DSLR + Beachtek adapter audio issues in post

[Philip Perkins]

A friend of mine who is an excellent filmmaker, who has produced a large number of high-quality productions (incl features), has a side business as a one-man-band producer of small videos for non-profits (...in these times...). He recently asked me to clean up and mix the audio for such a film he had done with his Canon 7D. I was expecting AGC artifacts, which can be dealt with in audio post somewhat, but instead got interviews that no matter how much I reduced the level of the file still seemed to be pinning all the meters. They also included huge numbers of tiny clicks. To see what was going on I exported a single clip to Renovator, and saw what you see below. In this image of the file, time is L>R, freq is up/down and volume level is color, with white being the loudest and black being silence. As you can see there is a tremendous very high volume level pitch running through the whole clip, at about 18.9kHz (and you can see the parade of clicks running through the audio as well--narrow vertical lines). The volume of that 18.9k pitch was so much louder than the voices that reducing the level of the file down to unacceptable levels for the voices still had the meters FAR too hot to pass tech eval. I realized then that without telling me, the filmmaker had used one of those DSLR audio adaptors that uses a mega-loud high-freq tone to paralyze the camera's AGC. But a track edited and mixed with this audio untreated (as done by most picture editors) would cause significant problems down the line for broadcast, streaming, etc. Even with the offending high freq pitch removed the audio still has a very odd quality to it--hard to describe, but definitely distorted and lowfi. The blizzard of click events had to be removed one at a time by hand, to avoid causing further damage to the voices by using an auto declicker plug in. My point in this is that I think this is a very bad idea for how to get sound into a DSLR, as prevalent as it may be. On this forum I know that sentiment very much runs towards using double system audio for sound with these cameras (with good reason), but people will nonetheless still try to cut corners and use the camera audio anyhow. Does anyone know if ALL the adapters of this ilk do this goofy 18.9k thing, or just the BeachTek?

Also, as you can see by looking at the image closely, the system isn't actually completely neutralizing the AGC--you can see the level of the 18.9k pitch "grow" in level with each pause in the dialog.

phil p

[Jim Feeley]

JuicedLink does this on some products. They stick the signal on just one channel of the camera, admit it's a kludge, and suggest it might be time to get/use a camera with manual control:

http://www.juicedlink.com/agc-disable/chapter/0

Kinda nice to see this approach and attitude.

Don't know recall about other such devices.

[Steve Foy]

"but people will nonetheless still try to cut corners and use the camera audio anyhow"

So you have to ask yourself, "Why do people who cut corners expect to get as good a result as those who don't "

I don't want to devalue a perfectly legitimate question or your desire to offer assistance to your friend but seriously, if you knowingly cut corners you should be aware of the limitations of and the consequences of that choice. In any other Industry knowingly cutting corners without regard to the consequences would be seen as negligent.

[Derek H]

I just worked with a "juiced-link" yesterday for the first time, provided by the production company for me to interface with their 5d.

For someone used to using Sound Devices mixer day to day it was... interesting.

I just kept thinking why are you guys doing this to yourselves? I urged them to take the time to sync the backup. I think they will. They seemed to allude to previous experience with 5d audio "issues".

[Philip Perkins]

"but people will nonetheless still try to cut corners and use the camera audio anyhow"

So you have to ask yourself, "Why do people who cut corners expect to get as good a result as those who don't "

I don't want to devalue a perfectly legitimate question or your desire to offer assistance to your friend but seriously, if you knowingly cut corners you should be aware of the limitations of and the consequences of that choice. In any other Industry knowingly cutting corners without regard to the consequences would be seen as negligent.

Actually, in other industries you'd be rewarded for cutting corners to save costs. Look at the auto industry, food industry, building industry etc etc.

As a postie one has no control over this sort of thing--the deed is already done when the files come to me.

But help me--there has to be a better way.....

phil p

[Neil Hurley]

Hi,

It seems Canon have a firmware update for 7D on the way.It includes manual gain control.

www.usa.canon.com/eos7dfirmwareupgrade

Regards,Neil.

