[SOLVED] ML-equipped 5D M2 w/AGC and gain off - low BeachTek input volume

Started by cudak888, October 16, 2013, 05:00:53 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

cudak888

EDIT: Turns out to be an oddity with the BeachTek DXA-SLR; has nothing to do with ML; hence my marking it as solved. Granted, there remains a question about the preamp, for which I'm waiting for a reply from Harry at BeachTek, but for those curious as to a perfect workaround, read on.

I've been Googling about and tearing through the archives here at the ML forum to no avail (though one user in the audio sticky did seem to suggest a similar issue) hence, I am posting my specific quandary:

I'm presently running a Canon 5D Mark II with Magic Lantern. I also have a BeachTek DXA-SLR which I use in conjunction with it, but have had some difficulty attaining good audio levels without running the BeachTek's with the preamp gain on the "HI" setting (15db gain), and the trim controls around ~7 or ~8. Goes for both my cheapo Audio Technica battery-powered lavalier mics, and my phantom-powered Rode NTG2.

I was led to believe it was a problem with the BeachTek, but it isn't. I plugged the BeachTek directly into my computer and monitored its audio levels. As I thought, the BeachTek's preamps are putting out enough range that I can easily go straight past -6db and peak way past 0db if I jack up the trim controls even close to full. The green lamps on the BeachTek correlate this.

Therefore, it would seem logical that if I run the 5DM2 with ML set to 0db of analog gain (and 0db digital, obviously), and AGC off, I should get roughly (granted, 44.1kHz, but adequate in its own right) the same levels as the aforementioned good output from the BeachTek.

Problem is, I have to run Magic Lantern with a full 32db of analog gain (no digital gain, AGC turned off) to get roughly the same audio level as I hear clear from the BeachTek. It doesn't make sense, and introduces the hiss-prone preamps in the 5D which we're all trying to get away from to begin with.

For the sake of curiosity, I went back to 0db of analog gain and split the output from the BeachTek to both the computer and the camera. Whatever value was recorded by Audacity from the BeachTek output was recorded by the camera at a level roughly 20dB lower (-18db from the BeachTek becomes -38 in the camera). With the headphones on the BeachTek, I could jump from monitoring the XLR signal directly from the BeachTek, and jump to the A/V-to-RCA signal directly from the camera and hear the reduction in gain, live.

I realize the camera is limited to recording in 16bit 44.1kHz, but that doesn't explain the wild difference in volume.

I highly doubt that I've missed anything, for I've been up and down the settings, the manual, and the forum (both ML's forum and everyone else on Google) quite a few times.

I figured now is the time to ask the experts. I'm not sure if I've found a problem or not.

Many thanks for any help that can be provided.

-Kurt

dmilligan

Quote from: cudak888 on October 16, 2013, 05:00:53 AM
Therefore, it would seem logical that if I run the 5DM2 with ML set to 0db of analog gain (and 0db digital, obviously), and AGC off, I should get roughly (granted, 44.1kHz, but adequate in its own right) the same levels as the aforementioned good output from the BeachTek.
Why? I see no logical reason. Why should the amplifer/ADC in your computer behave exactly the same as the amp/ADC in the camera? I see no reason they would. It actually makes perfect sense that they would be different.

http://en.wikipedia.org/wiki/Impedance_matching

Isn't the whole point of this box you have to boost the weak mic signals up to the high impedance level of the camera's input, so you don't have to use the camera's crappy preamp? Are you not able to achieve that with this box? Why does it supprise you to have to crank up the output of this preamp to match the high impedance of the camera's input, and make up for not using the camera's preamp? This makes perfect sense to me. I think your +15db switch is there exactly for this purpose.

It sounds to me like you don't really have a problem at all. Are you not able to turn the camera's preamp down to 0, and turn your external premp up enough to get satisfactory levels? If so, I don't understand what your problem is.

cudak888

My mistake - let's throw the computer theory out the window for now.

Quote from: dmilligan on October 16, 2013, 01:35:48 PM
Isn't the whole point of this box you have to boost the weak mic signals up to the high impedance level of the camera's input so you don't have to use the camera's crappy preamp?

Correct.

Quote from: dmilligan on October 16, 2013, 01:35:48 PMAre you not able to achieve that with this box?

