Jump to content

How hot do you run your signal into Zaxcom TRXLA3?


Mike H

Recommended Posts

For sitting interviews, I normally run a wired Sennheiser MKH-416 mic into a Sound Devices 702 and adjust input level for max peaks at around -10dB below clipping. I have the 702 limiter on but it is rarely needed, only for sudden loud unexpected sounds (laughs, coughs, etc.).

For field interviews, I use:
DPA 4063 lav mic
- Zaxcom TRXLA3 transmitter
- Zaxcom QRX200 
- Sound Devices 702 recorder.

I recently experienced clipping with the TRXLA3 during an interview. Possibly I somehow incorrectly setup the compressor. I don't want any compression to occur during recording; if I do any, I can do it in post.

Zaxcom tech support's advice was "Set max peaks at -20dB." Well yes, that will probably do it but signal/noise ratio can become a problem at this level, at least in analog systems. I've done studio recording for years and never set peaks that low.

My questions are about how you set up your TRXLA3:
What max peak level do you adjust gain to?
What settings do you use for the compressor (most importantly threshold)?

 

Link to comment
Share on other sites

The comment that jumps out at me is "I've done studio recording for years".  When in the studio, I'll typically track vocal at around -10dBFs into ProTools - with the SSL board can ride the fader for an extra 10dB of "gain in hand" if I need a little boost or cut here or there.  Studio is a controlled environment, artist is glued to a spot in front off the pop filter or on a stool, mic on stand in the perfect position since there's no camera to deal with, and musical performance is typically fairly predictable.  Since your now talking about dialogue recording with lavaliers, totally different ballgame.  Now there's greater chances of being off axis or really "off mic" with head turns, the fact your dealing with spoken voice vs louder singing, when you gain up for that softer level, it does expose you to more risks from dynamic outbursts.  The way that I deal with field dialogue recording is to try to shoot for -10dBFs on my mix bus, but record ISO's somewhere in the -15 to -25dBFs range.  What this means in reality is that I will occasionally get peaks approaching 0dBFs on the mix bus as obtaining -10dBFs "RMS" is quite optimistic given the conditions.

Some notes...

1) when I say -10dBFs "RMS" that would be the person delivering a natural level in normal conditions.  If they are soft because of sensitivity of the subject matter, or moment of reflection, or are intentionally being soft, most likely it will be mixed that way too, so you don't have to force the -10dBFs technical goal in spite of the performance.

2) I personally rely heavily on Zaxcom TRX compression and like you don't use Zaxcom recorders.  I have a limiter set just below digital clip, but also enable compression.  I know a lot of people don't like to use compression, but for me personally, I've come to know and trust my compression settings.  From memory, my settings are -9dBFs threshold, very fast attack, slow release, 3dB knee, 2:1 ratio, no makeup.  In other words the mild ratio combined with soft knee provide for a very transparent compression, akin to soft tracking compression I'd give vocals in the studio, something to just take a little of the top to help level the take a bit, leaving any character compression for the mix.  This means that if I'm trying to keep ISO levels at -16dBFs for example, the compression won't even be in.  My mix bus will of course be a bit hotter at -10dBFs.  If the interviewee suddenly gets loud, they may start tickling the compression or if extreme outburst, will be "full on" but then at 2:1, I find it really hard to detect still, and then in very extreme situations, when the limiter does kick in, even it's effect, while noticeable, isn't as severe as suddenly hitting it without compression softening it up a bit first.

If I didn't have such trust in the compression of the TRX, I don't think then that it would be responsible for me to shoot for -10dBFs on the mix bus.

I attached a little test clip, which shows a typical unexpected outburst.  I edited so that you don't hear the first syllables, just in case there's an NDA concern, but the first hit (xx98_6) is around -3dBFs with a little spike, meaning that my TRX was in full compression and would have clipped the output otherwise.  The second (xx98_7) is my mix fully clipped.  The third (xx99_6) is me regaining down to avoid compression, fourth (xx99_7) is my mix (still uncompressed), and final (xx99_7_NORM) is that mix normalized to approach the level of the clipped take.

