Distorting my AT4073a?
 

We did some production work yesterday and I recorded it with an AT4073a going into an XLR "Y" into left and right of my Marantz PMD660 w/ the Oade Advance Mod. What we have is a club scene where a "squable" happens... the guy starts yelling (no music playing) and some friends pull him away... anyway here is the wav from my channel that NEVER came close to clipping (set one lower as safety). I had the mic a good 4-5 foot over head but still it distorted pretty bad... Needless to say, I'm not very pleased with this at all. Is this normal, is something wrong with my mod maybe? The mic specs say "Maximum input sound level 126dB SPL, 1khz at 1% T.H.D." The guy was yelling but I don't know if I would think he was that loud!

Caution, strong language in this clip!
1.32MB
http://gilligan.kicks-ass.org:1023/WA-Distortion.wav

Could it be the Marantz? I recently bought one and noticed a lot of clipping with, get this, a Sony lav mic. I could not get a decent level. I don't know if it was the mic or the Marantz. I was too depressed to continue...

Well, the Oade advance mod can even take line level signals well so I don't think so and if you d/l that wave you will see that it's VERY low(-7.7db peak). A buddy of mine said with cheap mic cables/faulty ones and phantom power you can have some things like this happen... any else agree with this?

I thought this was supposed to be dialog
and not an argument with your wife.

Anyways, sounds to me like the Marantz
input is clipping. The 4073 is one of the
hottest mic's around. Maybe put a 20dB
attenuator in the line.
MHO.

Dave nailed it. Just to clarify what he said, the output voltage of the mic clips the input circuit of the Marantz. You can turn down the recording level of the Marantz and record the clipped signal at a low level, but it's still a clipped signal.

Your owner's manual specifies an input sensitivity of 1.2 milivolts. But at shouting levels, the 4073 puts out closer to 70 milivolts.

Regarding the input sensitivity of 1.2, does this mean
a recommended average, or is this the max?

Also, how did you get that figure of 70?

Since you're using a "Y" cable are you sending phantom power into the mic from both inputs? Do you know for certain how the cable is actually wired?
Both of these could be a source for unexpected problems.

I believe that such a specification is usually the nominal voltage, the one that will give a good recording level (- 12 dB or so) with the level adjustment set to mid range. I expect the ciruitry to handle at least a 40 dB range centered on the nominal. I keep meaning to ask a manufacturer if this is what they mean, but have never gotten around to it. But I've never run into trouble acting on those assumptions.

AudioTechnica specs the AT4073 at 71 milivolts per Pascal. A Pascal is about 94 dB of sound pressure. According to various tables of typical sound pressures you can find around the internet, I estimate that a strong voice shouting at about a meter away is close to 94 dB.

Have any idea what a typical consumer line level
input sensitivity might be spec'd at?
Say, for example, a minidisc line in
jack?

That's what I'm thinking. Y-cords are generally NOT a good idea for audio.

Regards,

Ty Ford

I agree, but I have read to feed your mic to both lines as a safety for clipping and then I looked around on proper ways to do it for the sake of phantom power and everything I read acted like a straight Y was just fine. *shrug*

This is NOT a stock Marantz, this has the complete preamp section ripped out and replaced with a smoking preamp from oade brothers... they say you can run line levels into the preamp with out a concern for clipping so I'm not sure if the AT is too hot or not :(

What I'd like to know ... is 71 milivolts/Pascal
a line level signal?

Don't know but the AT4073a can't be putting out a signal hotter than line level?! Could it?

Nominal operating voltages are 1.228v for the pro +4dbU line level standard and 0.316v for the consumer -10dbV line level. So 71 millivolts is well under either one.

Any idea what putting a 10dB pad in the
line right after the 4073 would give
as an "effective" sensitivity rating
for the 4073/pad combo?

1. Whether it's stock or not, if you are abusing the laws of impedance, you're asking for trouble, and you're asking for trouble.

2. The at4073 has one of the highest outputs among shotguns.

3. A shrug mean, "I don't know" where I come from. The right way to feed two inputs is with a mixer.

4. Good luck


Ty Ford

Kevin, the reality of sound recording is that it is often difficult to get right. Expensive equipment is subject to most of the same rules as cheap equipment. I've listened to your clip, and looked at its waveform. You have clipping as well as room reverb. I believe that because of the limitations of the situation, which was staged for novelty in the midst of the audience and not recording, you would have done as well or better with a well chosen $150 mic going into a $200 recorder.

