BMCC Audio Performance - 1.2 vs 1.3 vs 1.4 vs 1.5

The place for questions about shooting with Blackmagic Cameras.
  • Author
  • Message
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

BMCC Audio Performance - 1.2 vs 1.3 vs 1.4 vs 1.5

PostSun May 05, 2013 9:33 am

Hi Everyone,

Ever since the 1.3 Firmware has been released, there have been a lot of fantastic blogs and forum posts that have covered the new firmware release in great detail. The general consensus seems to be although some things have been fixed (i.e. DC-offset, odd gain structure, automatic switching from MIC to LINE, etc.) - there are some other issues introduced, such as poor frequency response.

I did a lot of tests when we first got our camera, and then basically did the same tests again after 1.3 was released. However, all of these tests were basically to determine the best gain structure for connecting external mixers into the camera.

I'm now going to do some more tests to try and determine what exactly has changed from 1.2.1 to 1.3 - especially in terms of frequency response.

My testing methods are pretty basic. I've downloaded a whole heap of test tones/sounds from the good folks at AudioCheck.net, put them together into one long file in Adobe Audition, then exported out a 48kHz 24bit 320k MP3.

I've then put this MP3 onto a Sound Devices 552 for playback into the BMC. I would have preferred to use a WAV - however, I couldn't export a Broadcast Wave from my computer that the 552 would play back properly. However a 320k MP3 is pretty good quality all things considered - so I don't think it will be an issue at all.

Basically, all I'm going to do is play the file from the 552, then record it in RAW mode on the camera and compare the results between different firmware in Adobe Audition.

For this first test I'm going to use the latest 1.3 firmware.

Here are the settings on the BMC and 552:

BMC
Microphone Input: 0%
Ch 1 and CH2 Input Levels: Line
Ch1 Input: 100%
Ch 2 uses Ch 1 Input: Off
Ch 2 Input: 100%
Headphone Volume: 93%

Sound Devices 552
Tone Frequency: 1000 Hz
Tone Level: 0dBu
L/R XLR Output Switch: Line
Everything else factory default.

When sending out tone from the 552 (which reads 0dB on the meters), I get the following results:

Blackmagic UltraScope
bmc_13_line_us.png
bmc_13_line_us.png (605.6 KiB) Viewed 26859 times


Blackmagic MediaExpress
bmc_13_line_me.png
bmc_13_line_me.png (609.04 KiB) Viewed 26859 times


Adobe Audition
bmc_13_line_audition.png
bmc_13_line_audition.png (187.98 KiB) Viewed 26859 times


You can only attach three attachments per post, so I'm going do a new post with the results.

Best Regards, Chris!
Last edited by Chris Hocking on Fri Nov 15, 2013 12:40 am, edited 3 times in total.
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSun May 05, 2013 10:52 am

I then down-graded to the 1.2.1 firmware.

To keep things consistent, I decided to use Line Level for these tests as well - even though my previous tests have shown that Mic Level gets you better results. The main reason for this is the original tests I did were with a Sound Devices 302 Mixer, which allows you to attenuate the signal - whereas the 552 only gives you Mic, -10 and Line.

Here are the settings on the BMC and 552:

BMC
Microphone Input: 0%
Ch 1 and CH2 Input Levels: Line
Ch1 Input: 65%
Ch 2 uses Ch 1 Input: Off
Ch 2 Input: 65%
Headphone Volume: 93%

Sound Devices 552
Tone Frequency: 1000 Hz
Tone Level: 0dBu
L/R XLR Output Switch: Line
Everything else factory default.

When sending out tone from the 552 (which reads 0dB on the meters), I get the following results:

Blackmagic UltraScope
bmc_121_line_us.png
bmc_121_line_us.png (608.06 KiB) Viewed 26847 times


Blackmagic MediaExpress
bmc_121_line_me.png
bmc_121_line_me.png (617.63 KiB) Viewed 26847 times


Adobe Audition
bmc_121_line_audition.png
bmc_121_line_audition.png (183.57 KiB) Viewed 26847 times


It's worth noting that the screenshot in Adobe Audition shows the files straight from the camera - there has been no fix for the DC offset applied.
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSun May 05, 2013 11:36 am

Here are the results:

Original Test File
Download audio file here.
test-track.png
test-track.png (153.76 KiB) Viewed 26845 times


1.2.1 Firmware (with DC-offset Repaired)
Download audio file here.
test_1-2-1_dc_repaired.png
test_1-2-1_dc_repaired.png (142.38 KiB) Viewed 26845 times


1.3 Firmware
Download audio file here.
test_1-3.png
test_1-3.png (143.13 KiB) Viewed 26845 times


Again - big thanks to the guys at AudioCheck.net for supplying the test audio.

As you can see, I'm basically getting exactly the same results in terms of frequency response as the juicedLink guys.

I'm going to do some more tests - but at the moment, I think I'm going to stick with the 1.2.1 firmware, and just repair the DC-offset in post. You have to balance the compromises, but I think it's going to be easier and achieve better results if I use Mic Level with 1.2.1 and repair the DC Off-set in post, then using 1.3 with Line Level and have to try and apply and EQ to repair the frequency response curve.

Hopefully BMD comes out with a fix soon!

I'm still a little shocked that they haven't even made any public statement about the audio changes in the 1.3 firmware. Very odd.
Offline
User avatar

Peter J. DeCrescenzo

  • Posts: 2205
  • Joined: Wed Aug 22, 2012 6:53 am
  • Location: Portland, Oregon USA

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSun May 05, 2013 5:43 pm

Hi Chris: Thanks very much for the information.

I hope BMD addresses audio issues in their camera's firmware ASAP.

Cheers.

-
http://www.peterdv.com
http://HereForTheWeather.wordpress.com
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostTue May 07, 2013 2:33 pm

After much messing around - as of today, my new preferred settings are as follows:

BMC 1.2.1
Microphone Input: 0%
Ch 1 and CH2 Input Levels: Line
Ch1 Input: 77%
Ch 2 uses Ch 1 Input: Off
Ch 2 Input: 77%
Headphone Volume: 93%

Sound Devices 302
XLR Atten. Level: 0dB
Everything else is factory default.

This means I still have to remove the DC Offset in Post - but that's literally one button click in Audition.

When sending Full Scale Tone (where "full scale" is the same as "clipping level") from the 302 to the Camera I get the following results:

Blackmagic Ultrascope 1.6.3
line-77-ultrascope.png
line-77-ultrascope.png (604.08 KiB) Viewed 26649 times


Blackmagic Media Express 3.2
line-77-mediaexpress.png
line-77-mediaexpress.png (656.58 KiB) Viewed 26649 times


Adobe Audition CS6 5.0.2 Build 5
line-77-audition.png
line-77-audition.png (183.97 KiB) Viewed 26649 times


When sending out a "factory default" tone (1kHz tone at 0dBu) from the 302, it reads about -21dB in Audition after you repair the DC Offset - which is about right (ideally 0dB on the 302 should read -20dB in your NLE).

