Messages

Hi guys,

would you have the logical audio settings for working with a preamp/DT454 running ML on a Canon 5DII,
that go beyond the info being mentioned in the Juicedlink Manual (pages 6 and 7)?

I´m having a hard time figuring out if to work either with or without analog gain at all on the ML set-up, where to best disable AGC and which settings in the Canon menu (automatic/manual audio control) to best choose from.

Many thanks in advance, any link or hint to the correct set-up would be appreciated a lot. I was rushing and reading through forums but couldn´t find any detailed audio settings guide for working with pro-/lav-mics and using preamps on ML with the 5DII..
sorry for crossposting.

Best regards,
chris

Hi Okaro,

I just waded into a similar post  here.  My tests with my 550D showed that when you disable the AGC in a Canon that you're virtually back to a line level input.  So any good pre-amp running line level out will sound good in a Canon with ML -- just zero all analogue gain -- and never use digital gain as it just adds noise.

This brings us back to JuicedLink's claim that mixer pre-amps specifically (they say) the Sound Devices Mix-pre D was "padded down".  This, of course, is rubbish so it's suspected that JuicedLink's mic out signal is somewhat boosted, because they know full-well that you need more than mic level to get sufficient signal into a camera's mic input.  This can be verified by simply looking at why the Rode Videomic Pro has a 20db boost on its mic. 

So in summary with the JuicedLink I would try a couple of things:  Firstly to match a cable that delivers its line level output to the camera's input with the AGC disabled.  Secondly to use JuicedLink's mic output and simply add the necessary analogue gain in ML to bring the meters up to between -6 and -12db.

I'm not sure if JuicedLink has one of those AGC disable where it puts a phantom tone in one channel to fool the AGC -- but that certainly wouldn't be my choice when ML does this much better.

Cheers

Hi Mate,
I've used the H4N, but I'm not too familiar with the 5D2--I've got the 550D.  I suppose what you're trying to do is utilize the H4N's XLRs and mic-pres to get a good recording and then send that to the 5D2 to  use as your audio track -- rather than line up the H4N's audio.  I suppose you have to look at the rationale behind doing that.
You would be bringing a mic signal into the H4N and using its mic pres to boost the mic signal up to line level.
As you know what good mic pres do is make a nice clean boost to line level from a signal that is sometimes as much as -60db -- that's 60db below line level.  So if the mic pres aren't nice, noise-free amps they just boost any inherent noise as well.  That's the theory.   But look at what you are proposing: To boost that low level mic signal in the Zoom and then pad it back down to mic level and then boost it again in the Canon using its mic pres.
Well that's the theory but my tests have shown that it doesn't even work that way. Damn! 
In the 550D there's that troublesome AGC, which ML can disable.  I think you can manually disable this in the 5D2, but I'm not sure. 
But once you disable the AGC (Automatic gain control, which raises and lowers gain automatically) my tests have shown that the camera's input is back to line level.  So buying a cable that is padding your signal to mic level would be pretty useless.
Since you have the Zoom and the camera a simple experiment would be to disable the AGC in the camera and simply run line level into the camera (not sure if you can attenuate this from the Zoom, but that would help).

In summary, I don't have a problem with the 550D's audio recording quality, in fact, I've found it's quite good. But once you disable the AGC--because it's noisy--you need to pre-amp any microphones before going into the camera--like with a mixer or a mic pre amp.

And just a word on XRLs.  The only advantage of an XLR is that it is balanced.  Balancing simply means that the same audio is run in-phase on pin 2 and out-of-phase on pin 3, and pin one is earth.  What happens is that any noise picked up by the cable cannot be exactly the same  on both pins 2 and 3 -- because it's intermittent noise.
But any audio originating from the source (like the microphone) is, of course the same on both bins.  So a summing amp at the input stage of a balanced XLR input simply discards anything that is not exactly the same on both pins--like noise.  This is useful on long runs where cable become susceptible to noise--but generally on short runs this is not a problem. 

Of course you can't lock the aperture to a value that is not shared among all the focal lenghts but it is so evident that i thought it was unnecessary to highlight this fact.

Yes I know what you mean. It's like when they bring a weapon onto a film set, it's evident that it wouldn't be loaded.
I get it now.

Cheers.

No where does the OP say anything about this being an issue of a non-constant aperture zoom. It sounds like they are unsure how to set exposure to manual in video mode.

Hey Francis,
No he doesn't mention about non-constant aperture zoom, and it may simply be as you say a case of how to set the f/stop manually, but the information is very relevant if his lenses aren't constant aperture lenses.  For instance when I first started out with a DSLR shooting concerts I'd set the widest opening because of the lack of light and as I'd zoom the aperture would blow out to f/ 5.6 -- then a rival photog at a concert told me about constant aperture lenses, and I was thankful for the information.