As Steve says, the line level capability claimed by Oade Brothers is 316 mV with lots of "headroom," at term you don't normally associate with digital recording. Without more details, I'm skeptical. But even if you are getting that performance, the AT4073 is capable of 2400 mV before it clips itself, so I still think it creamed the Marantz input with the screeches from your actor. Send Oade Brothers the clip and see how they explain it.

If the sensitivity of the microphone is 71 mV/Pa it will drop to about 32% (square root of 10) of this or about 21 mV/Pa with a 10 dB pad installed.

Well, I was brickwalling the preamp... a quick email to Doug Oade cleared that up. Apparently you are supposed to have the software selectable -20db pad on at all times unless you can't get enough gain. This was contrare to the original Marantz circuit that sounded like poo if you did engage the -20db pad.

Just did some quick test and everything is perfect and I couldn't be happier again... well other than figuring out what I will do about that dialog that can't be reshot!

I wasn't insinuating that because it's not stock I can abuse anything... I was answering other people in the same msg (about being too hot, though I was incorrect anway). The *shrug* meant I don't understand why everything I read said to split the signal and when I read about Y's it seemed like no big deal... but this ended up not being the problem so I guess I'm "safe" doing it this way. The beauty of the marantz is not being teathered to other things, if I go with a mixer now I got more stuff strapped on me and power cords and such.

I'm all ears if you have any other suggestions or corrections to statements/assumptions I've made... I've done audio for years but it's always been band production stuff where everything is pretty straight forward... funny how 20+ mics and I'm at home but give me one mic to use and I'm screwing things up ;)

Not to be putting words into Ty's mouth, but his point still stands. While you might get your "Y" cable to work, the better way is to use a mixer, especially when you're dealing with sending phantom to the mic. Something like the Sound Devices MixPre or 302 would be ideal and they are small, lightweight, and battery powered so power cords etc wouldn't be an issue. Their audio performance is very high quality, up there among the very best professional tools.

I'm all about doing things right or not doing them at all but at 700-1600 bucks just to split my mic into two channels of a $750 recorder what is the real harm in just using a Y?

I understand if it were money no object but as we all know money is an object and for example this short is being filmed on ZERO budget and actors doing it for free as long as we pay thier plane tickets. I'm the producer/grip/AC(apparently)/electritian/AD oh, and when that is all done they go "ready for sound?" and I'm like umm where is my gear at anyway as I've been having to do everyone else's job. And trust me, it's not that I WANT all those credits as I will probably only credit myself as producer and sound it's just with such a small inexperienced crew you have to wear many hats. Honestly I wish I could trust someone else to set levels and do sound so I didn't have to do that!

Unless someone can explain to me a huge short coming to the Y cable I can't justify that expense at this point.

Don't get me wrong, I am VERY greatful for the advice!

I had to pass three FCC tests to get into radio a long time ago when you actually had to know something about electronics.

Impedance-wise, it's the wrong thing to do.
Regards,

Ty Ford

I do understand such concepts... my father is the chairman of the board for the state issued FCC liscences for working on Satellite and TV's so electrics, ohms law, and impedance matching aren't unfamiliar terms. I just don't know the specifics when it comes to mic inputs as you are making this out to be a dreadful thing.


Kevin

Well, I'm glad Doug was there to straighten that out--if you had listened to Dave in post #4 you might have thought you had brickwalled the preamp and needed -20 dB of attenuation ... 8>)

..and Morse code at 14 words per minute, right? Was that 1rst Class Radiotelephone?

There was no CW test for the 1st. That was for a Ham license.

Regards,

Ty Ford

I'm not an engineer so perhaps someone will correct me if I'm wrong but what I suspect happens is this...

The mic is the source driving a load at the inputs. Using a "Y" cable connects the two inputs in parallel. Connecting two loads in parallel halves their effective impedence when viewed from the source - ie, putting two 8-ohm speakers in parallel to an amplifier effectively puts a 4-ohm impedence load on it and doubles the output voltage it must develop. I would expect the same thing to happen with the mic - the 4073a is rated as 100 ohms and expects to work into a high impedence mic input. Driving a lower impedance means the mic's internal electronics must produce more power than they were designed to deliver, which could change its behaviour considerably.

None that I know of. I guess there is some experience of associated problems out there somewhere, and that's plausible, but I don't recall anyone reporting any specific problem around here in the past year or two. Personally, I think you'll be fine. Now that you know what your main problem was, it's worth a try. Please let us know how you make out.