Previously, I have said that I preferred using Mic Level with this firmware - but the automatic switch from Mic to Line when the signal is too hot, is simply too annoying. When I say too hot - it doesn't even have to peak. If you send a tone that's CLOSE to peaking in both channels, the camera freaks out, and automatically switches to line. You have to get your limiters JUST RIGHT to use this workflow - but the problem is, you can't send full scale tone to the camera when you're in Mic Level, because it will automatically switch over to line. Because of this - you don't really know where your levels are REALLY sitting, and your left trusting Ultrascope. But the problem with Ultrascope is that the DC Offset screws up the metering. So... on the 1.2.1 firmware, Mic Level is just too hard, because you can't force it to stay in Mic Level.

The only truly accurate way I've been able to determine levels is judge everything Audition AFTER the DC Off-set has been removed.

Setting the Line Level to 77% on the camera might seem like a bad idea at first - mainly because Robert from juicedLink originally said to use somewhere in the 25-30% range, and everyone seems to take this as gospel.

However, Robert from juiceLink is biased, because currently all of his juicedLink products (with the exception of the yet-to-be-released BMC366 Low-Noise Preamp - which looks awesome) use an unbalanced mic-level out. This makes sense, as most of his clients are using pro-sumer cameras such as DSLRs and Handicam's which only have 3.5inch stereo inputs. But, either way, all of his tests in his initial video were focussed towards Mic Level - not Line Level.

Given that I'm sending Line Level out of the 302 - and the BMC is correctly calibrated at 77%, I'm just going to have to assume that 77% Line Level is actually 0dB Gain / Attenuation. This sort of makes sense - maybe 80% is technically "line level" on the camera, and you have 20dB of gain and 70dB of padding to play with?

How I test for this... I have no idea. If anyone has any ideas, please let me know!

This is really truly guess work - but until BMD actually releases some useful information - there's not much we can do.

For now though... I'd prefer to avoid 1.3, if the only thing it really fixes is the DC-offset, and stick with 1.2.1, which gives us more "data" to play with in post.

Hopefully in the next release they'll replace the audio menu all together with something that allows you to select Mic vs Line PER CHANNEL, offer actual dB numbers instead of percentages, and document what is exactly going on behind the scenes (i.e. where the digital gain kicks in). Eventually... I'm sure they'll add software meters, which will be even more handy!

Sending a useable line level signal into the camera should't be this hard.

Christine Peterson - if you're reading this... what's the deal? Why is BMD being so quiet about this audio functionality on a product that's out there in the wild being used daily? I don't want to to discuss things in future firmware - I want you to please explain things in the CURRENT firmware! What do the percentages mean? Explain why you made the changes from 1.2.1 and 1.3? Do you actually consider 1.3 to be "fixed"?

The adventure continues...
Offline

ChrisBarcellos

  • Posts: 329
  • Joined: Tue Sep 04, 2012 3:34 am

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostTue May 07, 2013 5:37 pm

Chris:

I've come to the conclusion that Christine Petersen is like that "artificial actor" they had in an Al Pacino film a few years back. She must not be real and she is immune to sensible queries and discussions. In fact the whole Black Magic experience may be a plot by the Sony/Canon/Panasonic/RED players to show us how much we need them.

Just jesting, of course, but sometimes the lack of communication is maddening.....
Offline

Theodore Prentice

  • Posts: 591
  • Joined: Mon Sep 03, 2012 3:56 pm

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostTue May 07, 2013 6:24 pm

Who is Christine Petersen :?:

















.









(/end sarcasm)
Offline
User avatar

rick.lang

  • Posts: 7575
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostTue May 07, 2013 6:29 pm

Theodore Prentice wrote:Who is Christine Petersen?


Title for your next creative film?

Rick Lang
Sent from my iPad using Tapatalk HD
Rick Lang
Offline

Christine Peterson

  • Posts: 662
  • Joined: Mon Sep 24, 2012 9:00 pm
  • Location: Boston, MA

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostTue May 07, 2013 6:47 pm

ChrisBarcellos wrote:I've come to the conclusion that Christine Petersen is like that "artificial actor" they had in an Al Pacino film a few years back. She must not be real and she is immune to sensible queries and discussions.

Hahahahaha. I just honestly don't know much about this issue, but I'll make sure the product team is aware.
Christine Peterson

(Previously Community Relations Manager for Blackmagic Design)
Offline

ChrisBarcellos

  • Posts: 329
  • Joined: Tue Sep 04, 2012 3:34 am

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostTue May 07, 2013 8:45 pm

Just "funnin" with you Christine. You are our only point of contact at this point and it would be nice if they gave you more to work with so you could address the concerns of you consumers better.
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostTue May 07, 2013 9:57 pm

Christine Peterson wrote:Hahahahaha. I just honestly don't know much about this issue, but I'll make sure the product team is aware.


Thanks Christine - GREATLY appreciated!

Even if they just said:

A. We are aware of the issues with 1.3. Please use 1.2.1 for the time being until a future release.
B. 1.3 is the way we intend the camera to operate, and will not change things future firmware.
C. You're doing something wrong - please use these settings instead.
D. The audio changes slipped under the radar with the 1.3 release. We will update the documentation to reflect the changes we made in 1.3.

...and I'd be happy.

If BMD released the camera with the 1.3 firmware first this would be a very different discussion. But the fact that they had issues with 1.2.1, then released 1.3 WITHOUT TELLING ANYONE they changed the audio functionality, and let the users find out themselves that 1.3 has fixed some things but broken others is just unfriendly.

Both the Alexa and RED released their cameras initially without audio, and then when they eventually added audio, it had bugs... HOWEVER, at least they documented everything properly.

If the BMC is to be considered a SERIOUS PROFESSIONAL CINEMA CAMERA - then good documentation needs to be in place - end of story.

If the BMC is to be considered a SERIOUS CONSUMER CAMERA - then this is an even bigger issue, as Sony and Panasonic customers, etc. just expect their camera to work out of the box.

Please let me know if I'm out of line... but I think I'm being pretty reasonable.

I'm a MASSIVE fan of this camera. We took it for a test run before it was even publicly released. We're using the camera in the real world every day on jobs. But... this whole audio issue is a big stumbling block.

Yes... most people will use an external recorder, and this makes the most sense. HOWEVER, considering the camera costs $3K, and it's hire value is stupidly cheap (like AUD$126/day for a basic kit) it seems a little crazy sending out an expensive Sound Devices, Zaxcom, Cantar, etc. kit with the camera - and would make a lot more sense financially if we could just bundle something like a 302 with the camera. Personally, I consider the BMC a pro-sumer camera, just like the old HVX202 and Z1P. If I was using it for high-end commercials, television or long-form - then yes, I would definitely use external audio and just record timecode and a guide track to the camera. BUT... to be honest, I'm NOT going to be using this camera for as my main camera on these jobs... there are just too many other limitations. Don't get me wrong - it's a FANTASTIC camera that records beautiful images. But it's not Alexa or RED EPIC. I'll be using the BMC for LOW BUDGET jobs because the camera kit itself is cheap. I just need to find an audio solution now that matches the budget.

