GH4 GH4 with Factory Audio Fix Has 70 dB of Dynamic range !

[TheDingo]

Finally after waiting almost 2 months I was able to buy a GH4 camera with the factory audio fix, and from initial tests it sounds better than I ever expected.

Doing a quick noise-floor test with my Sound Devices MixPre-D mixer, I'm getting a noise floor reading of -70 dB RMS, which is 7 dB better than my GH3 or the previous un-fixed GH4 camera. This is better than the Zoom H4n portable recorder, and it means that there is at least 10 dB of audio "safety" range when recording.

i.e. you could set your GH4 volume levels 10 dB too LOW, then boost everything by 10 dB in post, and still have a great sounding low noise recording

...I haven't figured out what the best input signal level is for the GH4, but I expect I'll have this figured out within the next hour.

Woo hoo. Very happy with these results.

 

[TheDingo]

After a few more tests it appears that the best audio signal for the GH4 is a -50 dB MIC level signal with the camera audio level set to -12 dB.

This is a huge improvement over the GH3, which requires a -20 dB audio signal when recording with the camera at it's lowest audio setting.

So compact mics like the Rode Video Mic Pro can be left at it's default 0 dB gain setting when used with the GH4.

In case anyone is interested, the left audio channel in my GH4 is recording 0.1 dB louder than the right audio channel, and the right audio channel is showing a 1 dB lower noise-floor than the left channel.

 

[Bern Caughey]

Thanks TheDingo!

 

[Erik Naso]

TheDingo you are the man! Thanks for the report. Really thrilled it's better than previous models and even the new one! I just picked up a Juicedlink box for phantom power so I can use my shotgun and other power consuming mics.

I bought two more refurbished GH4's that have V 1.1 firmware installed. I hope these two have the fix already. How could I tell?

 

[TheDingo]

I bought two more refurbished GH4's that have V 1.1 firmware installed. I hope these two have the fix already. How could I tell?

If the serial number is higher than a certain number, then you've got a factory fixed GH4.

If you send a private message to the very helpful DVX member Emanuel, he can provide you with the details. ( Emanuel asked that this info not be made public, as I don't think he wants to upset Panasonic )

 

[danny dale]

If the serial number is higher than a certain number, then you've got a factory fixed GH4.

If you send a private message to the very helpful DVX member Emanuel, he can provide you with the details. ( Emanuel asked that this info not be made public, as I don't think he wants to upset Panasonic )

@Emanuel - I tried to send you a message to get this info but your box is full and rejects my attempts...

 

[Sol M.]

That's great to hear TheDingo, but also rather bittersweet for those who received the manual "fix" which seems to have made the camera noisier.

Can you share your testing and measurement protocol so that we can test our cameras for comparison with your results?

 

[GaryinCalifornia]

I bought two more refurbished GH4's that have V 1.1 firmware installed. I hope these two have the fix already. How could I tell?

where?

Thanks in advance...

 

[TheDingo]

That's great to hear TheDingo, but also rather bittersweet for those who received the manual "fix" which seems to have made the camera noisier.

Hmmm... That doesn't sound right. If anything I would expect the manual solution to not increase the noise-floor of the camera. You should try doing the same test I do, and see what results you are getting. ( instructions below )

Can you share your testing and measurement protocol so that we can test our cameras for comparison with your results?

I've been doing my own camera/recorder audio tests the same way for the past 6 years, which started because I found I could not believe the manufacturer specs any more, so I came up with my own simple noise-floor test.

To perform this test you need a high quality audio mixer that has a 1 kHz test-tone generator. I use my Sound Devices MixPre-D for this test. If you don't own a good mixer, then try and borrow one, and if you can't borrow one then rent one if you are able to. ( you might be able to do this test at the rental place, as it only takes a few minutes to perform, and there's no point in renting a mixer for 5 minutes of use )

I chose the Sound Devices MixPre because it has a real dynamic range of 110 dB, which means that the noise I will be measuring will be caused by my camera and not the mixer itself.

To adjust the signal strength of the 1 kHz test-tone from the Sound Devices MixPre, I use an old BeachTek DXA-2s passive audio adapter. ( I originally bought this for my brand new Canon HV20 DV tape camera many many moons ago )

The reason for adding the BeachTek adapter is that LINE level signal from my SD mixer is very likely too strong, while MIC level signal from the SD mixer might be too weak, so I need to be able to adjust the strength of the audio signal I feed my GH4 to obtain optimal signal strength. The BeachTek is a passive device, so it's unlikely that it will introduce it's own noise to the test. ( I have other gear I could use, but the BeachTek is very convenient )

HOW I TEST AUDIO NOISE-FLOOR

1- Connect the SD MixPre mixer to my BeachTek DXA-2s adapter using a high quality XLR cable, connect the BeachTek to my GH4 using the 3.5mm cable from the BeachTek.

