Calibrating 1khz tone analog on a digital device?

T

Toenis

Well-known member
I really don't remember anymore how would one set up 0dbu 1khz tone from analog equipment on a digital camcorder, just for reference. Somewhere at -20dbu?

I feel stupid here, it's night right here and when I wake up I'll most definitely have answers here on this basic question.

Thanks,
T
 
-18dBFS is the EBU standard.

In the USA they use -20dBFS.

As you are in Estonia, I would use the European -18dBFS standard.
 
Thanks, that's just what I wanted to know, it has been about two years when I did such a calibration myself.

Thanks,
T
 
Keeping in mind that I am still on the steep side of the audio learning curve I question whether recommending sending tone at -20 dbfs and mixing with confidence is a good approach for beginners. Many of us are still learning how to use the mixer read outs, and the camera readouts, and a play back, and especially our ears to lay down usable audio. As I spend more time with a wide array of cameras I am learning some hard lessons. I had a recent run-in with the JVC-250 where the -20 rule didn't seem to be enough. Adding Oleg's approach (I will leave him to fill in the blanks as I don't want to speak for him) of working back from the cameras upper limits was a different but better approach in that situation where the camera's metering did not reflect its capabilities.

I hope the old hands will once again chime in on this topic as this is a job breaker and yet no one approach seems to fit all cameras. I am running into owner/operators who are changing cameras as new versions come out. They use mixers intermitantly. They find comfortable recording levels for direct in and do not know (some times defensively) how their camera reacts to industry standard feed levels. It is up to us to walk in and hit the sweet zone. Sometimes I have doubts I am earning my paycheck. Every town is a small town when you screw up.
 
If I use XH-A1 and SD302, why would I use -18 dBFS here in Europe and -20 dBFS in America? Do these tools have a secret GPS reciever which tells them on which continent they are and adjust themseves accordingly?
 
-18 is certainly a safe bet.

Baker: in case there's any confusion, we're talking about a 1K calibration tone to represent the equivalent of a 0 VU analog meter reading. That's all. Record level for program material is another matter, and should probably average around -12, with peaks around -6.
 
Thank you Ted. Yes there is plenty room for me to stray in these matters. I understand this thread to refer to calibrating analog mixer meters (or other) to digital equipment where 0 dBFS is the maximum acceptable to the processor and the mixer achieves, in an ideal world, 18 to 20 db of head room above 0dbu. Correction is welcome.

Cutting to the chase, how do you handle a camera that you haven't worked with? I never have problems with the pro units. Its the new crop of pro-sumer cameras that are good enough for money making and yet not quite prime time. I can't always find info on line (factory or forum) that prepares me for a shoot. Some need hotter levels or a narrower band of acceptable gain. I am trying to build a ref sheet listing some of the quirks as I read on or work with different units. Having a technique to qualify each camera would be a huge plus.
 
Last edited:
Baker Park said:
Thank you Ted, yes there is plenty room for me to stray in these matters. Cutting to the chase, how do you handle a camera that you haven't worked with? I never have problems with the pro units. Its the new crop of pro-sumer cameras that are good enough for money making and yet not quite prime time. I can't always find info on line (factory or forum) that prepares me for a shoot.
Click to expand...

First of all, keep in mind the purpose of the alignment levels when recording is to be able to correlate what you see on the mixer's meters with what's going on inside the camera - that is, to align the mixer's meters to the camera's meters so that looking at one will tell you what's going on with the other. This is a different thing altogether than the purpose of reference tone on the final program deliverables.

SMPTE's -20dBFS and EBU's -18dBFS specifications fall into the category of standards for program deliverables and even there it's not exactly what it seems. The actual SMPTE specs call for -20dBFS digital to correspond to signal level of +4dBu but the EBU specs are -18dBFS = 0dBu. That means in turn that it matters whether your analog meters are calibrated for 0VU = +4dBu or 0VU = 0dBu. If you had a mixer such as the SD or a Mackie desktop and were aligning it exactly to SMPTE specs, you'd set 0VU tone to read -24dBFS on the digital meters since they both are sending a signal of 0dBu when their meters are reading 0VU!

