Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Red audio testing with SD 442 - interesting results

StephenHall

New member
Joined
Jul 10, 2009
Messages
1
Reaction score
0
Points
0
Location
Greater Seattle area, WA
Website
sound.facesproject.org
Hi, I'm new here at reduser.net, but I've been on a few shoots with the Red and some have been challenges.

I've had some puzzling results. I wanted to share the results of some testing I did to try and get the best results with single-system sound on the Red.

When sending line-level audio to the Red, the Red seems way too sensitive.
I used the same setup I use to record to Panasonic HVX-200s, HPX-170s, HPX-500s, DVX-100s, Sony PD-150s and 170s and JVC HD100s, but when setting up a "normal" gain structure, the Red's recorded audio is clipped.

So, with the gracious assistance of Sunshine Whitton, owner of Absolute Technology here in the Seattle area, and the use of his camera for an hour, I tried to find out what was going on and develop a workaround.

First off, we tested on a Red One with Build 20 software, an updated audio board and the newer XLR-to-mini-XLR cables that he got with the audio board upgrade.

For the testing, we only used audio channels 1 and 2.
Each channel was set to line level input and the input level sub-selection was set to 0 db.

I used my Sound Devices 442 mixer with a Sennheiser MKH-416 on channel 1 and a Sennheiser G2 wireless mic on channel 2.
Channel 1 was panned full left and channel 2 was panned full right.
The connection to the Red from the 442 was a Remote Audio breakaway cable with headphone return.

The SD 442's output limiter was set to kick in at +17 db on the meter, 3 db below the factory default of +20 db.

The 442's master output level pot was set to unity gain ( at the detent ).

The SD 442's output level on the 'A' output bus was confirmed to be line level.

Again, this is no different than my usual setup.

Sending tone to the Red from the 442 resulted in green bars on the audio metering display that stopped at the first vertical white line ( reading from left to right ).

This, I've read, should nominally be -20db on the Red.

I also recorded to my Edirol R4-Pro hard disk recorder via the SD 442's direct outputs. Once again, tone from the SD 442 was set to the -20b line on the recorder's LCD meters.

I then had Sunshine do some normal-level talking and some shouting.

Even normal conversation was clipped on the Red, while the Edirol handled it perfectly. Shouting resulted in the Red clipping badly and the Edirol still not clipping.
As Ty Ford would say: "It had hair all over it..."


First Workaround
-----------------

I had two Shure 15/20/30 db switchable in-line attenuators.
I set them at 15 db and put them between the output of the snake and the Red's mini-XLR adapter cables.

We recorded a second time with normal conversation and with shouting.

This time, the Red didn't clip.
It looked like levels were a bit lower than I would have liked to see, but at least the audio was preserved.
Just guessing here, but I bet a 10 db attenuator would have been perfect.

When I pulled the Red clips into Vegas 9, it seemed like the shouted passages were peaking around -7.8 to -9 when the 15 db attenuators were in use.


Second Workaround
-------------------

We pulled the attenuators out of the signal path and I tried turning down the master output level on the SD 442.
This could be made to work.
When I set it to about 9 o'clock, the Red levels seemed to work out ok.

The problem with this approach is that it's not really a calibrated, repeatable mechanism.
I tend not to like "set it about there and try it again" stuff.
It might work in a pinch.

It seems like it messed up my gain structure and I wouldn't feel comfortable doing that.

Conclusion:
-----------

The Red One's audio input isn't like the other cameras that are in common use.
1 khz tone may line up ok, but normal line-level audio is too "hot" for the Red's input section.

A good way to settle things down may just be to use 10 db in-line attenuators.

Perhaps the Red folks will add a line-level pot or something, someday.

Cheers,

Stephen H

P.S. I also posted this on another forum at dvinfo.net
 
Thanks Stephen for doing all this helpful work.

I think that we are all hoping that audio will be much easier to use on the new cameras.
 
Audio Levels

Audio Levels

I take a different set of conclusions from your test and would ask you to try that again.

The fact that you can send a -20dBFS 1KHz tone to the camera and it indicates that it lines up to the single reference bar tells me that you do have correct input gain set on the camera - if it were set to +10dB rather than 0dB it would be indicating to the right of that witness mark.

A full scale 0dBFS (equal to +20dBu) output from a Sound Devices mixer would clip, but selecting +17dBU as a maximum output level shouldn't. I have tried that same test at +18dBU On a Sound Devices 302 and not seen clipping.

So the re-test I'd suggest is to verify the output limiter of the Sound Devices unit really was / is providing that functionality?


If it is, then we should talk about what else may be happening....
 
No follow up, guys?
I'm not an audio guy but I'm on set with a very knowledgeable and experienced one, and he seems to feel the way that Stephen does.
Slightly befuddling...
Cheers,
H.
 
I recently had exactly the same experience. Line level out of my 442 was recorded at good levels on my recorder (-10db avg lvl, -3db peaks, no clipping), and pretty much everything going into the RED from the camera out snake of the 442 was mashed. The RED build was 18 I believe, and the audio input setting was at 0 (0 and +10 being the only options). So, I flipped the -10 attenutator switch on the 442, and the RED audio only mashed on the peaks. I wasn't going to modifiy my main settings for the sake of the camera. In offline, ultimately they had to conform my audio when they had locked picture because the camera audio was unusable. Agreeing with the previous posts, I think RED audio is probably +12 db too hot at 0 line level input.

Digital audio ins would be really nice (48k 24b AES), then you would never have to deal with this problem. You could send digital outs from your recorder and be done with it. RED should partner with Apogee for clean, ultra low jitter clock hardware so the sync is perfect.
 
First off, we tested on a Red One with Build 20 software, an updated audio board and the newer XLR-to-mini-XLR cables that he got with the audio board upgrade.

Sending tone to the Red from the 442 resulted in green bars on the audio metering display that stopped at the first vertical white line ( reading from left to right ).

This, I've read, should nominally be -20db on the Red.

You mention the "first vertical white line" Build 20 only has one vertical white line. Are you sure you were one Build 20? On build 17 there were 3 vertical white lines and after it only one, this has me confused. There have been several audio improvements in 20 if you were on 17 with the multiple vertical lines the test doesn't surprise me as much as it would on 20.
 
Audio Levels

Audio Levels

Another test I'd suggest is as follows.

The Sound Devices 302 (and I assume other models) allow you to modify the output tone level.

In addition to -20dBFs (0dBu) is a 0dBFS setting (+20dBu) but that can be adjusted to -2dBFs (+18dBu)

What do the meters say when you record and playback that tone level?
 
Same experience here. The RED camera really just should work like every other camera on the market. Stuart, please understand that 'usually' the sound department doesn't have the time or doesn't get the opportunity to play around with the camera and test this and that and what have you. Why can't RED just stick to the standards and all is fine?
With the 'normal' (standard!) setup and gain structure the RED over-modulates. I personally use RED camera's audio strictly as scrap audio. Which is fine and not really a problem - if the RED wouldn't sell itself having audio recording capabilities. Well, yes, but only kind of so far.
 
Back
Top