Anyway... My 2c.

Hopefully one of the product managers from BMD can jump in and offer some words of wisdom! I find it funny that they're willing to jump in when there are non-issues (such as dirt around the fans)... but are dead quiet when there's an actual real functional concern.
Offline

Soeren Mueller

  • Posts: 595
  • Joined: Tue Sep 04, 2012 2:21 pm
  • Location: Düsseldorf, Germany

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostWed May 08, 2013 7:53 pm

Chris Hocking wrote:If the BMC is to be considered a SERIOUS PROFESSIONAL CINEMA CAMERA - then good documentation needs to be in place - end of story.


Hey Chris, good work! And I don't think you're out of line at all. I'm also a very big fan of the camera and a happy owner since 2 months, however I'm totally with you here - especially with the projects I'm using the camera for the audio part is sort of important and it's kind of a big pain with these undocumented changes happening! :(
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostThu May 09, 2013 4:54 am

I've been in email contact with Robert from juicedLink, asking about his soon-to-be-released BMC366 Low-Noise Preamp product. The information he replied back with is so incredibly helpful and useful, I thought I'd share it here (with his permission):

Basically, each XLR input has its own MIC/LINE, GAIN, and Phantom swithes. Phantom power can be selected from either 12 or 48 volts. Each XLR input can be selectively power down, to extend battery life. MONO/STEREO, plus a switchable attenuator on the right channel for audio output bracketing (which is where you want to do it ... doing it in the camera will result in either pushing the gain artificially high in the camera on one track resulting in poor SNR, or reducing the gain (digital) in the camera on one of the tracks (where you have constricted headroom, defeating the purpose)). The signal level adjustment of the auxiliary output. All of the outputs are an amplified MIC level.

"Line Level will get you the best signal to noise ratio?" - This is not correct, assuming that using a low noise preamplifier for both the MIC and LINE inputs of a recorder. I would agree that this can be true if you're not using a low noise preamplifier (on the MIC input, and the camera's front end is setting the noise figure). From the principle of the the noise figure of cascaded amplifiers, a properly designed a low noise preamplifier will set the noise figure of the entire system. In this case, from a signal to noise perspective, there's no benefit for going in to a recorder LINE level. The recorder at LINE is just going to engage an input attenuator, and then subsequently go through the same signal chain as the MIC input.

The where there is a benefit for running line level, has nothing to do with signal to noise in the recording device. If there's any electromagnetic interference they can be picked up on a long cable run, it's relative affect will be smaller on a line level signal than a MIC level signal. But, for the purposes of a camera preamplifier like a juicedLink, the exposure that you have to electromagnetic susceptibility is quite small, since your cable run is only about 6 inches.

Regarding the camera results at LINE level, my little videos were more focused on an applications guide for users of a juicedLink, so I didn't really cover that. But if I recall correctly, the LINE sensitivity in 1.2 was off the scale (required 100% gain in the camera for a reference test tone from and ENG mixer like the MixPre). So, the SNR in MIC was actually better than in LINE (from what I tested). The sensitivity of 1.3 in LINE was within range, but I don't recall the cal gain level in the camera.

Hopefully, they will fix all of these firmware issues with the BMCC. Until then, I personally would probably use firmware version 1.3 with a juicedLink to avoid the camera the suddenly freaking out and switching to LINE level if it gets overloaded. The frequency response of 1.3 is much worse. But, the filtering in the camera is all done after the A/D, so it can easily be reconstructed without penalty in post production.


Very interesting stuff! Thanks heaps for all your help and information Robert - HUGELY appreciated!

Looks like I need to do some more testing! :)
Offline
User avatar

rick.lang

  • Posts: 7575
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostThu May 09, 2013 5:34 am

Chris Hocking wrote:I've been in email contact with Robert from juicedLink, asking about his soon-to-be-released BMC366 Low-Noise Preamp product. The information he replied back with is so incredibly helpful and useful, I thought I'd share it here (with his permission):

Basically, each XLR input has its own MIC/LINE, GAIN, and Phantom swithes. Phantom power can be selected from either 12 or 48 volts. Each XLR input can be selectively power down, to extend battery life. MONO/STEREO, plus a switchable attenuator on the right channel for audio output bracketing (which is where you want to do it ... doing it in the camera will result in either pushing the gain artificially high in the camera on one track resulting in poor SNR, or reducing the gain (digital) in the camera on one of the tracks (where you have constricted headroom, defeating the purpose)). The signal level adjustment of the auxiliary output. All of the outputs are an amplified MIC level.

"Line Level will get you the best signal to noise ratio?" - This is not correct, assuming that using a low noise preamplifier for both the MIC and LINE inputs of a recorder. I would agree that this can be true if you're not using a low noise preamplifier (on the MIC input, and the camera's front end is setting the noise figure). From the principle of the the noise figure of cascaded amplifiers, a properly designed a low noise preamplifier will set the noise figure of the entire system. In this case, from a signal to noise perspective, there's no benefit for going in to a recorder LINE level. The recorder at LINE is just going to engage an input attenuator, and then subsequently go through the same signal chain as the MIC input.

The where there is a benefit for running line level, has nothing to do with signal to noise in the recording device. If there's any electromagnetic interference they can be picked up on a long cable run, it's relative affect will be smaller on a line level signal than a MIC level signal. But, for the purposes of a camera preamplifier like a juicedLink, the exposure that you have to electromagnetic susceptibility is quite small, since your cable run is only about 6 inches.

Regarding the camera results at LINE level, my little videos were more focused on an applications guide for users of a juicedLink, so I didn't really cover that. But if I recall correctly, the LINE sensitivity in 1.2 was off the scale (required 100% gain in the camera for a reference test tone from and ENG mixer like the MixPre). So, the SNR in MIC was actually better than in LINE (from what I tested). The sensitivity of 1.3 in LINE was within range, but I don't recall the cal gain level in the camera.

Hopefully, they will fix all of these firmware issues with the BMCC. Until then, I personally would probably use firmware version 1.3 with a juicedLink to avoid the camera the suddenly freaking out and switching to LINE level if it gets overloaded. The frequency response of 1.3 is much worse. But, the filtering in the camera is all done after the A/D, so it can easily be reconstructed without penalty in post production.


Very interesting stuff! Thanks heaps for all your help and information Robert - HUGELY appreciated!

Looks like I need to do some more testing! :)


Hope his new product provides balanced TRS 1/4" inputs to the BMCC etc. Tonight, i am sorry I just can't remember if it does that. His comments sound like he is still providing unbalanced signals.

Rick Lang
Sent from my iPad using Tapatalk HD
Rick Lang
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostThu May 09, 2013 5:41 am

rick.lang wrote:Hope his new product provides balanced TRS 1/4" inputs to the BMCC etc. Tonight, i am sorry I just can't remember if it does that. His comments sound like he is still providing unbalanced signals.


