24-bit audio with mirrorless cameras

S

Samuel Dilworth

Well-known member
I’ve been struggling with various audio problems (as usual) and began to wonder if any mirrorless cameras have built-in 24-bit audio recording.

I know the GH5S does 24-bit recording when paired with the DMW-XLR1 adaptor. Anyone know if there are limitations to that?

For example, the XLR1 has two XLR inputs. Fine if you’re using an XLR mic. What if I’m using a wireless lav mic (e.g. the Sony URX-P40 receiver from the UWP-D21 package) with unbalanced output. Anyone know if I could use an adaptor cable to connect to the XLR1 and still get 24-bit depth?

Do the overpriced Sony equivalents to the XLR1 support 24-bit audio? I know at least one of those had digital output to supported Sony cameras with the Multi-Interface shoe (the ones with a bunch of pins at the front).
 
The Sony A7Siii, A1 and A7iv all have up to 24-bit internal recording with up to 4 channels as standard, depending on the mic setup used. They all now seem to be equipped with fairly decent pre-amps these days. All models will also work with the digital/analogue XLR-K3M audio interface. As always a top-quality mic into an analogue input will generally beat an average mic on a digital processing input. A point some people seem to miss. I've used the XLR module in both analogue and digital mode and my old ears really can't pick the difference unless the mic has a pretty high noise floor which none of my Sennheiser, Sony or Ramsa mics has. Mics obviously are analogue by nature as sound is analogue by nature. The XLR-K3M in reality has an internal A/D that converts the incoming audio into a digital stream to put out at its connection point, the hot shoe whereas if you use the analogue setting the analogue signal gets converted to a digital stream later after it enters the camera. Not a massive difference if the truth be told. This YT video gives a reasonably good overview of the ins and outs of the K3M module along with comparisons of its analogue vs digital performance characteristics. It's interesting in his case that he had what sounds like a ground loop hum when he had the third input, the 3.5mm input connected to a PC for music when the unit was in digital mode and the camera was being powered externally via its USB-C port. This ground loop sound was not present when the K3M was set to analogue mode. See this at 9:43 into the clip.

Chris Young

https://www.youtube.com/watch?v=hDTlgR28W1A
 
cyvideo said:
The Sony A7Siii, A1 and A7iv all have up to 24-bit internal recording with up to 4 channels as standard, depending on the mic setup used.
Click to expand...

Huh. How was I not aware of that? Sony should be shouting this from the rooftops. Especially with the weird meters and poor limiters in these cameras, 24-bit gives more room for error without a noise penalty, permitting levels to be set lower than with 16-bit.

I watched the video which went some way to explaining the higher price of the XLR-K3M versus the Panasonic DMW-XLR1: the Sony comes with a decent short shotgun mic (ECM-XM1) in the package.
 
Short cable runs of unbalanced connections are usually not a problem. For an unbalance to balanced input, it usually requires a custom XLR cable... XLR pin terminals 1 (shield) and 3 (cold) are tied together.
 
Rick R said:
Short cable runs of unbalanced connections are usually not a problem. For an unbalance to balanced input, it usually requires a custom XLR cable... XLR pin terminals 1 (shield) and 3 (cold) are tied together.
Click to expand...

Also, consider something a lot of people overlook which is correct microphone output impedance matching. The above method of 3 pin XLR to 3.5mm jack works. We've all done it. To be 100% correct though you should run balanced to unbalanced through a proper XLR Balanced Mic adapter when using balanced mics on mini-jack input cameras. There are numerous beneficial reasons for doing this as outlined below.

Chris Young

https://micbooster.com/leads-and-adaptors/70-test.html

EDIT:
The other thing you will notice on the Sony units as opposed to the Panasonic units is on the Sony they do have Auto AGCs if you wish to use them. Very useful if you are in a very noisy location with wildly fluctuating noise levels. The Panasonic adapter doesn't offer ACGs so you have to be a bit more careful in setting your levels to make sure they don't drive too hard into the limiters.
 
Last edited:
Samuel Dilworth said:
Huh. How was I not aware of that? Sony should be shouting this from the rooftops. Especially with the weird meters and poor limiters in these cameras, 24-bit gives more room for error without a noise penalty, permitting levels to be set lower than with 16-bit.

I watched the video which went some way to explaining the higher price of the XLR-K3M versus the Panasonic DMW-XLR1: the Sony comes with a decent short shotgun mic (ECM-XM1) in the package.
Click to expand...
I wouldn't get too crazy over 24bit. No reason not to record in it but don't expect too much of a difference. There are many more important factors that will ultimately determine the quality of your audio. It's a bit like saying hey my smart phone can film 4k 10bit yeah but it's still phone with a tiny sensor and highly compressed codec...
 