I am not, for with the BeachTek cranked up to full 15db gain and the camera's analog preamp turned off, I have to drone into my lavalier (battery assist) or NTG-2 (phantom, of course) not more an inch from my mouth before I get levels in the -20db range on the camera.

I should note that the BeachTek's green lamps turn are turning red occasionally, which means that it sees clipping when the 5D sees around -20db.

Quote from: dmilligan on October 16, 2013, 01:35:48 PMWhy does it supprise you to have to crank up the output of this preamp to match the high impedance of the camera's input, and make up for not using the camera's preamp? This makes perfect sense to me. I think your +15db switch is there exactly for this purpose.

What surprises me is that cranking the BeachTek preamp to maximum amplification - 15db with trim all the way up - still doesn't match the impedance of the camera.

Quote from: dmilligan on October 16, 2013, 01:35:48 PMIt sounds to me like you don't really have a problem at all. Are you not able to turn the camera's preamp down to 0, and turn your external premp up enough to get satisfactory levels?

No.

-Kurt

dmilligan

Quote from: cudak888 on October 16, 2013, 03:46:14 PM
I should note that the BeachTek's green lamps turn are turning red occasionally, which means that it sees clipping when the 5D sees around -20db.
What the clipping level is, is completely arbitrary and depends on the amplifer circuit design. Decibels are a relative quantity, relative to the reference voltage of the ADC or maxium voltage the amplifer circuit can handle, these are arbitrary design choices. Even if the clip voltage of your preamp is the same a the maxium voltage of the camera's ADC, the impedance in the circuit between the output of your preamp and the input to the camera's ADC is going to cause a loss (or gain depending on the circuit, like if there is an amplifer in the circuit; if the circuit is passive, then there can only be a loss), which means you can easily clip on your preamp and not on the mic input. An impedance mismatch between the preamp output and camera is going to make the loss worse.

Perhaps you do have a defective unit. Have you tried plugging your mic directly in to your camera looking at the signal, and then do it again with the preamp to see how much gain the preamp is actually giving you? Have you tried it on your camera without using ML to make sure it's not a problem with ML?

Perhaps this preamp is just not capable of giving you enough gain that is neccessary, though I would think it should have been designed to be capable of giving you plenty.

cudak888

Quote from: dmilligan on October 16, 2013, 04:24:11 PM
What the clipping level is, is completely arbitrary and depends on the amplifer circuit design. Decibels are a relative quantity, relative to the reference voltage of the ADC or maxium voltage the amplifer circuit can handle, these are arbitrary design choices. Even if the clip voltage of your preamp is the same a the maxium voltage of the camera's ADC, the impedance in the circuit between the output of your preamp and the input to the camera's ADC is going to cause a loss (or gain depending on the circuit, like if there is an amplifer in the circuit; if the circuit is passive, then there can only be a loss), which means you can easily clip on your preamp and not on the mic input. An impedance mismatch between the preamp output and camera is going to make the loss worse.

Makes sense.

Quote from: dmilligan on October 16, 2013, 04:24:11 PMPerhaps you do have a defective unit. Have you tried plugging your mic directly in to your camera looking at the signal, and then do it again with the preamp to see how much gain the preamp is actually giving you? Have you tried it on your camera without using ML to make sure it's not a problem with ML?

Perhaps this preamp is just not capable of giving you enough gain that is neccessary, though I would think it should have been designed to be capable of giving you plenty.

I haven't plugged the mic directly (will try that right now), but I did wind up connecting connecting the camera's input to the BeachTek's headphone output, rather than the BeachTek's intended to-camera ouput.

Voilà - the output impedance from the headphone jack could be matched in the camera with a bit of fiddling with the volume knob. The primitive V/U meters on the BeachTek (yes, I realize clipping is arbitrary, but it's good enough to know whether the camera should be getting a very hot signal or not) could be easily with those on the camera with no additional gain necessary. I recorded a clip and checked the audio in Adobe Audition for noise - it is clean, and what I expected given my trim settings.

Now, I'm no great audio expert - as has been undoubtedly noted - but it seems as if the intended output jack of the BeachTek sends out a severely attenuated signal, and the headphone output gives access to the full range of the same signal.


Let me know if I'm completely off track or not.