Unlike the Riggy Micro/Assist range, the BMC366 Low-Noise Preamp has two BALANCED mini-jacks which are switchable between Mic and Line Level. You can buy custom right angled cables that go from the balanced mini-jack on the BMC366 to the balanced 1/4inch jacks on the side of the BMC.
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostFri May 10, 2013 4:33 am

Any response from the product gurus Christine Peterson?
Offline
User avatar

rick.lang

  • Posts: 7575
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostFri May 10, 2013 4:05 pm

Chris Hocking wrote:
rick.lang wrote:Hope his new product provides balanced TRS 1/4" inputs to the BMCC etc. Tonight, i am sorry I just can't remember if it does that. His comments sound like he is still providing unbalanced signals.


Unlike the Riggy Micro/Assist range, the BMC366 Low-Noise Preamp has two BALANCED mini-jacks which are switchable between Mic and Line Level. You can buy custom right angled cables that go from the balanced mini-jack on the BMC366 to the balanced 1/4inch jacks on the side of the BMC.


Thanks, Chris. I am sure you are eager to try the BMC366!

Rick Lang
Sent from my iPad using Tapatalk HD
Rick Lang
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSat May 11, 2013 2:37 pm

After reading from Robert at juicedLink that the filtering in 1.3 release of the firmware is all done after the analog-to-digital conversion (and can easily be reconstructed without penalty in post), and that his recommendation was to use the 1.3 release (as opposed to 1.2.1) - I decided to go back to the drawing board and do some more tests.

For these latest tests, I ended up doing things slightly differently that the tests I did at the start of this thread - here's what I did this time round:

1. Download test material from AudioCheck.net

2. Create a new 48kHz 24bit WAV with the following content:

> 10sec Full Scale Tone (i.e. one notch down from clipping using FCP7 Generator)
> 10sec -12dB Tone (FCP7 Generator)
> 10sec Pink Noise
> 18sec Test Music
> 31sec High Frequency Check
> 30sec Low Frequency Check
> 20sec Logarithmic Sine Sweep - Full Spectrum (20 Hz - 20 kHz)

3. Put the test WAV on a Tascam DR-100.

4. Connect the -10dBV Line out from the DR-100 to the RTN inputs of a Sound Devices 302. I then activated the RTL L to Ch. 4 In and RTL R to Ch. 5 In options using the setup menu.

5. Calibrated the RTN levels on the 302 using the full scale tone on the test file as a reference.

6. Made sure that the BMC was running the 1.3 firmware. I then set the Ch 1 and CH2 Input Levels to Line and set the Ch1 and Ch2 levels to 100%.

7. I then connected the BMC to my laptop using Thunderbolt, so that I could use Blackmagic Media Express 3.2 to capture the test footage and also to monitor the levels on the camera.

8. I then bring in all the MOV test files into Adobe Audition for comparison.

When sending out "full scale" tone (i.e. clipping level) from the 302 I get the following results:

Blackmagic UltraScope 1.6.3
bmc_1-3_line_fulltone_us.png
bmc_1-3_line_fulltone_us.png (611.23 KiB) Viewed 26285 times


Blackmagic Media Express 3.2
bmc_1-3_line_fulltone_me.png
bmc_1-3_line_fulltone_me.png (627.45 KiB) Viewed 26285 times


Adobe Audition 5.0.2 (Captured from Media Express)
bmc_1-3_line_fulltone_aa.png
bmc_1-3_line_fulltone_aa.png (161.55 KiB) Viewed 26285 times
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSat May 11, 2013 2:39 pm

When sending out "factory default" tone (1kHz tone at 0dBu) from the 302 I get the following:

Blackmagic UltraScope 1.6.3
bmc_1-3_line_tone_us.png
bmc_1-3_line_tone_us.png (608.99 KiB) Viewed 26285 times


Blackmagic Media Express 3.2
bmc_1-3_line_tone_me.png
bmc_1-3_line_tone_me.png (617.83 KiB) Viewed 26285 times


Adobe Audition 5.0.2 (Captured from Media Express)
bmc_1-3_line_tone_aa.png
bmc_1-3_line_tone_aa.png (157.48 KiB) Viewed 26285 times
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSat May 11, 2013 2:45 pm

Now let's look at the results:

Test Clip
48kHz_24bit_test_track.png
48kHz_24bit_test_track.png (157.84 KiB) Viewed 26285 times


Result from Camera (without any processing applied)
bmc_1-3_line_test_track.png
bmc_1-3_line_test_track.png (158.98 KiB) Viewed 26285 times


Result from Camera (with EQ applied in Adobe Audition)
bmc_1-3_line_test_track_eq.png
bmc_1-3_line_test_track_eq.png (163.02 KiB) Viewed 26285 times
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSat May 11, 2013 2:54 pm

Here are the EQ settings I used in Adobe Audition:

bmc_FFT_filter.png
bmc_FFT_filter.png (61.28 KiB) Viewed 26886 times


It's not perfect - but it seems to do a pretty good job.

Based on these latest tests, until BMD come out with a new firmware release that hopefully fixes these issues, I'm going to stick with the 1.3 firmware for now, and apply this FFT filter to all of the audio.

Originally I was thinking that I'd just stick with 1.2.1 and remove the DC-offset in Audition - getting a decent Line Level signal into the camera using 1.2.1 proved to be tricky, and Mic Level is useless because of the automatic switching to line when the signal is "too hot".

Please keep in mind - that this latest tests ONLY takes into account Line Level inputs. If you're sending the camera a Mic Level feed - then I'd suggest referring to this video.
Offline

Margus Voll

  • Posts: 993
  • Joined: Wed Aug 22, 2012 10:31 am
  • Location: Tallinn, Estonia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostWed May 15, 2013 6:22 am

I just saw something like you try to compensate in audition on my last weeks shoot.

Audio devices files where super nice but the same mixer outputted to bmc ref sound and it
"looked" like audio with telephone effect on it. All the lower frequencies were lost so to say in comparison.


I could post some samples if anyone cares.
Margus Voll, C.S.I.

http://iconstudios.eu
margus@iconstudios.eu
@margusvoll

Resolve 14.1 W10
Resolve 12.5.4 OSX
BMC 2,5 K
Pocket camera
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostWed May 15, 2013 6:47 am

Yes, if you're using the 1.3 firmware the audio that is recorded has a lot of the lower frequencies reduced - and this needs to be fixed in post production. Luckily - it doesn't look like you're loosing any quality/information though. Mic Level is worse than Line Level.

If you're using the 1.2.1 firmware, you still need to do some audio post production - but it's not as bad.

HOPEFULLY eventually someone from BMD will actually comment on this. They've been deadly quiet - which is really disappointing, and extremely unhelpful.

After careful consideration, I feel safe saying that 1.3 is definitely an improvement on 1.2.1 - but it's still a long while off being useable straight out of the camera. This NEEDS to be fixed.

Anyone from BMD care to comment, or at the very least, release a frequency plot for the 1.3 firmware so that we can EQ the audio correctly?
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostThu May 16, 2013 1:14 pm

OK.. One last try before I give up...