Cheers

18-55 IS has a biggest common aperture of f/5.6 so i don't see the problem to lock the aperture to that value.

I suppose what you are saying is that the 18-55 IS has its widest opening aperture of f/5.6 when at its max focal length of 55mm, and yes of course you can fix that so it stays at f/5.6 with Canon's own firmware to stay at that f stop even at its 18mm setting pulled right back.   But its widest opening aperture at 18mm is f/3.5 and you can't lock that to stay when you zoom out to the 55mm focal length, as it will automatically go to f/5.6.
Since eightcore asked can you "lock the aperture during video recording," I think the information is relevant if he is not using a constant aperture lens.

Cheers

If you have an f/3.5-f/5.6 lens, you can always shoot at f/5.6 so that it remains constant at both sides of the focal length. If you're shooting at a lower aperture (for example at f/3.5 when the lens is at 18mm), don't use the zoom (besides, it's next to impossible to get a smooth zoom), especially with non-L lenses which I'm assuming you're using (from the non-constant aperture).

Hi Mal,
Just to clear things up.  I don't have a problem.  I'm simply stating that the author cannot fix a constant aperture unless the lens is a constant aperture. 

[Language]

Blame your lens then. The aperture will stay the same if that f/stop is available at each focal lenght you zoom through. Else, it will obviously change.
As said a million times, ML is a software, it can't do miracles

edit by nanomad: Language

Hmmmm (I like saying hmmmm)...The author asks, "Is it possible to lock the aperture during video recording."
And I'm very clearly saying that unless your lens is a fixed aperture throughout the zoom range, you can't!

Can someone please shoot me because I really don't see how that can work.  If your lens is a zoom and not a fixed aperture the iris is going to change every time you zoom in or out. 

It's already been said, but I'll try to put it another way.  1080p is 1080 lines of resolution each line having 1920 pixels. As a fraction 1920/1080 when you break it down to its lowest common denominator is 16/9. 
If you multiply 16 by 120 you get 1920 and if you multiply 9 by 120 you get 1080.  So 16:9 is the lowest common denominator of 1920/1080.  All LCD, Plasa etc screens are set up to display that resolution without cropping or 'letter-boxing'
Hope that makes sense.

I'm curious as to what's happening with the video signal in that cable.  Do they just terminate it somewhere?

The very first test I would do would be to turn on your 550D holding the camera's set button.  This will temporarily disable ML for that session.  Now that you have eliminated ML from your troubleshoot you should simply plug in only the 550D and test it in record mode and also playback mode.  Set the camera to 1080 / 25. 
I'm not sure how the Blackmagic displays the resolution -- like does it display what it's receiving or what it's playing out -- and do you know if it's doing any upconverting or downconverting?
But at least you'll know that it's not getting confused with ML or even the other inputs.

I did a bit of cheating...er...Googling, as I'm not familiar with the Blackmagic mixer.  But the general blurbs say that all the inputs should be receiving the same resolution for it to display all the units.

Sounds like your TV/Monitor is set to NTSC and doing the converting.  1080i/60 is the NTSC setting.  However your camera setting of 1080p/25 is, as you say, PAL.  Are you in NTSC territory?   It's not uncommon for a TV to convert an incoming signal to its own native setting.

The camera is capable of replaying 1080p when not recording, and a lower resolution via HDMI when in record mode (some say 480P).  This, of course, isn't a problem because monitoring is done on smaller monitors--and HD is only more pixels that are only appreciable on large monitors--like your LCD, Plasma or big screen.