Remember that decibels are a ratio of one level relative to another and do not indicate an absolute value. Also most DV camera meters and also those on modern digital recorders are NOT VU meters - they're peak reading meters calibrated in dB below Full Scale. The designers of a digital recording device decide what voltage will be represented by all ones digitally and that voltage, whatever it may be, in turn defines the 0dBF(ull)S(cale) value. Recording at -20dBFS means the signal you're recording has 20dB of headroom before it reaches the voltage that translates into all 1's and you clip; -12dBFS means there's 12 dB headroom before clipping, etc. By using dB relative to full scale it doesn't matter what specific voltage the designers chose as the full scale limit - if your signal is 20 dB below it you have 20 dB headroom before you hit the clipping point. By aligning the analog mixer's meters so that 0VU on the mixer equals -20dBFS on the camera's (or recorder's) meters you know that when the mixer meter is bouncing around 0VU with real world signals, you're recording with 20dB headroom before you clip. The sensitivity of the camera input, the specific voltage the designers chose for their 0dBFS reference point, becomes totally irrelevant.

What is more importance is the nature of the sound you will be recording - how high are its peaks over its average level. If you have a mixer whose meter allows you a choice of ballistics such as the SD mixers and you send it a tone reading 0VU when the meter is set for VU, then switch the meter to peak reading without touching levels you see the reading doesn't change much. But if you send it speech and adjust the levels to the meter reads 0VU when set to VU mode, then switch to peak reading mode, you'll see the peaks read signifigantly higher. (One of the nice features of the SD meters is they offer a combination mode that shows you both average VU and peak at the same time. On speech the peaks run about 6 to 8dB or so above the average. )

If you align 0VU on the mixer's analog VU meter to, say, -18dBFS using tone and then switch to speech metering the same 0VU on the mixer, you'll see the meters in the camera bouncing up around -10 or -12 dBFS. If you set the mixer's meter to peak reading on the identical signal it would be reading +6 to +8, 6 to 8 dB over 0VU just like the camera meter is reading 6 to 8 dB higher than the "tone up" level.

With 16 bit recording you want your average levels high to give you as much room above the noise floor as possible. At the same time you don't want them so high that you risk peaks hitting clipping. Getting within a few dB of full scale is starting to tickle the dragon's tail so you want to set an absolute never exceed limit with a saftey zone, of say -3dBFS. Assume speech has peaks 8dB over average, again to keep a safety zone. -3dB -8 dB = -11 dB. Setting average 0VU at -12dBFS (the red dot on a number of DV camera meters) lets you record as far above the noise floor as you can without your peaks getting too close to the clip point.
 
All right, some of this is finally sinking in. So assuming we are using a SD mixer and recording non-dramatic dialog and sending line level would you address the following...

1. When recording to a 16 bit camera can I set the limiters on the SD to an equivalent setting of -3dBFS (on camera meter), then hit it lightly every few seconds and end up with a track that still benefits from the relatively hotter settings?

2. How do you choose between running the camera gain almost full up or running the mixer gains very hot? I have been hurt by getting tone to where I needed it by turning up the camera only to find the camera distorting the incoming signal. If you have to guess the answer with out any tests which way do you lean? Is running a SD mixer gain and pots at 4/5s (or more) acceptable?

3. Getting back to 24 vs 16 bits. Am I correct to consider running 4-6 db hotter for the 16 bit camera? Especially with cameras that have a reputation of being noisy. I am a little confused as we have discussed average vs peak readings. My problem not yours.

Thanks for the extended answers. I am trying to walk away from this thread with these nagging questions under control.
 
Baker Park said:
All right, some of this is finally sinking in. So assuming we are using a SD mixer and recording non-dramatic dialog and sending line level would you address the following...

1. When recording to a 16 bit camera can I set the limiters on the SD to an equivalent setting of -3dBFS (on camera meter), then hit it lightly every few seconds and end up with a track that still benefits from the relatively hotter settings?

2. How do you choose between running the camera gain almost full up or running the mixer gains very hot? I have been hurt by getting tone to where I needed it by turning up the camera only to find the camera distorting the incoming signal. If you have to guess the answer with out any tests which way do you lean? Is running a SD mixer gain and pots at 4/5s (or more) acceptable?

3. Getting back to 24 vs 16 bits. Am I correct to consider running 4-6 db hotter for the 16 bit camera? Especially with cameras that have a reputation of being noisy. I am a little confused as we have discussed average vs peak readings. My problem not yours.