Christine Peterson - any word from the engineers about what exactly is going on with the audio in the 1.3 firmware?
Offline

johnjvogel

  • Posts: 22
  • Joined: Wed Aug 29, 2012 7:15 pm

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostThu May 16, 2013 9:05 pm

Chris Hocking wrote:OK.. One last try before I give up...

Christine Peterson - any word from the engineers about what exactly is going on with the audio in the 1.3 firmware?


Chris, if I may add. There is one expectation if this camera is offering L/R inputs, Line/Mic levels - That this works like any other professionally built camera. I'm not even going to concern this with the lack of phantom power, because there are work arounds for that.

This means that if I am on a shoot, let's say an interview and all I want to do is set up a simple Seinheisser ME66, which is self powered. On my other cameras, Panasonic HCX 900 or Sony Z7U even Canon 5D with ML, I can plug in my ME66 and it works - no phantom power needed since the mic is powered by a AA battery. This is an industry standard. If I plug my ME66 into my BMCC 2.5k with 1.3 firmware - unusable audio. The levels are so low that even when importing into FCP, I have to jack up the levels so much that there is an unbelievable amount of background "room" noise and I think the fan of the camera is mixed in. Yes, I'm aware that there are work arounds for this. I'm just pointing out that there is a certain level of basic expectation. Now I'm not buying a $3,000 camera, I'm buying a $3,500 because I have to buy a juciedlink adapter? Why offer any audio at all?

I don't expect to get audio to my Canon 5D (I'm aware and use magic lantern), because it doesn't offer two mic inputs with Line/Mic levels and find it a gift that I can record audio to the card in some way. I do expect to be able to record audio in a simple fashion to the BMCC because they are dressing the camera to do so.
Offline

Paul Stone

  • Posts: 26
  • Joined: Thu Dec 20, 2012 4:58 am
  • Location: Queensland, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostThu May 16, 2013 11:53 pm

We did our first test shoot using audio on our BMCC a few days ago, and my boss was mortified at the resulting sound. He's used a bunch of cameras over the years, from dinosaurs like the Panasonic 502 to the more compact 102, and now to the Canon 5d, and we've never seen anything like it.

We pretty much can't shoot any jobs requiring audio on the BMCC until this is sorted.
Offline
User avatar

Rakesh Malik

  • Posts: 1708
  • Joined: Fri Dec 07, 2012 1:01 am
  • Location: Tacoma, WA

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostFri May 17, 2013 12:01 am

Paul Stone wrote:We did our first test shoot using audio on our BMCC a few days ago, and my boss was mortified at the resulting sound. He's used a bunch of cameras over the years, from dinosaurs like the Panasonic 502 to the more compact 102, and now to the Canon 5d, and we've never seen anything like it.

We pretty much can't shoot any jobs requiring audio on the BMCC until this is sorted.


What was the sound like? Was it noisy, or distorted?

I'm just curious because I shot a short video using a loaner BMCC a couple of months ago using the BMCC's audio and it sounded quite good. I used a MixPre-D and an Audio Techncia shotgun mic for the audio.
Rakesh Malik
Cinematographer, photographer, adventurer, martial artist
http://WinterLight.studio
Offline

Paul Stone

  • Posts: 26
  • Joined: Thu Dec 20, 2012 4:58 am
  • Location: Queensland, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostFri May 17, 2013 12:11 am

Tamerlin wrote:
Paul Stone wrote:We did our first test shoot using audio on our BMCC a few days ago, and my boss was mortified at the resulting sound. He's used a bunch of cameras over the years, from dinosaurs like the Panasonic 502 to the more compact 102, and now to the Canon 5d, and we've never seen anything like it.

We pretty much can't shoot any jobs requiring audio on the BMCC until this is sorted.


What was the sound like? Was it noisy, or distorted?

I'm just curious because I shot a short video using a loaner BMCC a couple of months ago using the BMCC's audio and it sounded quite good. I used a MixPre-D and an Audio Techncia shotgun mic for the audio.


Really, really low. Boosting it up in After Effects was the only way to even hear it, and the increase in volume revealed a bunch of noise. We used a nearly brand-new pair of Sennheiser (ew112p G3) radio mics.
Offline

bhook

  • Posts: 1024
  • Joined: Wed Aug 22, 2012 8:19 pm

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostFri May 17, 2013 12:20 am

There has to be a reason BMD hasn't gotten VUs into the camera yet. I was thinking that it might have been because of the DC offset problem but now I'm not so sure.
Offline
User avatar

Rakesh Malik

  • Posts: 1708
  • Joined: Fri Dec 07, 2012 1:01 am
  • Location: Tacoma, WA

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostFri May 17, 2013 2:17 am

Paul Stone wrote:
Really, really low. Boosting it up in After Effects was the only way to even hear it, and the increase in volume revealed a bunch of noise. We used a nearly brand-new pair of Sennheiser (ew112p G3) radio mics.


I'm sure that would have been a lot easier to prevent if the camera had VU meters on it, so that you'd have some way of knowing what levels the camera was seeing in the first place. Sigh.
Rakesh Malik
Cinematographer, photographer, adventurer, martial artist
http://WinterLight.studio
Offline

johnjvogel

  • Posts: 22
  • Joined: Wed Aug 29, 2012 7:15 pm

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostFri May 17, 2013 3:08 pm

Tamerlin wrote:
Paul Stone wrote:
Really, really low. Boosting it up in After Effects was the only way to even hear it, and the increase in volume revealed a bunch of noise. We used a nearly brand-new pair of Sennheiser (ew112p G3) radio mics.


I'm sure that would have been a lot easier to prevent if the camera had VU meters on it, so that you'd have some way of knowing what levels the camera was seeing in the first place. Sigh.


Having VU meters wouldn't make a difference unless you have a way of boosting your audio which goes back to my post. If the meters are reliable they would tell you that your audio was unusable. If BM is going to offer audio, then the expectation is that it will record at a minimally professional way. In the current state other devises are needed just to meet bare requirements (Zoom, Juicedlink, etc) or other software such as Sound Soap or Audacity.

Now I'm looking at my Canon 5D and my BMCC and debating which is better for what it offers. I know what I'm getting with the 5D. The BMCC doesn't deliver what's promised. The picture from the BMCC is phenomenal, but the lack of expected audio is a huge problem.
Offline
User avatar

Peter J. DeCrescenzo

  • Posts: 2205
  • Joined: Wed Aug 22, 2012 6:53 am
  • Location: Portland, Oregon USA

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostFri May 17, 2013 5:47 pm

IMHO, “looking” at the Blackmagic Cinema Camera’s current audio performance in isolation, Blackmagic Design has fallen short of reasonable expectations.

According to its specs, the BMCC hardware appears to be capable of recording “48 kHz and 24 bit” audio, has balanced audio inputs, and is switchable between mic & line level audio signals. If true, this means the camera is capable of recording audio that should sound fairly good, especially from line-level sources.