SZ1798_6.wav

SZ1798_7.wav

SZ1799_6.wav

SZ1799_7 NORM.wav

SZ1799_7.wav

Link to comment
Share on other sites

2 hours ago, Tom Visser said:

The comment that jumps out at me is "I've done studio recording for years".  When in the studio, I'll typically track vocal at around -10dBFs into ProTools - with the SSL board can ride the fader for an extra 10dB of "gain in hand" if I need a little boost or cut here or there.  Studio is a controlled environment, artist is glued to a spot in front off the pop filter or on a stool, mic on stand in the perfect position since there's no camera to deal with, and musical performance is typically fairly predictable.  Since your now talking about dialogue recording with lavaliers, totally different ballgame.  Now there's greater chances of being off axis or really "off mic" with head turns, the fact your dealing with spoken voice vs louder singing, when you gain up for that softer level, it does expose you to more risks from dynamic outbursts.  The way that I deal with field dialogue recording is to try to shoot for -10dBFs on my mix bus, but record ISO's somewhere in the -15 to -25dBFs range.  What this means in reality is that I will occasionally get peaks approaching 0dBFs on the mix bus as obtaining -10dBFs "RMS" is quite optimistic given the conditions.

Some notes...

1) when I say -10dBFs "RMS" that would be the person delivering a natural level in normal conditions.  If they are soft because of sensitivity of the subject matter, or moment of reflection, or are intentionally being soft, most likely it will be mixed that way too, so you don't have to force the -10dBFs technical goal in spite of the performance.

2) I personally rely heavily on Zaxcom TRX compression and like you don't use Zaxcom recorders.  I have a limiter set just below digital clip, but also enable compression.  I know a lot of people don't like to use compression, but for me personally, I've come to know and trust my compression settings.  From memory, my settings are -9dBFs threshold, very fast attack, slow release, 3dB knee, 2:1 ratio, no makeup.  In other words the mild ratio combined with soft knee provide for a very transparent compression, akin to soft tracking compression I'd give vocals in the studio, something to just take a little of the top to help level the take a bit, leaving any character compression for the mix.  This means that if I'm trying to keep ISO levels at -16dBFs for example, the compression won't even be in.  My mix bus will of course be a bit hotter at -10dBFs.  If the interviewee suddenly gets loud, they may start tickling the compression or if extreme outburst, will be "full on" but then at 2:1, I find it really hard to detect still, and then in very extreme situations, when the limiter does kick in, even it's effect, while noticeable, isn't as severe as suddenly hitting it without compression softening it up a bit first.

If I didn't have such trust in the compression of the TRX, I don't think then that it would be responsible for me to shoot for -10dBFs on the mix bus.

I attached a little test clip, which shows a typical unexpected outburst.  I edited so that you don't hear the first syllables, just in case there's an NDA concern, but the first hit (xx98_6) is around -3dBFs with a little spike, meaning that my TRX was in full compression and would have clipped the output otherwise.  The second (xx98_7) is my mix fully clipped.  The third (xx99_6) is me regaining down to avoid compression, fourth (xx99_7) is my mix (still uncompressed), and final (xx99_7_NORM) is that mix normalized to approach the level of the clipped take.

SZ1798_6.wav

SZ1798_7.wav

SZ1799_6.wav

SZ1799_7 NORM.wav

SZ1799_7.wav

Thank you very much Tom for taking the time to explain this. It really helps.
Since, just from habit, I am more comfortable compressing in post, I will probably try setting the threshold closer to -2dBFs with say 4:1 ratio and treat it more as a limiter. But I will try your settings also; maybe I will become more comfortable with it.

