Need to identify this hiss

[Brian Maurer]

Hello all - I'm looking to identify this source of this hiss, and hope that someone here can help!

The tools:

Canon 550D
Sennheiser MKH416P48
BeachTek DXA-6VU

The audio:

http://www.mathmissionfilms.com/audiohiss.mp3

The problem:

As you'll hear, there's a very distinct hiss that happens each time I record video using this setup. I am nearly positive that the problem lies with the BeachTek box, but as I do not own any other recording devices that can phantom power my mic, I have no way to test to make sure. The box has a ground switch on it, but on either setting, the hiss remains. If I adjust the gain, the hiss remains. Even when not recording, I can hear the hiss via headphones. I've had to use Soundbooth to isolate the hiss, and remove it, but as one would guess, that does terrible things to audio in different environments. I've read and used a Zoon H4N with a friend (his tech), but haven't established if it has the hiss as well; dreading that it's a problem with the mic, and not the box. Anyone with thoughts?

[Richard Crowley]

So it seems easy enough to do the differential diagnosis experiment where you record "silence" with just the camera, and then with the BeachTek box plugged into the camera (with any controls in minimum,and then maximum position, etc.) And then with the mic plugged in but buried in the middle of your bed inside a big pile of all the blankets and pillows you have in the place (to make something as close to acoustic "silence" as possible.)

Seems unlikely that anyone can diagnose the problem with a simple, compressed data sample with no variables to compare. Remember also that camcorders in general, and DSLRs in particular, were NOT designed to be decent audio recording devices.

[Brian Maurer]

When I get home this evening, I'll go through those steps and repost the audio. In the meantime, I found a single sheet of paper "manual" that came with the BeachTek box. Under its 10 points of interest, number 7 and 8 are the most interesting with respect to the noise:

7) Ground Lift Switch - Set the G1/G2 switch on the DXA-6vu to the position that gives you the least amount of noise. This switch allows the input and output grounds to be isolated to prevent ground noise on some Sony camcorders.

8) XLR Inputs - The two XLR inputs can accept balanced or unbalanced connections. To convert the input to unbalanced, simply ground pin 1 to 2 on the XLR cable. We recommend the use of sensitive condenser type microphones for the best results. Dynamic micropphones usually do not have sufficient output levels and may result in an excessive amount o fhiss from the high level of amplification necessary by the camera's audio preamplifiers. Microphones with a sensitivity rating of greater than -40 dBV are ideal.

Not sure if that helps at all...

[Steve House]

The 416 is the type recommended and not noted for being noisy so barring a malfunction I'd expect the hiss isn't coming from there. The Beach is not a preamplifier so switching to something such as a JuicedLink or a field mixer where you can add some signal boost before sending it to the camera would help. I'd suspect that what you're hearing is the result of the noisy preamplifiers in the camera that virtually all video capable DSLRs are inflicted with.

[Dan Brockett]

I agree with Steve. Sounds like normal noise floor for a DSLR to me. You have to realize that all DSLRs and almost all video cameras have horrendously bad audio quality in general. Any audio professional listens to them and just shakes their head, that is why using something high-end like a Sound Devices recorder will always yield a superior result to any camera audio.

I have been testing the Juicedlink DT454 with my 5D MKII and FWIW, the noise floor is not that much better than your sample posted above. DSLRs may work fine in certain situations but when audio quality is important, it would behoove you to use a high-end, quiet recorder like an Sound Devices 702, especially when you have a nice quality mic like the 416. You get what you pay for and an $800.00 camera isn't going to have good quality audio, but neither is a $100k video camera, they all sound terrible. You should hear how bad the audio quality is on the RED too.

Dan Brockett

[Richard Crowley]

The ground lift switch on the BeachTek is unlikely to have ANY effect on the hiss you are asking about. OTOH, if you had a HUM issue, then ground-lift switch would be indispensable. Of course, you can do the very simple experiment of flipping the switch and see if it has any effect on your hiss (I would bet that it does not.)

Now, it is possible that you have a low-output microphone and/or you are using it too far away, and thus producing a signal level that is below what would be "normal". And the BeachTek docs are warning that they can't do anything about that since it is beyond the scope of what the BeachTek adapter was made for.

To be perfectly honest, I thought your sample sounded quite good considering it was recorded on a DSLR. Your issue may simply be unrealistic expectations.

[Chris Barcellos]

This sounds like the normal AGC (automatic gain control) issues we have all had with the Canon DSLRs, and others. If you can turn off AGC, and adjust your mic levels going in from a preamp aperatus, you can get much cleaner sound. As I recall, we don't have the capability of turning it off on the T2i. On the Canon 5D you do. With my Juiced Link CX 231, before we got control of sound in Magic Lantern on the 5D, I had constructed a tone to feed to one channel with an mp3 player, that defeated the automatic gain. Then on the other channel, I could feed my mono mic testing for level adjustement. This actually worked nicely and I could find pretty good recording zone. Of course, you were limited to one track.

[Brian Maurer]

Originally Posted by Richard Crowley

Now, it is possible that you have a low-output microphone and/or you are using it too far away, and thus producing a signal level that is below what would be "normal". And the BeachTek docs are warning that they can't do anything about that since it is beyond the scope of what the BeachTek adapter was made for.

To be perfectly honest, I thought your sample sounded quite good considering it was recorded on a DSLR. Your issue may simply be unrealistic expectations.