Thanks for the extended answers. I am trying to walk away from this thread with these nagging questions under control.
Click to expand...

Alignment lkevel and recording level aren't necessarily the same thing. First set up the mixer ao its nominal output level matches the camera's expected inputs: +4dBu professional line level, -10dBv consumer line level ("tape"), or a mic level. Most prosumer cameras expect -10dBv line level but check the documentation, there are some that are wildly different ( for example the Canon XH-A1 line input is rated at +8dBu - Lord knows where they got that! - and you get better results sending mic level from the mixer to the mic inputs). The SD 302 and 442 use slightly different methods to make that adjustment - the 442 has switches on the output to select mic, -10, or line while the 302 uses a setup menu setting to pad the output level down to the desired level. Mixer faders to midpoint, output fader to unity gain. Send 0VU tone to the camera and using ONLY the camera's controls, set its level so the camera's meters read the desired alignment point. Align to -20dBFS and you should find the controls are about 2/3 up when that happens. Tone off and do a sound check on the mic. Don't touch the mixer output fader. Leaving the channel faders at unity (midpoint) - on the 442 you can hold the PFL switch on, on the 302 set the headphone switch to PFL - use the channel gain input trim until the speech being recorded hovers around 0VU. If you're recording to miniDV with a -12dBFS nominal level increase the output fader so the meters average around +8 VU, the orange to red colour change on the 302, the first step into the orange colour on the 442. Use the channel fader knobs to ride gain during the actual take. Setting the meter to the combo PPM/VU mode gives you a good idea of what's happening to the peaks while monitoring the average level as well. This method sets things up so the mixer's limiters (default +20) will kick in at full scale in the camera - some users like to set the limiters down to +18dBu to give a bit more safety. Max headroom, max dynamic range above recording noise floor.
 
Last edited:
Steve House said:
Most prosumer cameras expect -10dBv line level but check the documentation, there are some that are wildly different ( for example the Canon XH-A1 line input is rated at +8dBu - Lord knows where they got that! - and you get better results sending mic level from the mixer to the mic inputs).
Click to expand...

Not so. You get 10 dB better S/N ratio (over 90 dB) by sending XH-A1 the full line level signal from SD302 mixer with camera potentiometers fully open, than by lowering the signal level and using mic input (S/N about 80 dB).

XH-A1 is designed to be used with potentiometers fully open when using line level signals. The level alignment is easy and perfect.
 
most professional cameras such as sony 970-900r or equivalent aligned to get +4dbu (0vu) for -20 dbfs and even then you can adjust input sensitivity ( attenuation/ or more gain ( i have to check it ones:-)) threw internal menu )
 
Steve, thanks for the time on these anwers. I am still flailing on this.

Doesn't Petrus post provide an example of the difficulty in finding the right amount of mixer gain vs camera gain. If I understand his point the XH-A1 needs to be turned full up. I recently did a running shoot with a PD-170. The self noise on that camera seemed very high when turned much past mid point. My bad experiance with a JVC-250 was the result of being too conservative with the on camera gain. I had my mixer at least 4/5s up, gain and pots. This has been mentioned in other posts.

So back to questions 1 and 2. Can I run with SD gain running near max on these cameras needing high input? How tight are you willing to run to the limiters. One book sold by Trew Audio says hit them often. Do SD mixers still run fairly clean even at high gain?

Lastly is running 4-6 dBFS hotter about right when recording with 16 bit settings? Is this only when using pro-summer cameras or true of most units including full pro cameras?

Sorry to flay the nag.

I may get beat up on this but it is worth it.
 
Baker Park said:
Steve, thanks for the time on these anwers. I am still flailing on this.

Doesn't Petrus post provide an example of the difficulty in finding the right amount of mixer gain vs camera gain. If I understand his point the XH-A1 needs to be turned full up. I recently did a running shoot with a PD-170. The self noise on that camera seemed very high when turned much past mid point. My bad experiance with a JVC-250 was the result of being too conservative with the on camera gain. I had my mixer at least 4/5s up, gain and pots. This has been mentioned in other posts.

So back to questions 1 and 2. Can I run with SD gain running near max on these cameras needing high input? How tight are you willing to run to the limiters. One book sold by Trew Audio says hit them often. Do SD mixers still run fairly clean even at high gain?

