unique test of the H2e where he compares it with a recording made with an expensive AEA R88A stereo ribbon mic at the same time (into an unidentified recorder).  The subject is a soprano with piano accompaniment.  He acknowledges that the Zoom is a bit brighter/thinner, but then he applies some eq (and a touch of reverb) and compares again, and the result is quite persuasive.  Given that he can claim to know what he is talking about, it's good to hear that overall he's impressed with this very cheap device.

https://youtu.be/W1NLVLp0qMA?si=X_rCuJdLYvvdaUKC&t=515

The comparison is at the 8:35 mark - the link should take you to that spot.

I've contacted him at his company to thank him for the video - and to point out that the H2e doesn't have dual ADCs which his introduction kind of implied that it does.  But that shouldn't be an issue unless the H2e was placed stupidly near to the soprano.

[rastasean]

Good for him doing more with less.

[Ozpeter]

Shamelessly copied from a comment on YouTube -
----------------------------------------------------------
UPDATE: It appears that Zoom had quietly rolled out a firmware update (currently version 2.01) for the H2essential.

This either preceded or coincided with their announcement today, April 24, 2026, of the TCA-1 timecode adapter and firmware updates, across the handy recorder lineup.

H2essential v2.00–2.01
• Support for SMPTE timecode input/output using a ZOOM TCA-1 timecode adapter
• Input gain setting function added
• AI Noise Reduction function added to analyze and reduce environmental sounds and other noise
• Functions added to enable recording and playback of MP3 files
• Functions added to export files in MP3 format
• "Off" added to the "LED Brightness" menu
----------------------------------------------------------

Most of these updates won't require explanation or testing.  But the "AI Noise Reduction" function certainly requires testing.  Experience with such stuff in other devices is, well, mixed.  Sometimes it's ok to use if essential, and sometimes it can be absolutely awful.  One could also take the view that it is often best to do noise reduction in post production rather than bake it into the recording, of course.  But I will post some samples on YouTube when I have updated mine. 

The "input gain" function is almost certainly just digital gain processing which could also be done as required when editing, but I guess having it onboard could be handy in situations where the user might need a recording to be available immediately with some change to the previous fixed gain setting.  And maybe this particularly applies to mp3 files, where processing afterwards is ideally avoided lest mp3 artifacts get created.

Lastly the timecode adapter may provide remote control from a phone.  But I have no plan to buy it to find out.

Zoom announcement -

https://zoomcorp.com/en/us/news/h2essential_v20/

[Ozpeter]

I've now done the update and tested the new AI noise reduction feature.

When you record with the AI NR turned on, it records to a "mix" file which contains the AI NR processed audio.  But it also records, in the same folder, to a separate file which is the unprocessed audio.  This is excellent as it means that if the AI processed sound is not good, you can revert to the original.  Total insurance.  Well done Zoom.

I did three quick tests - one in my 'studio' where the PC puts out a slight amount of fan and disk noise into the room.  Then in the bathroom with the extract fan on.  Then out on the balcony of my suburban home with a fair bit of traffic going past nearby.

Firstly, when you turn on the AI NR feature, a message appears indicating that the recorder is assessing the current noise level.  This tends to mean that it won't deal with significantly fluctuating noise (but I'm not totally sure).  Also, I imagine that it will be essential to ensure that you avoid creating handling noise while the noise level assessment is being processed.  There is a slight pause before the message appears, and it seem to just take a couple of seconds to undertake the assessment.  Then you can hit record.

Doing some analysis with Adobe Audition, it looks like the AI NR function reduces noise by around 12dB.  But what does it do to the required sound, in this case speech?  Well, in the studio the speech was slightly affected (in terms of sounding 'processed') and in the bathroom, the speech sounded more processed, to the point where it was only just acceptable (but that does rather depend on many factors, eg do you prefer to have the loud sound of the fan, or the processed sounding speech?).  In the outdoor test, I guess the same thing applies - is the 12dB reduction in background noise worth the moderately processed sound of the speech?

I think I would probably only use this AI NR feature if I wanted to record narration for something like a YouTube video in my slightly noisy studio.  Or even in a relatively silent place where I wanted to reduce mic and preamp noise by 12dB.  Otherwise it would be good for situations where it was essential to reduce background noise to the point where a bit of processing of the desired audio was an acceptable option.

Would this new feature have any use in the context of recording music in a taperssection environment?  I doubt it.

As for the new level setting feature, I haven't tested it in detail, but I did discover that the way to set the levels is simply to press the forward/back buttons on the recorder, which allow precise changes of the default "70" setting.  But once you press record, the levels can't be changed.  Useful?  Well, maybe handy in some situations.

mp3 settings are 128 / 256 / 320kbps.  Useful for some people.

All in all, this update helps position the H2essential as a multi-purpose recording tool with an impressive array of optional features - at a quite reasonable price.

[Ozpeter]

A final update - I did one more test, in the quietest room in my house.  This time, processing of the voice was very slight.  Perhaps the total silence between each syllable troubled the ear a little.  And background noise was reduced by 12dB again, across the entire audio spectrum.  After I had normalised the file, so that my speech hit -0.01dB, when I wasn't talking the 'silent' audio level was -66dB at 2kHz, -81dB at 10kHz.  The AI NR version figures were  -78dB at 2kHz, -93dB at 10kHz.  For a recording using a cheap device in a room which is not soundproofed, that's quite a dynamic range.

[Ozpeter]

Here is my video with extensive audio samples with the new NR feature disabled and enabled.

https://youtu.be/ewkmjsAN_04

(This doesn't make money, my channel is not monetized).

One very important thing that I discovered when creating the audio samples is that each time you start recording in a new environment, you have to turn the AI NR off, then back on, otherwise it will use the noise sample from the previous location - and in my original testing reported above, I may have got that wrong.  Used correctly, this feature does a pretty good job of reducing noise while leaving voice recordings about as acceptable as you could expect at the price of the device, if that makes sense.  Would it work with music?  I haven't tried but I'm not sure what noise you would be trying to remove at a performance.  Applause?