2- Set the SD mixer to output a LINE level signal, set the BeachTek to output a mono signal to both output channels.

3- Turn on my GH4, set it to record a format that supports uncompressed LPCM audio, and set the camera audio level to -12 dB.

4- Turn on the SD mixer, then turn on the SD 1 kHz test-tone generator, and you should see the audio signal on the GH4 audio levels display.

5- Use the BeachTek to adjust the signal strength to show a -12 dB level on the GH4 audio level display. ( I can't remember, I think it's 2 or 3 units BEFORE the red units of the audio display )

6- Record 10 seconds of this 1 kHz test tone with your GH4, then turn off the camera and pop-out the SD card so that you can check how loud the actual recording is with either your video editor or audio editor. ( usually you will be a little high or low of the -12 dB target you want to hit )

7- Adjust the signal strength ( I'm using my BeachTek adapter for this adjustment ) until you are finally showing an actual -12 dB audio recording level by your GH4 camera. ( or something close to -12 dB )

8- Now that you are calibrated, make another 10 second recording of the 1 kHz test tone, but after 10 seconds SWITCH OFF the test tone on your mixer, and KEEP RECORDING for another 30 seconds. Essentially you will be recording silence from the mixer, and with my SD mixer it's probably -110 dB of audio silence.

9- Now turn off your GH4, and check this new recording with your video editor or audio editor. You should see the 1 kHz test tone right at -12 dB, just as you calibrated it to be, but when the test tone shuts off you are now going to see the actual recorded noise-floor of the GH4 camera in dB RMS. ( I use Sony Vegas Pro and Sony Sound Forge for video and audio editing, so I had to adjust the editor settings to allow for an audio levels display that goes down to -90 dB, as I think the default only shows -60 dB )

When I perform the test as described above, my 1 kHz test tone sits right at -12 dB and my audio levels drop down to -70 dB when I shut off the test tone. ( the levels bounce a little below -70 dB, but I know that this noise-floor is made up of random white-noise )

So with any luck you should see a noise-floor on your manually fixed GH4 that is close to -70 dB or at least -63 dB. If it's higher than -63 dB, then there's something not right happening here.

 

[Ed Jeffersen]

What do you mean by "manual fix"?
I sent mine in and it came back with v1.1 installed. Sounds really quiet. Quality is noticeably better than my GH3.

That's great to hear TheDingo, but also rather bittersweet for those who received the manual "fix" which seems to have made the camera noisier.

Can you share your testing and measurement protocol so that we can test our cameras for comparison with your results?

 

[TheDingo]

What do you mean by "manual fix"?

I think he means that the audio hardware fix was applied by a Panasonic repair center instead of being done in the factory when the camera was originally made.

I would expect that the manual fix would be identical to the factory made hardware fix.

I sent mine in and it came back with v1.1 installed. Sounds really quiet. Quality is noticeably better than my GH3.

I don't think Sol M. is experiencing the same results. I think he sent his camera in to have the hardware fix applied, but for some reason his camera came back in worse shape.

 

[Sol M.]

I may have a chance to test with a MixPre-D tomorrow, but I just did the test with a Tascam DR-60D. The noise floor may not be as low -110 dB, but the noise floor results (measured in Sony Vegas Pro 13) were practically the same as my previous post-fix audio tests using a powered-off Rode VideoMic Pro plugged into the GH4's mic input (to isolate the GH4's own self-noise):

• Channel 1: -53.5 dB (previous test: -52 dB)
• Channel 2: -55.5 dB

For reference, the DR-60D's test tone was calibrated to register the following levels on the GH4:

• Channel 1: -12.2 dB
• Channel 2: -14.9 dB

I'm not an audio engineer, so please do poke holes in my results. If there's a way to get better performance out of this camera, I want to know.

Also, out of curiosity, what is the reason for recording the test tone rather than just silence to measure the noise floor?

 

[TheDingo]

• Channel 1: -53.5 dB (previous test: -52 dB)
• Channel 2: -55.5 dB

For reference, the DR-60D's test tone was calibrated to register the following levels on the GH4:

• Channel 1: -12.2 dB
• Channel 2: -14.9 dB

That does not look good at all. Please post your results with the MixPre when you have them.

