In Camera sound with Canon EOS 5D Mk. II- Is it that bad- You Judge
 

We have heard a lot of disgruntled talk about the in camera sound capabilities of the Canon EOS 5D Mk. II. Yet I haven't really heard any actual sound tests posted out of it.

This morning I decided to link my 5D2 up to my Beachtek adapter, DXA-4 and ME66 mic. I set the mic and camera up about 4 to five feet away from me. I initially set the levels on the Beactek at 6 oclock position. Toward the end of the recording I moved one side of the levels to 9, just to see if it would make a difference. I think I detected clipping once when I did that.

Chris was nice enough host on DVInfo site here, and here is the link so you can hear for yourself. Here is the link:

http://www.dvinfo.net/media/canon/5d2/soundtest.mp3

What do you sound gurus think ?

Chris, good test... lots of hiss, but that can be taken care of in post if need be... and the voice sounded well, no breakups, pops or such...

and of course you got to test out the auto limiter during the yell... and it does work :-)

A good and cheap program to clean up the hiss would be audacity... and its free to boot.

Audacity: Free Audio Editor and Recorder

give it a shot and do an A/B test :-)

You can't even TRULY monitor what's being recorded.

Right there, for any serious project, is a nonstarter.

Not to say it's not usuable for some purposes, just saying...

Chris, thanks to this test I now realize that the 5D is not up to snuff sound-wise -- when even with $400+ worth of sound peripherals won't make this camera sound better than that. Adding to this, as Bill pointed out, it's also unfortunate that you can't even monitor the sound. I wonder what the engineers at Canon were thinking.

-- peer

It's still a forward step in evolution with this camera even existing. It will only get better in the future.

I would never expect XLR on this camera, not even in future releases, and that doesn't bother me one bit. It's a STILL camera primarily, let's not forget. And I also wouldn't expect any sort of compressor/limiter options either, not now, not in the future. If you need that stuff, you're not doing in camera IMHO.

Now despite the obvious crappy A/D, crappy preamp, and unbalanced inputs, I could have lived with those things for some projects, but not having (1) manual level control, (2) a meter, and (3) monitoring, now those are things that at some point should be added to any camera that's going to have audio inputs. Without those three things -- although I could probably deal with either (2) or (3), but I'd rather have both -- the audio section is dead to me. Completely useless for any serious project. I was so stoked when I found out it recorded uncompressed sound, only to be completely let down by these shortcomings. I mean just think if you could have turned down the 5D2 gain and then ran your audio through an external preamp, you get get beyond a lot of the shortcomings of the audio section, but nope, no dels on that front.

That all said, I still think this cam is an amazing leap forward and a sign of things to come. But yes, it's early-adopters-ville in the a/v department. For me, a primary stills guy who was already incorporating ambient audio into multimedia pieces before the 5D2 came out, this camera is simply AMAZING.

It already has a compressor & limiter. I don't understand what you are talking about here.

-- peer

I have to say that this test wasn't so much about audio quality, as it was about seeing if I could input sound through an XLR adapter, and not end up with clipping at some point in the chain. This is particularly important because of our lack of ability to actually monitor camera's sound. We should be able to rig monitoring to the point of entering the camera, but not what camera actually is recording. If this was about sound quality, the choice of microphone (ME66) in an interior shoot is itself an issue, because it is notorious for picking up room room echo, and is not recommended for that. It was the only powered mic I had available, however, and since my XLR adapter does not provide Phantom, I lived with it.

What this test does show, in my mind, is that you can put sound in to the camera, setting some limits with the XLR adapter so you can avoid worry about clipping. Normal room conversation for a narrative shoot will generally come through okay. You are still left with dealing with auto gain and the "hiss", but for the "rebel" narrative film shooter, the option to shoot this way is there.

The sound man I work with, Chris Swanberg, suggests that double system is in our future if we shoot this camera for a serious film project, but that we can work with this set up in less demanding shoots.