In actual practice, although this is somewhat true when using a BMCC with line-level audio sources, it’s definitely not true when working with mic-level sources.

I use Sennheiser ME66 & ME64 microphones with a variety of video cameras. Though not the best-sounding mics available, they’re quite popular, and when used properly can deliver results easily good enough for many professional applications. These mics can be self-powered (using a AA battery) or they can use external phantom power (such as from a camera or preamp), and they have much higher than average output levels — they are relatively “loud” mics (which is a good thing).

However, when directly connected to a BMCC (without an external preamp), the resulting sound isn’t very good compared to using them with most other cameras, including cameras that cost a fraction of the BMCC.

The following video is an example of what my ME64 sounds like when directly connected to my several-year-old Panasonic GH2 “DSLR”, which cost me less than 1/3 of the BMCC’s $3K price:


Not perfect, but quite usable, especially considering it was recorded using a $800 "stills" camera!

I don’t own a BMCC yet and I don’t have a sample video online that demonstrates what the ME64 & ME66 mics sound like when directly connected to a BMCC. However, I tested these mics with a BMCC I had on loan for a few days, and believe me, it didn’t sound anywhere near as good as a GH2 or any other pro or prosumer video camera I’ve ever used these mics with. Not even close. In my informal tests with the BMCC running firmware 1.2, the ME66 & ME64 (and also my Sennheiser wireless mics) sounded so bad (noisy! awful! terrible!) when directly connected to the BMCC (without aid of a preamp) I decided I couldn’t use them that way, at all.

Instead, I used my ME64 mic with my old SoundDevices MixPre preamp together with the line-level settings on the BMCC, and got the following result (see my notes on the Vimeo page for more info):


I think the results are quite good — not perfect, but completely usable for many professional applications. Based on what I’ve heard others achieve using various preamps (@ mic or line level), similar or better results are possible even when using preamps that cost far less than the MixPre (or current model MixPre-D), such as those from JuicedLink.

Going forward I hope BMD updates the audio portion of the camera firmware ASAP to “get out of the way” of what the BMCC hardware appears to be capable of.

Among other things, I think the low-frequency roll-off (hi-pass) filtering BMD added to the current version 1.3 BMCC firmware was a mistake, because BMD applied it to all BMCC audio, including audio from external audio inputs. The previous v1.2 firmware had other issues, but at least it didn’t feature an always-on hi-pass filter. BMD might have added the hi-pass filter to make audio from the camera’s built-in internal mic more usable (to reduce handling & fan noise), but it’s completely inappropriate for this hi-pass filtering to be always applied when using an externally-connected audio source.

Although I’d love it if BMD would please, please add audio level/VU meters displayed full-time on the BMCC LCD, I’d settle for a possibly simpler-to-implement “audio peaking” display, perhaps as simple as a small icon or dot that alternates between green & red to show when audio is either OK or clipping. That, combined with monitoring with good headphones, should be enough to get good, distortion-free audio recordings.

I don’t expect the BMCC’s mic-level performance to be as good as most standalone pro audio gear, but I do expect it to be at least as good as an old GH2. I don’t know for sure, but I suspect BMD can adjust the BMCC firmware so that using it with a good self-powered mic (or other self-powered mic-level audio source) without a preamp can result in good-quality audio recordings. But I could be wrong; maybe the current BMCC hardware can’t do it.

-

Having said all that, let me go back to what I said at the beginning: “Looking at the BMCC’s current audio performance in isolation” …

Given the BMCC’s considerable (awesome!) video capabilities, I actually don’t have a problem paying an additional few hundred dollars or so to add an audio preamp to a BMCC kit if that’s what it takes to also to get high-quality audio.

The way I look at it, the BMCC is an incredible value “as-is”, and for the user to have to add a few hundred dollars to its kit cost (buying a preamp to get really good audio) doesn’t really reduce its overall value all that much.

Yes, I definitely think BMD should fix the remaining firmware audio issues ASAP, but changing the actual camera hardware design — such as adding built-in high-quality mic transformers — is probably never going to happen to the currently-shipping or currently-announced BMD cameras. Hopefully BMD will do that in some future, unannounced camera, but at this point the hardware specs for the current BMD cameras (BMCC-EF, BMCC-MTF, BMPCC, and BMPC-4K) are pretty much set in stone.

I don’t think it’s unreasonable for BMD to expect users to add an external preamp for high-quality in-camera sound recording. But I also think it’s reasonable for users to expect that the BMCC should be capable of getting at least good-quality mic-level audio without a preamp, too.

-
Last edited by Peter J. DeCrescenzo on Fri May 17, 2013 11:51 pm, edited 2 times in total.
http://www.peterdv.com
http://HereForTheWeather.wordpress.com
Offline

Marshall Harrington

  • Posts: 412
  • Joined: Tue Sep 04, 2012 5:48 pm
  • Location: San Diego, California

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostFri May 17, 2013 6:16 pm

Well put Peter, thanks.

With all the new products being released I'm waiting a little longer continuing to use my H4n separately. I'll be curious to see the results of the new H6... if it uses the same preamps as the H4n then it won't work out. But if they are improved perhaps it will be useful as both a recorder and a mixer as it does have a separate line-out this go around. My experience trying to use my H4n as a preamp through the headphone jack that worked out very badly as many people around here also reported.
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSat May 18, 2013 2:52 am

Paul Stone wrote:We pretty much can't shoot any jobs requiring audio on the BMCC until this is sorted.


I disagree. We've been able to get "usable" audio from the BMC on both the 1.2.1 and 1.3 firmware using an external pre-amp. Unfortunately though with both firmware releases you need to do some post-processing to make it sound "right".

An external pre-amp/mixer is a given if you want to get good audio, regardless of what camera you're using. I wouldn't connect a ME66 directly to an Alexa ever if I was being serious with the audio - so you shouldn't expect you can just connect a ME66 directly to a BMC.

If you're doing anything that requires CRITICAL audio - then I'd still strongly suggest using an external audio recorder, and just use the BMC audio as a guide track and timecode reference. Even then, you still need a good external pre-amp/mixer.

Peter J. DeCrescenzo wrote:Given the BMCC’s considerable (awesome!) video capabilities, I actually don’t have a problem paying an additional few hundred dollars or so to add an audio preamp to a BMCC kit if that’s what it takes to also to get high-quality audio.


You can't just take the camera out of the box and start taking pretty pictures. You need a LENS. This is not included. Same deal with audio. If you want to record decent audio, you need to invest in good microphones and good pre-amps. Pre-amps are like filters. Yes, you can sometimes get away without them - but if you want to achieve the best results, you need the right tool for the job.

johnjvogel wrote:If I plug my ME66 into my BMCC 2.5k with 1.3 firmware - unusable audio.


This surprised me, so I decided to do some tests with the 1.3 Firmware...

