modifying a digital camera mic (Linkwitz) - my next move - like-minded advice sought?

  • Thread starter Thread starter murrayatuptown
  • Start date Start date
M

murrayatuptown

New member
Hello:

I hope this is interesting to someone other than me. I hope I can pick up some people here who can give some thought to the impedance questions below.

I have been experimenting with ways to improve horrible audio in digital camera video. I have recorded audio at live performances and only from the back of the room where overload was prevented did I get usable audio. I also have recorded with a separate Tascam DR-08 but have had great difficulty resyncing the removed audio. I have a software method that works, but it has terrible sync, and no responses on the recommended forum or from the software author who offered his address.

So I've moved back to trying to record decent audio in-camera.

After 3 or so approaches to the 'Linkwitz mod' on the internal electret mic capsule, I finally added a jack and separate 9V battery power. I have accomplished overload-free audio, at least so far, but have the problem of omnidirectional pickup...in a noisy bar, coffee shop, or restaurant, the chatter of patrons is really loud. An iPhone that impressed me from back of the room overloads up close (roughly 10 feet from instruments), and the camera does not.

It seemed ridiculous earlier, but since I have an outboard electret wired to a jack, I may as well add an XLR socket next. I have several percussion dynamic mics that are reasonably broad in response, but tend toward lower sensitivity. I have two that are -56 dBV and -66 dBV. This is quite a bit lower than any electret condenser capsule. One of the candidate mics is a cardioid and the other a supercardioid. I bought them because they were cheap. The low sensitivity and odd mounting bracket also make them 'orphans' for conventional vocal/guitar mics needs.

What I plan to do is probably capacitor couple the signal so I am certain I don't apply DC to the mic element. Despite the general rule that phantom power won't both a dynamic mic, there is little conventional about my wiring. For me to end up with a foolproof scenario that works with both the electret or the dynamic, I am going to play it safe.

One mic is nominally 300 ohms, and the other 600 ohms. The camera input impedance as best I can estimate (Thevenin calculations) is dominated by the bias resistors...With the additional external resistors I used for the electret, I have about 1.8 k looking into the camera. Based on this, I used a pair of 50 uF electrolytic caps back-to-back for 25 uF. I did this because I didn't want to shop for non-polar electrolytics, and the polarity of DC
across the mic coupling cap does reverse if the 9V external battery is connected and the camera is turned off. For now, the electret mic JFET is wired as a source follower with 2.2 k resistor...I'm not where I intended to be but am heading down a different path based on the noise and directional pickup.

My main question is if I feed a 300 or 600 ohm dynamic mic into a 1.8 k input through the 25 uF capacitor, do I need to 'load' the dynamic mic with a 300 or 600 ohm resistor, or do I really want something higher than that anyway since I don't need to maximize signal 'power'.

Extending my thoughts to other options, I thought I might use a low-to-hi-z XLR-1/4" transformer - to solve a number of 'problems'...I think...the step-up would solve some of the sensitivity difference, but mismatch the mic stepup transformer's secondary looking into 1.8 k camera impedance. I have taken enough risk dissecting the camera...changing the internal bias resistor is too risky. This may be a problem for the dynamic mic, but it is what it is.

Next thing if results are still good, a small mixer feeding a 14:1 DI box would allow me to adjust the level going into the camera.

The small mixer has phantom power so I could eliminate the battery, wiring and external bias/coupling horror stuck on the side of the camera...it gets a lot of freaked-out responses and will probably never be able to travel in an airport again, unless I like TSA attention. I'm afraid the DI box may be going the wrong direction impedance-wise because I still have a relatively low-Z mixer output heading into the DI stepdown transformer...it may not be too happy feeding 1.8 k either.

So, anyway, if you don't think this is insane and care to reflect on the impedance concerns I've expressed, I would appreciate hearing your thoughts...maybe what is the best option, regardless of mess or inconvenience.

If you DO think it's insane, oh well. That has never stopped me before. Pushing past the 'why bother?' issues is opening up other things I might never have gotten around to trying.

A separate topic might be the challenge of syncing external video in software.

Even if I end up with low levels, it's better than overload. I have exported audio from a video, saved the video with NO audio, post-processed the audio in Audacity, and re-muxed it into the video with usable results. At least I had compatible audio & video formats and knew up front they would 'fit' the original timing. It's frustrating to have external-source audio grossly out-of-sync with the video despite trimming to the same length within 10 milliseconds, and having the software time-shift settings ignored or overridden by some unseen parameter.

If anyone has their own questions instead of answers to mine, I guess join in!

Maybe I am the only one who would do this instead of buying another camera, Zoom or whatever. Can't help it...

Thanks
 
WHEW!!!!


How about just using a better external mic, plugged into the camera.
If the camera has no option for this....you need a more "pro-sumer" camera then.

