with award winning documentary team tony wilson and leo sullivan


Canon 5D
The CameraClass workshop


The aim of the exercise is to establish the viability of sound recording on the Canon 5D stills camera while shooting HD video in the Movie Mode.


A simple test was designed to compare a recording on the Canon 5D with a 'control' recording on the Sound Devices 744T Harddrive.

I also wanted to look at some other devices that could be used in conjunction with the Canon 5D to make it more practically functional. These included a Sennheiser G2 ew100 Wireless System, a Beachtek DXA 5Da mini mixer and a Zoom H4 SD card recorder.


A simple interview situation with a cardiod microphone overhead was fed into a mixer and a mic level output of that mixer was fed into the following devices;

1) Via cable to one channel (CH. 2) of the 744T recorder. This is the 'control' or master reference.

2) Via cable to Zoom H4 SD card recorder

3) Via Sennheiser G2 ew100 radio link to the other 744T channel (CH. 1).

4) Via Sennheiser G2 ew100 radio link to the Canon 5 D mini-jack input.

Recordings were conducted using the Canon 5D set in Manual recording mode and in Auto recording mode. When in the Manual mode, a reference tone was set at -20db.

A similar recording was conducted using the Beachtek DXA 5Da mixer which claims an Auto-override function. Tests were conducted with Canon 5D in Auto recording mode with the Beachtek override 'on' and with the Beachtek override 'off'.


The recordings were imported into Pro-tools and all similar tracks were put in sync to make comparisons very easy and clear.


As expected, the cabled recording into the 744T was very good. The other recording onto the 744T, via the Sennheiser link was also very good. A very slight increase in high-end 'hiss' was detected, as well as a slight increase (surprisingly) in the lower frequency range. Neither of these variations was considered significant in terms of broadcast quality.

The recording via cable onto the Zoom H4 maintained good frequency response and clarity, but was marred by a low level 'buzz' that I have termed 'engine noise'. In a noisy environment this may not be noticeable and could possible be cleaned-up in a mixing situation, but was clearly audible in this test.

The recording via radio link into the Canon 5D (in Manual mode) was inferior to the 'control' in two ways. There was a distinct increase in 'hiss' and there was a noticeable drop-off at the bottom end. I felt the recording had a sharpness or edge to it that gave it a 'digital' quality.

When in the Auto mode, the overall level was boosted significantly and the background levels varied noticeably in between moments of speech. The effect is quite pronounced and would hinder use of the recording.

Further recordings via the Beachtek DXA 5Da mixer (which enable more than one source into the camera) produced some interesting results. When the Canon was in the Auto mode, the Beachtek auto-override device seemed to have no effect, and the background variations continued. With the Canon in Manual mode, the Beachtek mixer enabled me to put separate mics to separate channels but the device added more 'hiss' to the recording. (Note: Since this test, Beachtek have released a DXA - SLR version with a 15db gain switch on mic inputs, which should lower the noise floor considerably.)


The Canon 5D stills camera has evolved into an alternative production device for shooting high quality video at a low cost. By all accounts the picture qualities are extraordinary for the price of the unit, but it's clear that the sound recording quality is compromised.

Based on the results from this test, I recommend that the way to achieve the best quality sound when shooting on the Canon 5D is to record separate sound on a higher quality device, like the Sound Devices 744T. Further, I think there would be significant advantage in recording a guide track onto the camera via one of the small wireless systems, like the Sony or Sennheiser, to assist in the editing process.

If in the event that no separate system is available, a powered (with its own battery) microphone like the Rode NGT 2 with the correct mini-jack adapter could be used direct into the camera. It is strongly recommended that the camera audio setting is in the Manual mode and an approximate level is set. However, not having monitoring is a disadvantage.

Although the Beachtek DXA 5Da mixer gives the added advantage of headphone monitoring and two separate inputs, the combination of the 'hiss' produced by it and the Canon recording system, in my opinion, make that option unsatisfactory. (See earlier note about new model from Beachtek)

There are significant issues with the practicalities of operating this camera in the field. Some, as we have seen, are about the sound quality. Single mic input, no headphone monitoring, short recording times (approx 15min per shot). There are also some other issues about the camera operation, not sound related, that present problems for documentary work, particularly in hand-held operation.

The Canon 5D is bound to take its place in the arsenal of video production camera options, but the in-camera sound recording quality leaves a lot to be desired and certainly doesn't match up to the standard that we would expect for film or television broadcast.