Last edited:
Peter C. said:
I wouldn't get too crazy over 24bit. No reason not to record in it but don't expect too much of a difference.
Click to expand...

Well, 24-bit gives huge room for setting levels safely low and raising in post. That can translate indirectly to better quality results.

The difference between 16-bit and 24-bit is massive in this regard. If more people realised how low they could set levels with 24-bit recording without incurring a noise penalty, they’d probably be less interested in 32-bit float. (Not saying 32-but float isn’t nice to have …)

With 16-bit recording you have no such freedom.
 
Peter C. said:
Ok I'll bite. Please explain these problems and tell me how 24 bit will solve them.
Click to expand...

Slow reply because I was travelling …

I’m recording audio from a Sound Devices MixPre-3 into a Sony α7C.

When I send a test tone at −20 dBFS (the lowest available setting for the tone on the MixPre) to the camera, there isn’t an “Audio Rec Level” on the camera low enough.

The camera has arbitrarily labelled levels from 0 to 31.

0 = silence.
1 = too hot for the test tone.

Moreover, there’s some kind of bug or shortcut taken by Sony in the α7C that means you don’t want to use an Audio Rec Level of 1 anyway. You need to be up around 3–5 for best results:

https://www.youtube.com/watch?v=Bh_gfEsFGtA

So let’s say I set the α7C levels to 5. How do I set the MixPre output? It seems to be guesswork, and since the α7C is limited to 16-bit recording, it’s guesswork with severe consequences if your guess is too low. And since this camera also has a limiter that cannot be turned off, you also have to worry about hitting that limiter when the levels look okay on the simplistic meters.

In short, how are people setting levels on these cameras?
 
Last edited:
Samuel Dilworth said:
In short, how are people setting levels on these cameras?
Click to expand...

Yes, a few of us have been chasing this problem over the last few years when it comes to feeding the A7 series cam. And yes the video you linked to is correct in its information about the artificial low clip at 9dB if you use anything lower the setting 3 in the Sony's. The solution that worked for me was to follow the advice I found in this "A Guide to Audio and the Sony A7s" outline, link below. Whilst it's a few years old and specifically refers to the A7s I've found it to work fine on all A7 models since. I've used these settings on an A7iii, A7siii and an A7iv and found they all respond the same way and to date, this is my preferred setup when feeding A7 series cams from a mixer.

Note: You may wonder why he refers to -18 as opposed to -20 as the tone set level. The reason for that is this is a UK article and -18 dB is the standard used in the UK and Europe are the EBU standards as follows:

1khz Line up tone would be 0db=-18dbfs=4ppm=-4vu=test on the EBU scale
Max broadcast level is +8db=-10dbfs=6ppm=+4vu=+8 on the EBU scale

This was always a pain for us in Oz when supplying programs to the UK as like the USA -20 is the Aussie standard. One TV series we did for Granada in the UK had to have the audio re-laid as the audio on the first two eps had all been mastered to -20... what we were used to! I should have known better being ex Beeb.

Chris Young

http://danmears.tv/a7saudio/
 
Last edited:
Samuel Dilworth said:
Slow reply because I was travelling …

I’m recording audio from a Sound Devices MixPre-3 into a Sony α7C.

When I send a test tone at −20 dBFS (the lowest available setting for the tone on the MixPre) to the camera, there isn’t an “Audio Rec Level” on the camera low enough.

The camera has arbitrarily labelled levels from 0 to 31.

0 = silence.
1 = too hot for the test tone.

Moreover, there’s some kind of bug or shortcut taken by Sony in the α7C that means you don’t want to use an Audio Rec Level of 1 anyway. You need to be up around 3–5 for best results:

https://www.youtube.com/watch?v=Bh_gfEsFGtA

So let’s say I set the α7C levels to 5. How do I set the MixPre output? It seems to be guesswork, and since the α7C is limited to 16-bit recording, it’s guesswork with severe consequences if your guess is too low. And since this camera also has a limiter that cannot be turned off, you also have to worry about hitting that limiter when the levels look okay on the simplistic meters.

In short, how are people setting levels on these cameras?
Click to expand...
I've been busy myself. I don't do a lot of work with mirrorless cameras but that was what I was getting at in general despite their specs that have historically not handled audio well. I would only use it to record the scratch track and use the mixpre recording as the primary audio. I understand from a production stand point it's nice when the audio is embedded into the video it eliminates a lot of work dealing with separate audio files.
 
You must log in or register to reply here.
Back
Top