Well, if both mic inputs are feeding phantom into the Y cable that could be as much as 96 volts to the mic. Or, it could be a lot less, depending on Marantz & Oade.

But it's probably way more than that little capacitor wants to see.

Does the Marantz allow you to turn off phantom individually per channel?

Lots of people record a second channel at -12 to -20db from their first for backup. No engineer I know uses a Y cable to do it. If the device doesn't do it natively (like a lot of sony cams), or no mixer, why risk good audio on both channels to impedance mismatch and double phantom to make a backup? Two bad channels is not twice as good as one good channel.

Just do a good job on the first channel. Use the mic like it was designed to be used. In my opinion that's appropriate risk management. Most people who stick with audio professionally develop a "best practices" sort of approach. Using a Y increases risk. Not having a backup channel? That risk can be managed by good engineering.

Two 48 volt sources in parallel supply 48 volts, so the smart money says the mic will see 48V.

And in terms of signal voltage, we're trying to distribute voltage these days, not power, so impedance mismatches are less significant. For transferring voltage, you just want the load to have a lot higher impedance than than the source. Besides, the 4073 has voltage to spare.

I don't disagree that this is not best practice, but I think the odds are in Kevin's favor.

Phantom POWER is POWER, not just voltage. Power = Current x Voltage. We already have existing examples of distortion due to low milliamp capability of some Phantom Power supplies.

I'm bemused by the "how to get by with less" attitude exhibited here. I once had a client suggest that he reduce his ad budget by 10% to see if his sales went down 10% or hopefully less (there's at least one in every crowd.) I suggested he shoot one toe off and see if it had any effect on his abilty to walk properly. If not, shoot another!

If you're doing this for yourself, it really doesn't matter. If you are getting paid as a professional, it does.

Thanks,

Ty

Thanks A.J. And am I guessing correctly that the
drop with a 20dB pad would be the square root of 20
i.e. 4.5% of the mV/Pa?

No, -20 dB is a drop to 10% of the reference level. dB is 20 log (new level / reference level). Going the other way, the ratio of new level to reference level is 10 raised to the power (dB / 20).

So for the earlier example of -10dB that was 10 raised to the power of -10/20 or 10 raised to the power of -1/2, which is 0.316 (not actually the square root of 10. +10dB is the ratiothat happens to be equal to the square root of 10).

For the case of -20dB it's 10 raised to the power of -20/20, or 10 to the -1 power, or 1/10

Right. Parallel, not series. Thanks for the correction.

***edit*** why did the forum server make 2 of these?

I knew it was possible I just misunderstood the mod and there is no addendum added to the package to clarify things... with a stock Marantz the -20db pad is useless as it just makes things so noisy... I thought they got rid of the pad with the mod so that's why I wasn't sure what the problem was.

But yes Doug cleared it up, they are a great company making some great products with very good support.

This is correct which is why it's not a problem for supplying voltage... as far as current is concerned, I am combining it yes but the mic will only draw what it needs at least in this case. That is why when you add batteries in parellel to supply something it just powers it for longer as it still draws same amount of current and you are just giving it a bigger current source to draw from.

The voltage factor for 10 dB is sqrt(10). In this case as you are padding (reducing) level you divide by sqrt(10) = 3.1622... Thats equivalent to multiplying by 1/sqrt(10) = sqrt(10)/10 = 0.31622.... I wasn't very clear on that.

There is no code requirement for any of the commercial radiotelephone licences. The commercial radiotelegrah licences did have a requirement. I think it was 25 wpm random groups of 5 for the first class.

Impedance matching isn't too critical at audio frequencies as is clear by the fact that the sources (microphones) tend to have low output impedances and sinks (mixers, cameras, recorders) tend to have high input impedances. Thus it's clear the mic is intended to serve as a voltage source. So if, for example, the mic has an impedance of 100 ohms and a mixer 5900 (to make the math easy) 5900/6000 = 98.3% of the microphone's open circuit output voltage would appear across the mixer's input. If two mixer channels are paralleled (with a Y cable) the mic is loaded by 2950 Ohms so that 2950/3050 = 96.7% of the mic voltage appears across the input. This is a very small loss (0.14 dB) and all should be fine assuming that the Y cable is properly wired.

True - I was more concerned that it might have some adverse effect on the mic's own internal preamplifier and impedance transforming circuitry that might make it more susceptible to clipping in the mic.