Double system without timecode..?! Yea, that's very "serious".

-- peer

For me, the audio solution is about as good as we could have expected. The onboard mic makes it easy to sync external sound accurately. Sure, it's not timecode, but I don't think anyone would realistically expect the 5D MkII to support that.

It would be cool if it had balanced inputs, settable gain, and and output monitor. Hopefully that will make in into the Mark III. Timecode? Maybe in the 1D Mark IV.

PLENTY of very high end movies were made with a clapper and logging (and no timecode) back in the day. It ain't exactly rocket science.

Excuse me, I should have said "controllable" I guess. So, OK technically that's true, but just look around, there are tons of cameras with AGC and manual modes but not actual user-controllable compression/limiting. All I was really saying is that I could care less about compression/limiting, without manual levels it sucks flat out as far as I'm concerned.

At the end of the day, you and I are basically on the same page as far as the audio section is concerned. So, no flamewar with me. I agree with you on that front. Where we disagree is I still think the camera is phenomenal despite its shortcomings.

For the record, I record a lot of live music, and I'd never, ever, consider using the 5D2 in those circumstances. My typical setup would be:

SBD & AKG480+ck61/ck62/ck63 > Oade-Modded Edirol R44 > 4xWAV(24/96)

On my shoot last weekend we ran a few different setups, with either the onboard mic or a feed from a zoom h4 into the body. The onboard mic was surprisingly decent for music - levels tended to stay consistent overall. Dialogue wasn't good of course but that's as much to do with not having the mic close to the speaker as anything. Unfortunately the location of the mic pretty much made everything unusable as even with music you could hear the constant ticking of the lens' image stabilizer. As a reference track though the built in mic was perfectly sufficient. Of course avoiding image stabilized lenses would help too but I don't know that the tradeoff of more rolling shutter artifacts would be worth it.

With a feed from the H4 the interview sound quality was perfectly acceptable - I would have used it except for the auto gain, which pumped the preamp noise way up every time the speaker paused, and then immediately clamped down on the first word. If not for this I wouldn't have any problem with the onboard sound quality, but in the end I found the onboard audio, even with an external mic, unusable - except as a reference track for rough cuts and visually syncing to the zoom's audio.

Syncing the zoom audio was a bit of a hassle as it does slip sync over time - although it seems to do so consistently at least. I had to cut and slip the audio 2 frames forward every minute to keep it in sync, now that I'm not under a hard deadline I plan to just calculate the precise percentage so that in the future we can pre-stretch the audio before syncing it. Or, with luck, the H4n's clock will more closely match that of the 5D.

Okay, let's re-start this thread. Keep it professional and keep it technical, please. Many thanks to those who are taking this discussion seriously and are contributing more than just snide one-liners, most of which I have now withdrawn from public view.

Yea, you surely are the enemy of fun. (just kiddin' Chris)

Anyways, to reinstate and to clarify a bit of my take on the test that Barcellos did (which is actually a good test) -- as you can hear, the compressor searches for a signal and hence pushes the BG noise to an (to me) unacceptable level and then slothfully compresses the signal when he speaks, as any dodo Radioshack equipment would do. The limiter might be better than expected, I dunno (hard to tell since it all depends on what Barcellos fed from the DXA).

So obviously, what I'd like to see in the next revision is some onboard adjustments for the audio (although still utilizing a pad limiter) and preferable balanced stereo (mini jacks).

To haul along a double system without time-code (as suggested by some here) is just silly, and to me reflects bad engineering. It's like mending it with duck-tape instead of welding. And this especially since the camera already has the capacity to record decent audio.

So no, I have not yet gotten my 5D in the hopes that those obvious snags are rectified.

-- peer

Is it just me, or shouldn't we all be begging Canon not to add audio/video features to the 5D, but rather shouldn't we be begging Canon to put the 5D features into a proper video camera?