Tuesday, August 26th, 2008 | by matthew mcglynn
I bought a Zoom H2 handheld flash recorder for interviewing. I picked the H2 because the form factor was appealing — it’s small — and for its lack of moving parts (hello mini-disc!). The reviews I read raved about the feature set, and although users were generally positive about the sound quality, I wasn’t shopping for a high-fidelity field recorder so much as a simple interviewing tool.
And it has four microphones! They might as well have put my name on the case.
As an interviewing tool, the H2 works wonderfully. Without touching the manual, I was able to set it to record 192 kHz MP3s with a “vocal” compression patch inline, and it made a couple extremely clean interviews for me with no fuss and no discernable noise. The LCD display of the input level was very helpful as a sanity check that it was still working throughout the interview.
My old interviewing rig was a laptop: I was using the built-in mic on a Macbook Pro. Designed for sources two feet away rather than 10 feet away, it was barely adequate, and not very portable, so it got left behind when, during an interview, we’d wander around studios to look at gear.
For a while I also carried my old Walkman-size cassette recorder with a little external mic, as a backup recording device. Despite being ancient technology, it worked really well, and it let me open interviews with a wisecrack about how the laptop was recording but “I just can’t get enough of that analog sound!”
No, nobody ever laughed.
The other day I had an opportunity to record a band rehearsal with the H2. Not wishing to futz around with it, I simply turned it on, pressed the red button, and set it in the middle of the room.
The sound quality was decent. My goal was to capture the jam, not to make a live CD, and for this it was ideal. It created a mixed, ready-to-play MP3 file that I could easily share with the other band members.
The only problem? Audio dropouts. I made three recordings of 10-20 minutes apiece. Each track contains 2-3 audio dropouts. They last from 1 to 1.5 seconds each. Hear two of them here:
The H2 had all four mics active, and was set to record a stereo MP3 at 192 kbps, with the COMP2 compression patch still inline. I was using the supplied 512 MB SD card, and version 1.30 of the OS.
The designers of the H2 foresaw the potential for dropout problems. Page 62 of the manual is dedicated to this topic.
Depending on the combination of operation mode and recording mode, data transfer to the SD card during recording may momentarily not be able to keep up with the data stream, leading to a brief sound dropout. When this occurs, the indication “Data write Error” is shown on the display during recording and after the recording is finished.
I didn’t notice any such warning on the H2 during rehearsal, but I can’t be certain it wasn’t there. Unfortunately I also can’t check the H2’s marking system, which bookmarks points in the audio stream where dropouts occurred, because I’ve already purged the files.
However, the behavior here is not what is predicted by the H2 manual. A chart in the manual suggests that dropouts only occur under “heavy processing load,” such as when recording 24 bit, 96 kHz stereo WAV files, or 4-track 24-bit 48 kHz WAV files. Either way, that’s something like 25 times as much data as I was recording (based on the H2 SD-card capacity chart at Oreilly.com).
Let’s do the math. I used Bias’ Peak Pro software to upsample and convert a 10-second song sample to two formats, for this size comparison:
- 10 seconds of 24-bit, 96-kHz WAV audio consumes 5625 KB of space
- a 192 kbps, 16-bit MP3 of the same audio consumes only 240KB
That’s a 23x difference in filesize, and therefore a 23x difference in data bandwidth.
So, whatever problem this was was not likely to be due to the SD card’s throughput rate; if the card is supposed to be capable of 23x more data throughput, why would it be choking on my puny MP3s? I speculate that the problem was actually CPU overload.
I’ve since ordered a faster SD card, not so much because I think it will matter, but because I needed a bigger one anyway. I’ve upgraded the OS to 1.4, which also should have no effect. And next time I’ll disable the compressor patch. Given that none of these changes is likely to fix the problem, though, I suspect the problem will continue. (Maybe I need to record WAV files instead?)
I will post an update after our next rehearsal. Based on the comment from Stefano (see below), I did a quick test: recorded 5 minutes of music twice, once with the compression filter enabled and once without. All other settings were identical: 4-mic, 2-channel surround mode, 192 kHz MP3. Both tests generated dropouts. Neither test generated warning messages. Maybe this unit is defective.