TEST ONE:
- Sennheiser ME66/K6 and Rode NT3 plugged into a Sound Devices 552
- Both microphones powered by Phantom Power from the 552
- The 552 was set to record 48kHz 24-bit WAV's
- XLR Output from the 552 (set to Line) connected to the BMC
- BMC set to Line Level 100% on both inputs
- BMC connected to MacBook Pro running Blackmagic Media Express 3.2 for audio meters and the actual audio recordings
- Once the footage has been captured to a QuickTime File via Media Express, I then simply drag the file into Adobe Audition 5.0.2 (Build 5) for trimming and exporting
- For the EQ'ed examples below I just used the preset I came up with previously. HOWEVER, this EQ was developed when testing the Line Level performance of the camera, not the Mic Level, so some additional EQ'ing to make the audio sound "right" might still need to be done.

Here are the results:
- ME66 recorded on BMC (Unprocessed)
- ME66 recorded on BMC (EQ'ed)
- ME66 recorded on 552 (Unprocessed)
- NT3 recorded on BMC (Unprocessed)
- NT3 recorded on BMC (EQ'ed)
- NT3 recorded on 552 (Unprocessed)

TEST TWO:
- Next up I connected the ME66 and NT3 directly to the BMC
- As the BMC doesn't supply Phantom power, I had batteries in both microphones
- BMC set to Mic Level at 80% on both inputs

Here are the results:
- ME66 recorded directly to BMC (Unprocessed)
- ME66 recorded directly to BMC (EQ'ed)
- NT3 recorded directly to BMC (Unprocessed)
- NT3 recorded directly to BMC (EQ'ed)

TEST THREE:
- For comparison I also re-recorded the same dialogue to a Tascam DR-100
- The DR-100 does actually have the ability to supply Phantom power, but I decided to power the microphones off their internal batteries
- The DR-100 was set to record 48kHz 24-bit WAVs

Here are the results:
- ME66 connected directly to the DR-100 (Unprocessed)
- NT3 connected directly to the DR-100 (Unprocessed)

TEST FOUR:
- To test the performance of the ME66, I decided to put it up against a Sanken CS-3e (which is one of my favourite interior microphones)
- Both the ME66 and CS-3 were connected directly to the 552
- Both microphones powered by Phantom Power from the 552
- The 552 was set to record 48kHz 24-bit WAV's

Here are the results:
- ME66
- CS-3

Rather than give my opinion on the results, I'd love to hear what you think!

Best Regards, Chris!
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSat May 18, 2013 3:21 am

Here's another quick test comparing the Mic Level performance to the Line Level performance on the 1.3 Firmware.

In this example I have a ME66 on the Left Channel and a CS-3 on the Right Channel of a Sound Devices 552, which is then connected to the BMC.

Here are the results:

- Mic Level at 80%
- Line Level at 100%
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSat May 18, 2013 5:17 am

I've also just completed some more tests, comparing the frequency response between Mic Level and Line Level.

For this test, I downloaded a whole heap of test tones/sounds from the good folks at AudioCheck.net, put them together into one long file in Adobe Audition, then exported out a 48kHz 24bit 320k MP3.

You can download the test MP3 here.

I then put this MP3 onto a Sound Devices 552 for playback into the BMC. I then used Media Express to capture the audio from the camera.

Here are the results:

Benchmark Test Track
test.png
test.png (193.21 KiB) Viewed 27408 times


Line Level (100%)
line.png
line.png (189.45 KiB) Viewed 27408 times


Mic Level (98%)
mic.png
mic.png (190.76 KiB) Viewed 27408 times


As you can see, there is definitely quite a difference in terms of frequency response between Mic Level and Line Level.

You can download the results from the camera here:

- Mic Level @ 39%
- Mic Level @ 82%
- Mic Level @ 98%
- Line Level @ 100%
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSat May 18, 2013 5:29 am

It's not perfect by any stretch, but here's an example EQ setting you can use in Adobe Audition to make Mic Level recordings sound a bit more accurate:

BMC_MIC_EQ.png
BMC_MIC_EQ.png (56.72 KiB) Viewed 27406 times


Here's an example of Line Level recordings on the BMC after EQ'ing:

BMC_TEST_LINE_100_EQ.png
BMC_TEST_LINE_100_EQ.png (199.46 KiB) Viewed 27406 times


Here's an example of Mic Level recordings on the BMC after EQ'ing:

BMC_TEST_MIC_82_EQ.png
BMC_TEST_MIC_82_EQ.png (196.78 KiB) Viewed 27406 times


In both examples I've only used a single FFT filter.

You'd probably be much better off using multiple parametric EQ's, but sadly I'm not good enough with Audition/Audio Mastering to offer any useful advice in this department. If there are any audio gurus here that thing they can "level things out" - then please, by all means try and share your results!

Best Regards, Chris!
Offline

Paul Stone

  • Posts: 26
  • Joined: Thu Dec 20, 2012 4:58 am
  • Location: Queensland, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostSun May 19, 2013 11:35 pm

Chris Hocking wrote:
Paul Stone wrote:We pretty much can't shoot any jobs requiring audio on the BMCC until this is sorted.


I disagree. We've been able to get "usable" audio from the BMC on both the 1.2.1 and 1.3 firmware using an external pre-amp. Unfortunately though with both firmware releases you need to do some post-processing to make it sound "right".


Sorry, I was actually just referring to myself and my boss. It's never been necessary for us to use external audio gear in the past, so we don't have any such equipment on hand.

It honestly never crossed my mind that we'd have more trouble with the audio on the BMCC than we did on the Canon 5D, a stills camera.
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostMon May 20, 2013 1:45 am

Paul Stone wrote:It honestly never crossed my mind that we'd have more trouble with the audio on the BMCC than we did on the Canon 5D, a stills camera.


Yeah - it's pretty sad.

However it would be really interesting to compare the results of a Canon 5D versus the BMC running 1.3 firmware. I've shot lots of stuff with the 5D but have never actually recorded audio to it because it only records 16-bit/44.1kHz Linear PCM. If I get time, I'll try do some comparisons.

My gut feeling is that the final results will be very similar.

The BMC audio's not "bad" - it just needs a lot of extra post-processing, whereas the 5D gets it pretty close straight out of the camera (however you have no room to move).
Offline

johnjvogel

  • Posts: 22
  • Joined: Wed Aug 29, 2012 7:15 pm

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostMon May 27, 2013 8:46 pm

Chris Hocking wrote:
Paul Stone wrote:It honestly never crossed my mind that we'd have more trouble with the audio on the BMCC than we did on the Canon 5D, a stills camera.


Yeah - it's pretty sad.

However it would be really interesting to compare the results of a Canon 5D versus the BMC running 1.3 firmware. I've shot lots of stuff with the 5D but have never actually recorded audio to it because it only records 16-bit/44.1kHz Linear PCM. If I get time, I'll try do some comparisons.

My gut feeling is that the final results will be very similar.

The BMC audio's not "bad" - it just needs a lot of extra post-processing, whereas the 5D gets it pretty close straight out of the camera (however you have no room to move).