[atheisticmystic]

The last three post gigs I've had were "clean up" jobs on sound recorded directly into the 7D.

The Beachtek seems like a bit of a ruse, with the "control knobs" needing to be cranked and set maximally clockwise to achieve unity gain, but being described in the manual as controls for levels... just a passive adaptor... not surprising that so many shooters still struggle with it.

While I appreciate the gigs that the 7D shooters create for me, it has been supremely saddening at times to watch really important material with significant speakers sound so shitty (most recently, civil rights pioneer Vernon Jordan being interviewed about legal trailblazer Jewel LaFontant...tragically off-axis and horrible 7D audio).

I beg these clients to call me the next time they go out BEFORE recording.

Best

[Philip Perkins]

Hi,

It seems Canon have a firmware update for 7D on the way.It includes manual gain control.

www.usa.canon.com/eos7dfirmwareupgrade

Regards,Neil.

I guess in a menu, via that wheel/controller thing? That would be ok for most interviews--better than the high Hz AGC "suppression" thing.

phil p

[Marc Wielage]

The Beachtek seems like a bit of a ruse, with the "control knobs" needing to be cranked and set maximally clockwise to achieve unity gain, but being described in the manual as controls for levels... just a passive adaptor... not surprising that so many shooters still struggle with it.

I think it depends on what level you send it, Steven. My memory with the DXA-5DA is that I fed it -20dB audio straight out of my mixer, and we turned it up about to 1:00 on the knob (putting a -20dBfs tone at about the halfway point, line in). They do tell you in the ads and manual that it's a "Passive DSLR Adapter," so that sounds to me like it's just a simple attenuator. I've only sent it mono, usually with a Y-splitter to both channels, assuming that only one channel gets the high-frequency tone that kills the AGC. (And thank you to Neil above for the info about the AGC defeat software update for the 7D.)

Beachtek does have a "more advanced" model, the DXA-SLR Pro, which I believe has actual mike preamps, auto or manual gain, and better metering. But I agree with you: to me, it's just a device to get scratch audio to camera. Anybody silly enough to try to use a DSLR for real sound is going to be sorely disappointed.

[Steve Foy]

"Actually, in other industries you'd be rewarded for cutting corners to save costs"

Not if the wings fall off !!!

[RScottATL]

The Canon 5D Mark III includes audio meters, a headphone out for monitoring and the ability to adjust levels while recording. Admittedly it is NOT pro audio by any stretch, but when you compare it to the Mark II, it is usable at the very lowest of quality standards. The older products are simply abysmal to attempt to use in any professional capacity. Those who do so should heed the MANY warnings that are VERY available in every professional forum. The old 5D/7D workflow is QUICKLY on its way out. Canon never really meant for it to be a professional video workflow to begin with, but at least they are moving in a direction that gives video producers a "minimal" set of options for how to not ruin their projects.

I'm a big fan of using whatever you need to use to make the products that you need to make. If you can do it with less and your project isn't that demanding, fine. The thing is, if you're used to using professional production products, they can make it very easy to do your job well. If you want to take ill-fitting products in to work, it pays to be doubly versed in the workflow needed to shoehorn them into behaving. Of course I'm preaching to the choir, but fixing an audio file like this manually is going to be an insane waste of time.

The better way is to get a piece of equipment with audio recording capabilities. Like a Zoom H4 for $60. For interviews on a shoestring it should be more than enough. I'd cut out the nonsense of trying to cram audio into a system with no way to monitor the recording and with a VERY bad reputation for ruining projects.

It would also be wise to do a full workflow test when using this kind of system where there are a lot of places that things could muck up. That way, all the mistakes get sorted out with 30 second test clips.

IMO, there is no hope for this project. There is hope for the next one. Just test the workflow and if something is just a dangerous and bad thing to do, find a way not to do it. There are so many ways to avoid recording audio into these junk camera inputs. Pick one and give it another go.