If  you're talking about headphone monitoring -- although I haven't had a great deal of DSLR experience -- I've monitored cameras way back to the analogue Betacams, and you certainly can hear if you are overloading, clipping, distorting, square-waving or whatever you want to call it.  Distortion is heard in different forms. Often if you overload a signal it will become quite shrill ( sharp in the high-mids). 
First thing that comes to mind about monitoring via meters is that you constantly have to watch them, when perhaps you should be focused on the visuals.  Another point is that meters don't always accurately reflect the sound -- and there are different types of metering, like PPM and VU etc (some averaging, some reflecting peaks).
And speaking of trusting meters, I recently used a JVC camera with balanced XLR inputs -- and meters, and those meters are absolutely useless -- even after lining them up with a slate  tone.
The professional way to monitor audio that's going direct to camera is usually done via a mixer that's connected via a snake cable to the camera. Part of that snake is a connection to the camera's own headphone jack.  You plug your headphones into the mixer but you can also switch to listen to the camera's headphone output -- that's because the mixer might sound fine but you may have overloaded the camera's input and as I've said, you certainly can hear that.
There's also that other thing that always amuses me about audio: I was once told by a filmmaker that a good audio recordist had to "listen for certain things".  Ah yes, that's the mystique of an audio guy, he can hear "certain things" that lesser folk can't -- and that's a big load of BS, but always good for a laugh.  I've been on shoots where actors drag their chairs across the floor whilst delivering their lines, and even one next to a railway line where nobody noticed that a train was passing (and let's not talk about aeroplanes).  Of course sticking a pair of headphones on simply keeps you in-tune to those "certain things"  that shouldn't be on the audio track: air conditioners, generators, refrigerators and even the talent slurring their lines.
Lastly, I'm fastidious about room reverberations. Reverberant or 'echoing' audio -- like empty rooms-- is the true mark of the amateur.  Very often the human ear (my preference) will subconsciously filter out room resonance: have you ever seen those  audio guys that walk into a room and clap their hands?  They're really not sure are they?  But you soon get the feel for troublesome rooms, and stick a pair of headphones on and you'll wonder how you missed it in the first place.
   

Go to the audio menu, Press MENU, then highlight 'Input Source' and press SET.

Hey, that worked!  Thanks Francis.

At first the audio INPUT SOURCE showed up on the menu.  Then it went into hiding.  I fished around and found where it was to mess with the settings.  But is there a way to make it appear permanently on the menu like it was in the first instance?

Cheers

Is it possible to make a setting "Audio" in the section "Input Source", reversed L: int R: ext, so that a record was with L: ext (external) and R: int (internal)? Is it possible to implement a new version of the firmware? If not why? The fact that I have an external microphone, which provides mono sound on the right to the left channel.

Cameras always provide a stereo input for the microphone, but, of course the majority of mics are mono.  But most mono mics that terminate into a 3.5mm jack simply wire the hot (signal) to both tip and ring (instead of left and right stereo).  You can easily rewire your 3.5mm jack to the tip and ring (sleeve is ground).  I'm guessing (because I haven't tried it) that ML simply takes only one side of the ext input in EXT/INT mode and dedicates the other side to the internal mic.  If it doesn't (which I doubt) you can simply switch your hot pin on your 3.5mm plug.

I did have the Mic gain up to minus 23db.

Hi,
I'm not being pedantic, but for those following this, that's +23db.  That's ok if the AGC is off, but if the AGC is on it's going to give you a lot of gain.  Do you remember what the meters were doing, and was the AGC on or off?

I think I've located the problem. After freaking out that I'd damaged my camera I read in the manual that you have to use the canon a/v cable coupled with an adapter to hear anything out your headphones. I wasn't doing that and was trying to go directly into a/v jack with my headset.
The issue of very crackly audio being recorded on the mic circuit with new rode video mic may have something to do with using old batteries. Im going to try out a new one tonight and see if that improves the situation. Thanks heaps for the kind responses.

It might also be worth checking some other stuff -- especially when you might have had 2 separate problems, which is often hard to solve  unless you deal with one at a time.  I'm not sure how the 5D works but you'd certainly want to disable the AGC because that not only causes noise (mostly hiss) but adds a heap of gain.  I'm saying this because I'm not sure if you have the Rode Video Mic pro, which also has a +20db gain switch.  And to add to that if you left the ML's analogue gain up around 23db it might all be overloading the input. 
I would start by reinstalling the ML. Disable the AGC, and also zero all gain including the ML output gain -- especially the digital gain, which is noisy.  Then try the rode -- if it's got the 20db gain switch it on.  If it hasn't got that you might need to add some analogue gain on the ML.