Chris, first I want to commend you for doing such a thorough job on testing the audio on the BMCC. My comparisons are not nearly as detailed. Having done sound for many years for TV shows like Extra, Access Hollywood, basically any ENG style show, I've worked it. My background is basically plug and play. So if the camera has line/mic, I expect it to work as any other camera that offers these inputs. I also expect certain situations to work when needed. For example if I don't have a soundman, can I place a powered boom mic to obtain audio.

So here's my short and skinny of this.
Firmware 1.3 - No - you will not be able to plug in a powered mic and record audio. You will need a booster.
Firmware 1.2.1 - Yes - you will be able to use a powered mic and record audio. You will not need a booster.

I rolled back my firmware to 1.2.1 for this reason. Audio sounds great and works like expected.
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostMon May 27, 2013 9:02 pm

johnjvogel wrote:So here's my short and skinny of this.
Firmware 1.3 - No - you will not be able to plug in a powered mic and record audio. You will need a booster.
Firmware 1.2.1 - Yes - you will be able to use a powered mic and record audio. You will not need a booster.


Sorry, but I'm not sure I agree.

The only difference between 1.2.1 and 1.3 is the scale of the levels - you still get exactly the same performance.

For example, Line Level at 65% on 1.2.1 is roughly the same as 100% on 1.3. In terms of performance, the quality will be identical - in both cases you're looking at 0dB attenuation and gain, so even if you could go past 100% on 1.3, you would just be adding digital gain which you can do in post anyway.

Remember, you can't just rely on meters with 1.2.1 to judge level because of the DC-offset. If you have UltraScope connected when you're doing audio recordings with 1.2.1 the meters will appear bigger (aka louder) than they actually are. You need to remove the DC-offset then review. You also can't really rely on the meters for 1.3 either, because a lot of the low frequencies are attenuated. Once you bring those frequencies back, the meters will appear bigger (aka louder).

The only differences between 1.2.1 and 1.3 are the DC-offset, frequency response, and the scale of the actual meters in the camera. The pre-amp performance hasn't changed at all. So if you're happy with the results in 1.2.1, then you should be able to get exactly the same end result with 1.3. In both firmware releases you still need to do post-processing to make the audio "correct".
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostWed May 29, 2013 11:43 pm

1.3.1 was just released, and I was HOPING that maybe the BMD developers slipped in another audio fix under the radar. Unfortunately, it doesn't look like they have, as we're seeing the same frequency issues as before...

BMC 1.3
Image

BMC 1.3.1
Image

Oh well... I guess we keep on waiting...
Offline
User avatar

Thomas Schumacher

  • Posts: 665
  • Joined: Tue Nov 06, 2012 11:14 pm
  • Location: Germany

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostThu May 30, 2013 6:32 am

Thanks for letting us know, Chris!
My award-winning doc: https://vimeo.com/91622063
ARE YOU HAPPY? https://vimeo.com/72724154
http://www.gernemehrfilm.de/
https://berlin3500.wordpress.com/
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostFri Jun 07, 2013 1:34 am

For anyone's who's interested, I finally got a response from Blackmagic earlier in the week. They are very much aware of the issues in the 1.3 firmware, and are planning to fix them in the future. When... who knows, but at least they're noted as bugs. Fingers crossed 1.4 will be released soon with these issues fixed!
Offline
User avatar

LDS

  • Posts: 14
  • Joined: Tue Dec 11, 2012 8:48 pm

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostTue Jun 11, 2013 10:55 am

Thanks for all of your work on this Chris. I'm on a shoot tomorrow where I will be setting the camera to LINE IN @ 100% level. I've done some tests recently and can recover the bottom end, but I'm still hearing a lot of hiss on the high end. Wondering if you noticed that as well?
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostTue Jun 11, 2013 1:14 pm

LDS wrote:I've done some tests recently and can recover the bottom end, but I'm still hearing a lot of hiss on the high end. Wondering if you noticed that as well?


Yep - as you can hear in the tests I've done previously, there's still a fair bit of noise in all of the recordings. Luckily, if you've got good clean dialogue, you can clean this up to a useable state in post.

Really though... for anything where you need really great audio - just use an external recorder (and send a guide track and/or timecode to the BMC).

Good luck with your shoot!
Offline
User avatar

John Bartman

  • Posts: 351
  • Joined: Wed Aug 22, 2012 3:15 pm

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostMon Jul 22, 2013 3:26 pm

Really though... for anything where you need really great audio - just use an external recorder (and send a guide track and/or timecode to the BMC).


Hi Chris,

whats the best way of
"....sending a guide track and/or timecode to the BMC" ?

(I have a Marantz PMD 661)
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostThu Jul 25, 2013 7:42 am

John Bartman wrote:whats the best way of "....sending a guide track and/or timecode to the BMC" ? (I have a Marantz PMD 661)


Sorry for the delayed reply John - I'm currently interstate on a job.

In general, I would send a line level guide track to the Left Channel of the Camera and a line level timecode signal to the Right Channel of the Camera.

In regards to the Marantz PMD 661 however - it's a prosumer device, so unfortunately it doesn't handle timecode at all. If you're mounting the 661 on the camera, then I'd just take the signal from the RCA outputs on the 661 and connect it directly to the camera (ensuring you have the right cabling - as you need to remember that the BMC has BALANCED inputs - otherwise you can run into phase cancellation issues). If you don't want to mount the 661 on the camera, then I'd just use a wireless transmitter/receiver to get the signal to the camera. You can then use something like PluralEyes, FCPX or Premiere CC to sync the camera's guide track to your 661 master audio.

Hope this helps!
Offline

Jules Bushell

  • Posts: 1023
  • Joined: Wed Sep 05, 2012 3:06 am
  • Location: London, England

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostThu Jul 25, 2013 11:07 am

Firmware 1.4 is out now. I'm currently on 1.2.1

I'd be interested to know if they fixed the recording external mic levels and the DC offset issue? Or have introduced more audio problems?

Jules
Jules Bushell
url: www.nonmultiplexcinema.com
url: www.filmmeansbusiness.com
url: www.blurtheline.co.uk
Offline
User avatar

Chris Hocking

  • Posts: 659
  • Joined: Wed Aug 22, 2012 4:23 am
  • Location: Melbourne, Australia

Re: BMC Audio Performance - 1.2.1 vs 1.3

PostThu Jul 25, 2013 11:43 am

I'd be interested to know if they fixed the recording external mic levels and the DC offset issue? Or have introduced more audio problems?


The DC Offset issue was fixed in 1.3. As far as I know, there's never been any issues with mic levels - however they did change the gain structure in 1.3 so that 100% in 1.2 is different to 100% in 1.3.

I'm currently away from the office, so sadly I can't check out the 1.4 release yet, however I'm extremely hopefully that they've fixed the frequency attenuation issue in this 1.4 release. If they haven't, then that's just bizarre.

Hopefully someone else is able to do a test and let us know if anything's changed!
Next

Return to Cinematography

Who is online

Users browsing this forum: Andrzej Kryt, Dmitry Shijan, Jamie LeJeune, Kim Janson, Majestic-12 [Bot] and 18 guests