Lastly is running 4-6 dBFS hotter about right when recording with 16 bit settings? Is this only when using pro-summer cameras or true of most units including full pro cameras?

Sorry to flay the nag.

I may get beat up on this but it is worth it.
Click to expand...

I think Petrus' criticism is regarding my suggestion to send mic level to that particular camera rather than running line level and opening the pots all the way. In fairness, I've seen both strategies recommended for that particular camera with no definitive arguments one way or the other. My suggestion is based on the idea that manufacturers usually design amplifiers to run their best with the gain about mid-range. Some cameras and recorders have 2 separate preamps for mic and line level, other have only one and switch a pad into the input when it's set to receive line level - line level comes in and is padded down to mic level and sent to the mic preamp. I honestly don't know which system that camera uses. I'd love to see a block diagram of its audio circuits to see what is really going on.


"Hotter with 16 bit" depends on the source material. If you're not tickling full scale on peaks, hotter puts your average signal farther above the noise but HOW hot you should go depends on the ratio of peaks to average. And remember that peaks can be too short to register on even a peak reading meter and still produce audible distortion. And further complicating matters is many camera manufacturers don't tell us how their meters are calibrated in the first place: is it VU, PPM, or something else?

SD's advice is to set the gain staging so the mixer's clipping point equals the camera's clipping point. Both the 302 and the 442 can output +20dBu and that's the factory set output limiter as well. The input limiters kick in at +18dBU. Remembering that the mixers are calibrated so 0VU tone equals 0dBu output level, if we can assume (!) that full scale on the camera meters is legitimately the camera's clipping level, setting 0VU tone output from the mixer to read -20dBFS on the camera aligns the 2 clip points and the mixer's input limiters give a 2dB safety margin.
 
Steve House said:
SD's advice is to set the gain staging so the mixer's clipping point equals the camera's clipping point. Both the 302 and the 442 can output +20dBu and that's the factory set output limiter as well. The input limiters kick in at +18dBU. Remembering that the mixers are calibrated so 0VU tone equals 0dBu output level, if we can assume (!) that full scale on the camera meters is legitimately the camera's clipping level, setting 0VU tone output from the mixer to read -20dBFS on the camera aligns the 2 clip points and the mixer's input limiters give a 2dB safety margin.
Click to expand...

There certainly is clear logic in setting the +20 dBu level from mixer to 0 dBFS on camera, because then you only need to keep an eye on the mixer meters and the camera leves will be exactly the same. What is the point of setting level on camera higher, then having to do mental arithmetic all the time with mixer meters as not to overdrive the camera inputs? Why make life complicated and possibilities for failure greater without any benefit from this risk????

Factory setting on SD mixers for limiter is +18 dBu, I have it at +17 dBu. With 1:20 slope I get extra 60 dB headroom, totally impossible to blow it totally no matter what.

About this XH-A1 with SD302 thing: I made carefull real life tests and the analysis function in Adobe Audition showed the line in with pots full open to have over 90 Db dynamic range, the same ultra quiet mics straight to mic inputs was 10 dB worse (but still terrific). So I see absolutelly no reason to shy away from opening the pots all the way with line input setting. Some other camera has hiss, this is not that camera. Why do you regarad these pots as amplifiers, maybe they are attenuators, which work best when fully open? That looks and sound more like it to me. And it really does not matter what they do, they do it really fine.

------------------

addendum: Why rely on hearsay and past (bad) experiences with other cameras? Audio is easy to DIY test. All you need is a quiet room where you can do the video recordings and some instruments to make loud noises (I used a steel dog food bowl kindly lended by our german shepard, and a potato smasher). After capturing the video you just separate the audio into a wav file and analyze it with Audition or similar editor. If you have a CD test disk with pink noise and decent stereo system, you can even analyze frequency spectrum, at least the cut-off points.
 
Last edited:
Petrus said:
...

About this XH-A1 with SD302 thing: I made carefull real life tests and the analysis function in Adobe Audition showed the line in with pots full open to have over 90 Db dynamic range, the same ultra quiet mics straight to mic inputs was 10 dB worse (but still terrific). So I see absolutelly no reason to shy away from opening the pots all the way with line input setting. Some other camera has hiss, this is not that camera. Why do you regarad these pots as amplifiers, maybe they are attenuators, which work best when fully open? That looks and sound more like it to me. And it really does not matter what they do, they do it really fine.