EDIT: I connected for mic levels (lavalier connected directly to the camera with it's own power source) directly to the camera. Levels were too low - that mic needs amplification to line level.

-Kurt

1%

I don't have this prob with a few preamps/mixers I use but I notice that the external mic on AK4646 is real quiet... how does it work with +6db digital gain to each channel (both channels at once), what happens to the noise floor since the beachtek is attenuating -20DB... IMO the "hiss" from this chip is mainly raised floor or uneven digital gain.

Also should add.. do like 10db analog and +6db digital.

dmilligan

Whoa, I just looked this thing up, $285? That's way too much for something as simple as basically just a preamp. You could get a Zoom H4n for that and be able to record much better quality, even do 4 channels (2 built in + 2 line in). Even if it was an incredibly high quality preamp that's worth that much, you're throwing away all the quality by recording with dslr audio input. I have a H4n myself for when I don't want to lug around my giant 4U recording rack, or when I need to be discreet.

Send it back! get a H4n (or the new H6)!

Syncing external audio is not nearly as hard as people make it out to be, and you can use RAW video and still have sound!

1%

If its the powered preamp then its required.. and I dunno about this talk of 44.1... all canons are 48.1

But yea, the recorder will let you record with raw, not just no video + H264.

cudak888

Quote from: 1% on October 16, 2013, 05:24:41 PM
I don't have this prob with a few preamps/mixers I use but I notice that the external mic on AK4646 is real quiet... how does it work with +6db digital gain to each channel (both channels at once), what happens to the noise floor since the beachtek is attenuating -20DB... IMO the "hiss" from this chip is mainly raised floor or uneven digital gain.

Also should add.. do like 10db analog and +6db digital.

I haven't tried adding digital gain to each channel from the camera, as I see no reason to add digital artifacts to the noise floor when the analog setup can compensate for it.

If nothing else, it's a problem with the BeachTek's audio output.

Quote from: dmilligan on October 16, 2013, 05:30:45 PM
Whoa, I just looked this thing up, $285? That's way too much for something as simple as basically just a preamp. You could get a Zoom H4n for that and be able to record much better quality, even do 4 channels (2 built in + 2 line in). Even if it was an incredibly high quality preamp that's worth that much, you're throwing away all the quality by recording with dslr audio input. I have a H4n myself for when I don't want to lug around my giant 4U recording rack, or when I need to be discreet.

1. Bought it used.
2. All those extra steps are easy to throw around when quick workflows are not a consideration.
3. Not everyone uses ML for RAW. While I am filmmaker by hobby, it doesn't pay the bills, nor does it improve my workflow. People who have no idea of the tech problems behind using DSLRs for ENG-style interviews and documentaries request exactly such setups. So be it; they pay well.

That said, seeing as the problem has nothing to do with ML, I'm going to send Harry at BeachTek an email to see if he can explain the reason for the extensive attenuation on the intended output signal.

Thanks all.

-Kurt

1%

I'm gonna mess with the de-emphasis filter and dgain and see if that fixes the floor, right now its not too useful.


cudak888

Harry at BeachTek just wrote:

QuoteWe designed the DXA-SLR adapter for use with the Canon firmware – not the Magic Lantern. With a setting of 25% gain in the camera, the output level of the adapter is perfectly matched for the camera. Note that the output level is not line level – it is an amplified mic level signal which is about -36 dBu.

I would not recommend that you connect the headphone output to the camera. While this will give you a much stronger signal, the headphone amplifier will compromise your signal. The signal from the OUT jack goes through low noise amplifiers and is much cleaner than the headphone output.

The DXA-CONNECT appears to be the best unit for use with a 5D running Magic Lantern. Can't wait :D

-Kurt

1%

Lol, they told me the same thing when my DXA was feeding gnd through the tripod screw and introducing a hum... yay buy something else. Sure...


cudak888

Quote from: 1% on October 16, 2013, 10:06:17 PM
Lol, they told me the same thing when my DXA was feeding gnd through the tripod screw and introducing a hum... yay buy something else. Sure...

We'll see how the shoe design of the Connect fares. I have good feelings about it.

As for your unit - did you swap from G2 to G1 to compensate for the ground loop?

-Kurt

1%

I have a 4P, all analog .. I put E-tape on the screw and the ground loop is gone  :) Not a prob on SLR, just HDV