I spent some time testing audio with my new ML software and 550D.  Firstly, after you disable the AGC, you are virtually left with a line level input -- not really a mic level input -- in the Canon's mic 3.5mm input.  This can either be good news or bad news, depending on how you look at it.  I tested this by first plugging my battery operated Sony ECM-979 electret condenser mic into the input with the AGC off.  It's nice, but the signal is way too low.  I boosted the analogue gain up to 32db (max) and the output volume to max.  The digital gains just add noise, so there's no point in using them.  The analogue and output gains don't add any noticeable noise, but the signal is still way too low. So, again, it's not much help calling this a mic input without the AGC, which seems to do all the work--badly.
This simply means that the camera's input is not set up to correctly boost a mic signal up to line level (sometimes a boost of some 60db is required).  This mic (Sony) also has a nice signal with the ACC on and lots of gain, but the AGC is too noisy--and not only noisy but adds a not-so-nice pre-amplification.
Next test was to take the MAIN OUT L/R from my little Mackie 402-VLZ3 mixer at LINE LEVEL into a 3.5mm jack for the camera's input.  I used a studio Rode NT2 condenser (phantom power on).  I set the Mackie to Unity gain at the mics INPUT GAIN, at the VOLUME FADER, and at the MAIN MIX out: Unity gain simply means that after the mic is pre-amped by the Mackie's mic preamp  that the input and output are at the same gain -- like it's not adding level.  This test recording was beautiful: The camera was quite happy to accept the LINE LEVEL signal.
I did set zero on the ML analogue gain, digital gain and output volume (and yes, the output volume does affect the main signal strength, and is not just for headphones as some have stated in Internet blurbs).
So the good news is that if you use an ENG mixer ( not the Mackie as it isn't battery powered) you can simply come out at LINE LEVEL if you keep the faders down: punching the faders up too far would cause overload.
And the bad news is that unless you have something like the Rode Videomic with its 20db boost you aren't going to get quite the required level from just plugging in a microphone with the AGC off.  You can get away with it if you boost the signal in post production, but I've noticed that this does add a little noise.
And it's a good idea to wait a couple of seconds after pressing the camera's record as the ML shoots out a big burst of audio whilst it's getting its act together.   

I am right in thinking that I should be able to use the a/v to listen to camera audio?

As Malcolm said, make sure you use the correct colour plugs--as in red and white for audio, because if you use the yellow plug (video) in an audio circuit it will certainly give you a similar kind of noise.  Also, the general blurbs say that you must only use the Canon supplied AV connector.  Canon say that the USB connection is a 'digital terminal' but to me that doesn't make any sense because those red, white and yellow leads deliver an analogue signal.  I can't believe that there is some little D/A converter in the plug, so I'm thinking that the only reason Canon say to use their cable is cause of the pin configuration.  I'm guessing it's all analogue and they just used a USB for convenience.

If no opinion is given it means that it doesn't matter. You can either keep everything or delete the pdf and the other .fir. I usually keep everything because so that I can demo and use ML on other cameras

Ok, so I go with the 550d fir (my camera). There's also an autoexec.bin file -- and the ReadMe file more or less says that you need this so the camera will boot.  So I guess I'll go with that.  I'm not sure about the ML folder, but I guess I'll go with that too.  A lot of guessing, but what the hey.

Pretty sure you don't need the other firmwares and the PDFs. I just unzipped the whole zip file to the card as I couldn't bother deleting them.

Thanks for your reply, Malcolm,

I suppose given that the whole file is only 5Mb it wouldn't hurt to dump the lot onto the card.  It's just a bit off-putting that there aren't any other opinions that could clarify this.  Wonder what else is left up to guesswork?

Cheers

Zug

Hi Guys,

This one's got me puzzled.  Every tutorial I see on installing ML 2.3 says to install all the downloaded files.
I've got the 550D and I see the 550D-109.fir file.  So do I need the other .fir files?   There's also an autoexec.bin, the two PDF install and user guide files and then there's the ML folder with all its stuff.  So exactly what do I need to copy to my SD card?

Cheers

AGC should have nothing to do with the amplifier. An amplifier simply "boosts" or "pulls up" the signal level of a fixed amount (i.e. +5db). AGC acts like a variable amplifier, aiming for an average "audible" signal level throughout the whole recording thus boosting silences and producing horrible hisses. So, unless the engineer was dumb, AGC off should simply disable AGC while leaving mic boost untouched. You may need a higher boost though (due to AGC being off and not trying to pull up the signal by itself)

Yep, understand all that.  Used to use AGCs or Levellers + mic pres in the analogue world.  Actually mics can be down around -60db and the mic pre usually pulls that up to around +4 or 0db line level. 
I'm quite happy with the Canon mic pre but the AGC is giving me a prob.
But thanks very much for your response, as I'm getting more confident that what you are saying is correct of the Canon.  What threw me off is the actual ML video, which wants you to use a JuicedLink pre-amp -- but when I listened to the JuicedLink video instructions, although they keep singing the praises of their mic preamps, you ultimately have to connect the unit's microphone output to the camera's mic input.  Of course this runs against everything I've ever done in my life: Like why use mic preamps -- which were always designed to boost to line lever -- and then pad the signal back down to mic level for the camera's mic preamp to re-boost.  Doesn't make any sense at all to me. 
So I think you have answered the question.  Thanks.

I'll try another approach:  Has anyone used the ML audio feature of disabling the AGC and then plugged a mic straight into the camera -- and what was the signal like?