Also, out of curiosity, what is the reason for recording the test tone rather than just silence to measure the noise floor?

The test tone is a reference signal that allows you to calibrate the strength of the signal that you feed your recording device. This allows me to confirm that my recording device is recording the reference signal properly, and that the amount of gain is set correctly for the signal being recorded. ( without the test tone I could set my camera audio level to +6 dB and I would not know if this was the correct amount of gain to use while recording )

 

[Sol M.]

The test tone is a reference signal that allows you to calibrate the strength of the signal that you feed your recording device. This allows me to confirm that my recording device is recording the reference signal properly, and that the amount of gain is set correctly for the signal being recorded. ( without the test tone I could set my camera audio level to +6 dB and I would not know if this was the correct amount of gain to use while recording )

Yes, calibrating devices is standard practice for recording. What I'm not clear on is how this comes into play when measuring the noise floor of a single device. Are you saying that changing the output level of the input device (e.g. a mixer), say where the test tone registered -3 dB or -20 dB on the GH4 (at -12 dB mic gain setting) would change the GH4's own noise floor at its -12 dB gain setting?

 

[TheDingo]

What I'm not clear on is how this comes into play when measuring the noise floor of a single device.

If I apply too much gain with my recording device, then I am effectively raising the recorded noise-floor of that device.

Say a properly calibrated device shows a recorded noise-floor of -75 dB, but if I apply 20 dB of extra un-necessary gain while I am recording, the recorded noise-floor will now show as -55 dB and I've lost 20 dB of the device`s dynamic range. The 1 kHz signal enables us to determine if the signal being supplied to our recording device is the correct strength for the input/gain settings we are using on the device.

I also use the test tone to see if the actual recording levels match between my two stereo channels in the recorded audio file.

 

[Stuart U]

Many thanks TheDingo, you have been extremely helpful and generous with sharing your experiences.

 

[Sol M.]

If I apply too much gain with my recording device, then I am effectively raising the recorded noise-floor of that device.

Say a properly calibrated device shows a recorded noise-floor of -75 dB, but if I apply 20 dB of extra un-necessary gain while I am recording, the recorded noise-floor will now show as -55 dB and I've lost 20 dB of the device`s dynamic range. The 1 kHz signal enables us to determine if the signal being supplied to our recording device is the correct strength for the input/gain settings we are using on the device.

I also use the test tone to see if the actual recording levels match between my two stereo channels in the recorded audio file.

Yes, agreed. That's why I noted the -12 dB gain setting on the GH4 (the "recording device" in this situation) in my previous question. The minimal gain is being added by the GH4, so the noise floor should not be raised beyond the GH4's minimum level.

To be as succinct as possible: does using a test tone have any effect on GH4's noise floor (at its -12 dB mic gain setting), and why is it better than a test that isolates the GH4's own noise by using a powered-off input device (e.g. self-powered mic, etc.)?

 

[TheDingo]

The minimal gain is being added by the GH4, so the noise floor should not be raised beyond the GH4's minimum level.

Yes, essentially we want to eliminate the GH4 preamps effect on the recorded audio.

does using a test tone have any effect on GH4's noise floor (at its -12 dB mic gain setting)

It`s simply a calibration device, so we know that the signal strength being delivered to the camera is correct.

why is it better than a test that isolates the GH4's own noise by using a powered-off input device (e.g. self-powered mic, etc.)?

I don`t know if it`s better, but I prefer to have the test-tone at the beginning of the recording to check that the recording level is set correctly.

 

[Sol M.]

Thanks for the clarification! Final question: how are you measuring the noise floor via Vegas' level meter? When you say "-70 dB", is that the peak level of the noise during silence or are you arriving at that number through some other means?

 

[TheDingo]

TFinal question: how are you measuring the noise floor via Vegas' level meter? When you say "-70 dB", is that the peak level of the noise during silence or are you arriving at that number through some other means?

With Sony Vegas or Sony Sound Forge the audio level meters are calibrated using a dB RMS scale. When you play your audio track you will see the green level bars bounce up and down as your audio plays back, and the very top of the audio levels display there is a numerical peak dB value that shows the highest dB value for the portion of the audio that you play. When trying to assess noise-floor values, the audio level display will rapidly bounce up and down over a 2-4 dB range, and I go with what appears to be the average of the highest values the green level indicator displays. ( I find the peak numerical dB value listed at the top of the levels display to be a bit random when evaluating noise-floor audio )