I'm using a Sennheiser MKH416 P48

I've recorded sound to the DSLR, as well as a Canon HF100 and directly to my computer via the mic input in the back; all have resulted in the exact same hiss. Leads me to believe that it's not the camera, and that it's a problem with the BeachTek box. Flipping the ground switch doesn't do anything.

I don't think I'm expecting something unrealistic; unless this is an attribute of the microphone, the simple solution is to dump the box, get something like a Zoom H4N and just deal with audio that way.

Originally Posted by Chris Barcellos

With my Juiced Link CX 231, before we got control of sound in Magic Lantern on the 5D, I had constructed a tone to feed to one channel with an mp3 player, that defeated the automatic gain. Then on the other channel, I could feed my mono mic testing for level adjustement. This actually worked nicely and I could find pretty good recording zone. Of course, you were limited to one track.

Yeah, limited to one track. Again, I think an answer may just be to get a stand alone recording device, and move forward from there. At least, until the T2i has gain control; not sure if there are firmware updates coming that would address that; doubtful.

[Brian Maurer]

Did some more reading and found a great video that I'd like to share with owners of the T2i

550D T2i AGC Automatic Gain Control Tests

It demonstrates the the AGC is the problem and that it increases the gain based on ambient noise. When the guy starts shouting, you can hear the noise drop down for a full second before it brings itself back up. From the literature that I've read, the T2i and 7D do not allow turning off the ACG. I read about a work around very similar as what you posted, where you flood one channel with sound so that the ACG drops the gain just enough to kill the noise, and record your audio with the other channel.

I'm not comfortable with a hack to disable it; and don't have the tools to pull off a dual channel thing to drop the gain. Guess I'm just looking at an external recorder.

[Jon Fairhurst]

To reduce the hiss, you need a preamp with gain. To keep the hiss from "breathing", you need a device with AGC defeat.

The Beachtek adapter you are using has neither gain nor the AGC defeat circuit. They later came out with a unit that has an AGC killer, and now have one with AGC killer and gain.

juicedLink came at it from the other side: their preamps (the CX series) have clean gain. The new DT454 has even more available gain and two modes of AGC defeat. These amplifiers are exceptionally quiet.

I've tested the Beachtek with AGC killer, but not the one with gain, so I'm not sure how clean it is.

[Richard Crowley]

Originally Posted by Brian Maurer

the simple solution is to dump the box, get something like a Zoom H4N and just deal with audio that way.

The BeachTek box is NOT causing the hiss. The audio path in the BeachTek is passive and does not contribute to the noise level.

I wouldn't bet on the H4N necessarily solving the problem. It isn't noted for stellar mic input preamps either.

[Brian Maurer]

Well, it looks like upgrading to the newer DXA-5Da may be the best bet for now then. Any thoughts on the market value of a used DXA-6VU?

[Jon Fairhurst]

Originally Posted by Richard Crowley

The BeachTek box is NOT causing the hiss. The audio path in the BeachTek is passive and does not contribute to the noise level.

While that's technically true, a passive adapter doesn't help reduce system noise, while a clean, active preamp can.

Think of it this way. Let's say you have a signal that goes through a passive, straight wire, then feeds a cheap preamp that adds 10 "pieces" of noise and then multiplies both the signal and the noise by 100. Now you have "1,000" pieces of noise, along with a full strength signal.

Now imagine a system with a super-clean preamp. It multiplies the signal by 50 right up front and in the end it only adds one "piece" of noise. Next you send it into the cheap preamp. It adds its own 10 pieces of noise (plus 1 from the preamp makes 11) , but you only need it to multiply the signal by a factor of two. In the end, you get 22 pieces of noise along with the full strength signal. 22 is a lot less than 1,000.

The bottom line is that noise in the mic and any other components through the first round of preamplification are critical. Every little bit of that up-front noise gets amplified. If you can go through a squeaky-clean preamp and apply gain right up front, you can avoid much of the noise of a cheap preamp like you'll find in the 550D.

And, actually, the audio chip in the 550D (assuming that it's the same as is in the 5D), isn't all that bad. It's just programmed that way. ;)

[Jon Fairhurst]

Originally Posted by Brian Maurer

Well, it looks like upgrading to the newer DXA-5Da may be the best bet for now then. Any thoughts on the market value of a used DXA-6VU?

The DXA-5Da doesn't have gain. You'll still have noise, as I show here:

5. Canon 5D Mark II Audio Exposed - Noise on Vimeo

I'd recommend the DT-454 (which I've tested to be clean), or the DXA-SLR (which I haven't had a chance to try), or an external recorder. Both the DT-454 and DXA-SLR have an AGC defeat circuit as well as gain.

[Brian Maurer]

Originally Posted by Jon Fairhurst

The DXA-5Da doesn't have gain. You'll still have noise, as I show here:

It more importantly doesn't have phantom power options, so it's useless to me. The DXA-SLR seems the best choice for what I'm interested in doing. The $400 price tag will keep me away for a bit, unfortunately, as I don't have the cash for it (unless I can find a buyer for my DXA-6VU to fight some of that cost).

Just to be clear, you tested the DT-454 with the T2i? I'd really be interested in hearing a clip, if you have one available. No worries if not; just interested in hearing samples before I (try to) buy. Thanks everyone for your help. You guys are wonderful.