...
Click to expand...

I certainly do wish that Canon was forthcoming on all the real specs and published a block schematic of the camera's circuits so we wouldn't be debating on partial knowledge. I've been assuming that the recording level pots are controlling the gain of the recording amplifiers (bacause that's the way most recording chains behave, AFAIK) rather than padding the inputs but without that info there's no way of knowing for sure and you may well be correct. The level controls on the Beachtek adapters are actually attenuators, for example, so it certainly happens that way. Whichever it is, my initial point was that one needs to be cautious about saying the controls should always be XX of the way open or whatnot because there are exceptions to the rules of thumb regarding input sensitivities and the optimal settings of the levels involved in gain staging, that particular Canon camera being an example of such exceptions. Most prosumer camera's use the consumer -10dBv line level, most professional camera's expect +4dBu professional line levels, but the XH-A1, for better or worse, is in a world of its own defining line level as +8dBu.

The SD 302 manual used to outline a procedure for aligning with full-scale tone. There was a button combination that resulted in the mixer sending tone at +20dBu, right at the output limiter threshold. You'd adjust the camera inputs until the camera meters were full scale and its clip indicators (if it had them) were just coming on. Now the camera and mixer clip points were aligned and you could rely on what you saw on the mixer's meters as indicators of what's happeing in the camera and the mixer's limiters would prevent driving the camera into clipping. The instructions seem to be missing from more recent versions of the manual though, don't know why as it really seems useful.

It would be an intersting experiment to put an external tone generator such as a Tone Plug one one of the mixer inputs and adjust the gain so it's putting out +20dBu on the meters, then adjusting the camera levels until its meter hits full scale and see where the controls end up.

I've been wondering, since the SD 302 and 442 allow you to set the 0VU point on the meters up to +8dBu output and also the output level of the tone up to +8dBu, if making those setup menu adjustments would make them easier to align to the Canon camera?
 
Last edited:
Steve House said:
I've been wondering, since the SD 302 and 442 allow you to set the 0VU point on the meters up to +8dBu output and also the output level of the tone up to +8dBu, if making those setup menu adjustments would make them easier to align to the Canon camera?
Click to expand...

Why, when the normal settings work perfectly?

I aligned my sd302 and XH-A1 with the sd302 instructions, first 0 dBu tone to -20 dBFS on camera, then double checked with +20 dBu full level tone, and lo and behold, it came to 0 dBFS on camera. To get this I had to open the pots all the way. If any of you care to do the same thing you are likely to find out exactly the same thing with these tools. As some of you voiced concern about the hiss & noise levels when pots fully open, I analyzed the available dynamic range. It was over 90dB. Good enough for me, case closed. What does it matter how the electronics work if they are this good.

If it looks like a cat, it acts like a cat, it smells like a cat and it feels like a cat, it is likely to be a cat (even if somebody mistakenly calls it a dog).
 
This conversation isn't just the Canon. There are a lot of cameras coming on line.

I wish I could get time with each one before I get on set. Tomorrow morning I am doing a walk and talk with Sony EX3. A glance at the online manual doesn't quite cut it. Maybe the trend is towards pro-sumer units meeting pro standards and I shouldn't get in a twist. How ever this noobie has been thrice burned and is looking for a little better technique.


As for getting +20 to camera, Oleg (if I got it right), uses the tone out on Lectro RX to double check scale on camera.
 
Last edited:
Baker Park said:
This conversation isn't just the Canon. There are a lot of cameras coming on line.

I wish I could get time with each one before I get on set. Tomorrow morning I am doing a walk and talk with Sony EX3. A glance at the online manual doesn't quite cut it. Maybe the trend is towards pro-sumer units meeting pro standards and I shouldn't get in a twist. How ever this noobie has been thrice burned and is looking for a little better technique.


As for getting +20 to camera, Oleg (if I got it right), uses the tone out on Lectro RX to double check scale on camera.
Click to expand...

I've got a Tone Plug, a 1kHz generator built into an XLRM connector, that works like a champ for that sort of testing. The little Behringer CT100 cable tester also has a tone generator, output selectable to roughly +4dBu, -10dBv, or -50 mic level that one could use. Both tools add value to one's kit far beyond their trivial cost.
 
You must log in or register to reply here.
Back
Top