AFA as synchronizing when using a separate audio recorder....that's pretty simple.
Record at 48kHz wit the standalone audio recorder...and also record the audio in the camera with the camera mic, as crappy as it is.

Then pull the video into something like Vegas or whatever video editing program you have, and also transfer the audio from the standalone recorder into the video editor.
Then just line up the two audio tracks and delete the one from the camera.
Done.
Just make sure you have the sample rate set at 48kHz....they should line up 1-2-3
 
Miroslav - thank you for reading the post. Most people probably wouldn't bother ;O)

It's a point & shoot that has decent video and died once on a trip, so opening it up was a nothing-to-lose situation. I was amazed I got it back together still working, not once but twice! All previous camera surgeries I have attempted ended in expiration of the subject :O(. On the second try, I dragged the ECM and wiring outside because another disassembly could be the last.

External mics are exactly what I'm attempting, but it's the homebrew interface I have to work around...out of necessity.

Other than the low impedance concerns, I think an external mic and a mixer have a lot of potential, and I just like doing weird stuff like this.

I also have a cheap shotgun mic (Chinese, <$30 delivered +O) I took apart to consider possibilities. I've heard nothing good about them other than voice in a hurricane. Mine has two electrets spaced (collinearly?) with two active bipolar transistor circuits I haven't deciphered yet...ugly circuit board, illegible components, all running on a 1.5 V battery (not phantom-capable). I had a one-time opportunity to run it with a B&K sound calibrator and Audio Precision ATS-1 and it was horrible...minimum 14% THD and only got worse at higher levels. I recorded with it into the DR-08 and it was pretty sad...ripe for a complete Franken-mic project...once I decide what I want. Maybe pull one of the electrets, replace the existing built in distortion effects and try to make use of the back openings...or reverse one of the electrets and see what it does differentially...the backward one capturing the back-opening sound and subtracting it. Hey, it's good for nothing else at the moment...even the plastic carrying case is awful ;o0.

Regarding bit rate, I recorded with the DR-08 at either 44.1 or 48 kHz (96 wasn't going to improve the fact that I was at the back of a theater), but the camera stream info says the audio is mono 16 kHz, 128 kBps, 2:1 mu-law.
Last time I re-muxed post-processed audio, I saved it as close to the original stream as possible...I felt it shouldn't matter but the re-mux software kept telling me there was 0.5 second interleave. Trying to match stream format was a desperate approach to solve the drift problem...it was seconds off and ignored and settings allowing a/v time shift.

Sure a better camera would be cool, some day...but I apparently have this problem doing things the easy way ;O).

I'm learning a few things (like what to not bother with, empirically determined, instead of being told 'you can't do that'), and keeps me off the street (except when I'm recording
street musicians).

It's still slightly satisfying to not have wind noise or an aggressive motorcycle driver totally cover up the desired content...

Vegas? Ha - what happens in Vegas stays in Vegas...maybe 'real' software would help as well.
 
Blimp with a Boom Pole- Camera shoulder mount 201

I like way the entire scene is covered. Like the video.....

Have a look at
 
With regard to the dynamic and transformers.

If you take the mic as a 150 Ohm source then into 1k8 there is little advantage in more than a 1:2 transformer so only ~6dB "gain" but every little helps and at least you get DC isolation and of course balanced operation (any chance of a shcemo?) .

However, you might want to keep the caps in circuit? Two of them in series with an R to deck to give a second order high pass filter. If you want to capture organ pedal notes you will need better kit anyway!

Personally I would go for a higher traff ratio, up to 1:6 and feed an NE5534 or similar and have a gain pot with say a 30dB range? Note this level will then need to be attenuated to drive the mic input circuit but HAVING the gain to start with is always handy. Note, any such attenuator needs to use very low resistances, the NE will drive as low as 820R and this keeps thermal noise low.

Dave.
 
You're recording in a room that is so loud the electret bottoms out and distortion wins. You idea to reduce the sensitivity, or just use your drum mics is technically sound - but the audio quality will still be pretty poor, even if your camera records it faithfully. Camera mounted mics, when working perfectly, within their and the preamps capabilities often just let you hear how dreadful it sounded at that place in the room. You will still have wooley, probably distorted and very un-professional sounding audio. Your mods are great ideas - and fitting an XLR, and the pad switch make sense if the camera has no real value - but they won't make it sound good. In most cases when sound is that loud at the back of the room, then it is so loud at the front and on stage that the poor tuning of the instruments and vocalists get concealed. Your mod may well let you hear it. A proper feed from the desk, if unavailable to you, robs you of any chance of a decent product. Great for the crowd noise between numbers, but unless you really enjoy the electronics, then maybe it's a bit pointless. Shotguns in big spaces like this sound very weird as you move the camera - thin weedy and phasey. A simple cardioid dynamic with a pad is the most simple, and probably the poor frequency response doesn't really matter that much.
 
You must log in or register to reply here.
Back
Top