Audio difference between HV20 and HV30 - it exists, sort of

[nolonemo]

A recent thread cited a review by a German magazine which stated that the Canon engineers had addressed the motor noise problem by limiting the high frequency sensitivity of the camera's audio circuits, citing tests done with the MKE400: http://hv20.com/showthread.php?t=11187 (see Fantomaz's post about half way down the first age)

Since I have an HV30 and still haven't managed to sell my HV20, I decided to follow up. I set up a tripod about 3 to 4 feet from a chair, set each camera to manual audio gain with the green bar at the -12db mark (chosen only because it was easy to match on each camera) and recorded some comparisons (record on HV20, replace camera on tripod, transfer tape and mic to HV30, and record). I did not do any level adjustment or normalizing to the audio files. The samples were recorded in 3 different sessions, so there is some variation in the camera-subject distance. All the comparisons are contained in an mp3 file at the end of the post. Below are a number of screen captures of Sound Forge's spectrum analysis of the various audio files (I selected the voice ID and about 5 seconds of "silence" (i.e. motor noise) after the intro for each file). Each graphic has the HV20 spectrum (light blue) and the HV30 spectrum (magenta) overlaid (where the spectra overlap, they show as purple).

First, the internal microphones, which sound pretty very similar, the HV30 sound a little less bright, but I'm not sure if that's because the highs are attenuated or the low/mids are boosted and mask them. But the spectra are very close. In any event, note the crappy high frequency response of the internal mics. It might be that the Canon engineers did put a high cut filter in the internal mic circuit, since the dropoff is so abrupt:



Then, the MKE400. I set the MKE400 sensitivity on "+" since that's where I normally would set it. In this case, the HV20 sounded much brighter, and, indeed, the spectra showed that the high frequencies fell off earlier on the HV30:



However, comparions made using the Audio Technica AT-822 and the Sony ECM-908 did not display the high frequency dropoff. On the contrary, the HV30 audio sounded brighter, and the spectra appear to show better high frequency response in the HV30:





This got me wondering if what was happening was an artifact of the way the mics were mounted on the camera, since the MKE400 was attached using its integrated shock mount, and the AT822 and ECM908 were attached using a Rode SM3 shock mount.

So I reran the MKE400 comparison attaching the MKE400 via the Rode shock mount. This time the audio and the spectra showed the same pattern of slightly better high frequency response as the AT822 and the ECM908:



Finally, I recorded some pink noise off a reference CD played in my laptop and connected straight to the camera:



I leave you to draw your own conclusions from all of this. Having said that, it appears to me that while Canon may have cut high frequency response (in both cameras) on the internal mic circuits, that does not seem to be the case with the audio circuits in general, as witnessed by the pink noise spectra.

Here is the link to a 320 kpbs MP3 containing all the comparisons:

http://www.nolonemo.com/files/hvcomp...omparisons.mp3

And, for those of you wondering if the physical motor noise is different between the two cameras, it isn't. Here's a comparison recorded with a Zoom H2 mounted on a light stand about 6" away from the tripod-mounted cameras while each was recording (HV20 followed by the HV30):

http://www.nolonemo.com/files/hvcomp...comparison.mp3

Finally, here are some lower-frequency plots for the various mics:

[GiriGauri]

wow! thanks for this.I have been meaning to delve deeper into this issue myself, what you demonstrate lays all my doubts to rest.

Lazy canon engineers just attenuated the highs. So no more motor buzzing. Cheap and cheerful but at the cost of crisp highs.

[fantomaz]

great test nolonemo, thank you
it definitly looks like the high frequencies get cut out inside the internal mic! it's kind of funny that just the mke400 (with internal shockmount) shows the effect they were talking about in the magazine. however imo this shows that they didn't crosscheck with different setups or different mikes and thats pretty weak for a 8 EURO magazine
thanks again nolonemo