[Philip Perkins]

The Canon 5D Mark III includes audio meters, a headphone out for monitoring and the ability to adjust levels while recording. Admittedly it is NOT pro audio by any stretch, but when you compare it to the Mark II, it is usable at the very lowest of quality standards. The older products are simply abysmal to attempt to use in any professional capacity. Those who do so should heed the MANY warnings that are VERY available in every professional forum. The old 5D/7D workflow is QUICKLY on its way out. Canon never really meant for it to be a professional video workflow to begin with, but at least they are moving in a direction that gives video producers a "minimal" set of options for how to not ruin their projects.

I'm a big fan of using whatever you need to use to make the products that you need to make. If you can do it with less and your project isn't that demanding, fine. The thing is, if you're used to using professional production products, they can make it very easy to do your job well. If you want to take ill-fitting products in to work, it pays to be doubly versed in the workflow needed to shoehorn them into behaving. Of course I'm preaching to the choir, but fixing an audio file like this manually is going to be an insane waste of time.

The better way is to get a piece of equipment with audio recording capabilities. Like a Zoom H4 for $60. For interviews on a shoestring it should be more than enough. I'd cut out the nonsense of trying to cram audio into a system with no way to monitor the recording and with a VERY bad reputation for ruining projects.

It would also be wise to do a full workflow test when using this kind of system where there are a lot of places that things could muck up. That way, all the mistakes get sorted out with 30 second test clips.

IMO, there is no hope for this project. There is hope for the next one. Just test the workflow and if something is just a dangerous and bad thing to do, find a way not to do it. There are so many ways to avoid recording audio into these junk camera inputs. Pick one and give it another go.

We know all that already, but thanks! There is definitely "hope" for the project I mentioned, because I've worked a whole lot of shows just like this in post, we have our ways! You might want to consider checking the many other threads on this forum re these cameras, we've been discussing them since they first came out. BTW, the Canon 5D style workflow is NOT going away anytime soon despite what you say--producers like the "zero line item" budget aspect of the cameras too much, and there are way too many of them out there working already. The diffs to a DP between Mk2 and Mk3 are not really enough to bother with mostly, esp since they are mostly doing double system sound with their cameras already.

phil p

[studiomprd]

" There is definitely "hope" for the project "

but at what cost ?

Perhaps you are doing this project pro-bono for your friend, but then you are paying (at least absorbing) this cost.

if the cost of fixing it in post is less than the cost of doing it right during production, this should continue.

OTOH, if they are tripping over the post production dollars by saving a few dimes in production, then the problem is that these folks must not know what they are doing, friends included (Philip, you make it sound like your friend should have known better!)

[Olle Sjostrom]

I'm starting to get the feeling that Senator is really just a very advanced spambot. At least he's been writing the same posts since I started coming here now close to a year ago...

Regarding the OP, that just looks scary. I'm sure you'll do good things with it though phil!

[Marc Wielage]

IMO, there is no hope for this project. There is hope for the next one.

Wow, that's a T-shirt logo if there ever was one!

[OmahaAudio]

I'm starting to get the feeling that Senator is really just a very advanced spambot.

Nope. Not advanced.

[RScottATL]

Well, if this style of audio workflow is here to stay, it seems post audio should have no problem finding work. Finding pay however may encourage the Zero Line Item producers to invest in the "Zoom Adapter" over the Beachtek. I only see one problem that has an easy solution. That is good news.

[Philip Perkins]

I'm well compensated, thanks. My friend should have known better, he believed what BeachTek etc told him. I've shown him what that thing is doing to his audio, basically making matters worse. I'm trying to get him to get that Fostex bolt-on recorder or a Zoom etc etc. One of those jobs I actually did some location sound for him on, w/ my usual rig. The editor cut the piece with the 7D audio, so I sent the elements back and said I'd be ready to mix the film when they cut in the master audio tracks, which they did. Slow progress, but progress nonetheless.

phil p