Two questions please:
(1) You say you use a limiter also. I assume this is a limiter in your recorder, as I do with the 702. Is this correct? 
(2) Can you tell me how NeverClip fits into what we are talking about? I have seen several threads on this, but I honestly don't understand how it works and what it does. Is it intended to avoid clipping during the wireless transmission? Any help would be appreciated. 
Thanks again. 

Link to comment
Share on other sites

I would not adjust any of the dynamics settings they are at the optimum setting by the factory for the unit to never clip . I would let the peaks on the transmitter meter come up to -10dB from full scale. Also be sure to use a mic without the -10dB red band. This kills the signal to noise by 10dB.

Link to comment
Share on other sites

2 hours ago, Mike H said:

 I have seen several threads on this, but I honestly don't understand how it works and what it does.

Essentially never clip is comprised of a dual A to D converter where one converter is running lower than the other. Then just before the signal clips a special look ahead algorithm will force the signal to be routed to the lower converter. Then the never clip software mixes both signals together so there is a seamless transition.

Link to comment
Share on other sites

2 hours ago, Mike H said:

Two questions please:
(1) You say you use a limiter also. I assume this is a limiter in your recorder, as I do with the 702. Is this correct? 
(2) Can you tell me how NeverClip fits into what we are talking about? I have seen several threads on this, but I honestly don't understand how it works and what it does. Is it intended to avoid clipping during the wireless transmission? Any help would be appreciated. 
Thanks again. 

I predominantly use an Aaton Cantar X2.  My Zaxcom comes in via AES, so any limiting from the wireless side is done on the body pack.  I believe in addition to the compression I listed before, I have a limiter at -1.  When I say that I use compression, I'll qualify that by saying I try not to record in such a manner where compression is being used.  In the event that it does get tickled, it's done in a way to be as transparent as possible.  If heavy compression is active, I don't necessarily consider that good, but on the other hand, it is better than the alternative of being hard limited or clipped.  I'm not sure how the Cantar handles limiting on the digital mix bus, but I'm sure something is done rather than just truncating the MSB side of the digital word.  The inclusion of the clips that I attached before was basically to show that even in almost worst case over modulation, the use of Zaxcom digital delivers something that could be used if it needed to and doesn't sound crunchy like it would have with a simple analog input to an analog mixer.

Most all my Zaxcom gear is older, so don't really rely on NC.  This is not a choice by me, just the reality given the vintage of my wireless and the fact that I don't utilize Zaxcom recorders and have the full power of Zaxnet at my disposal (just rudimentary and rather slow remote control via my IFB200).  I'm hoping to see something at NAB that might make me think of a strong investment in new wireless.  (QRX / RX12 replacement with integrated Zaxnet)

Link to comment
Share on other sites

29 minutes ago, Jack Norflus said:

Essentially never clip is comprised of a dual A to D converter where one converter is running lower than the other. Then just before the signal clips a special look ahead algorithm will force the signal to be routed to the lower converter. Then the never clip software mixes both signals together so there is a seamless transition.

Jack,
Thank you. A very simple explanation that I can understand.

On the last job I got a waveform that was flat at -6dBFs, with a few spikes above this. The audio did not sound like clipping, but it was muffled and not useable. I found the compressor threshold set at -6dBFs with attack on fast. I don't know what happened; I guess I somehow set the transmitter gain above -6dBFs and missed it on the check. I don't understand the peaks above -6dBFs that "blew through" the compressor. So this made my aversion to compression during tracking even worse.

Based on what you said, am I correct that I could:
(1) Switch the compressor off by setting the threshold at 0
(2) Monitor the 702 level during sound check and set the transmitter gain so that the peaks are around -10dBFs (or lower if the dynamic range is high)
(3) Switch the 702 limiter off and simply rely on the NeverClip to prevent clipping if I have an occasional spike.

This would keep me out of possible trouble with the compressor. You are confident enough in NeverClip to rely on it alone?

 

 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...