volume level: soundcard vs monitors for mixing

[Noiz2]

Hi Noiz, Digital Cinema Initiative (DCI) is built entirely on open standards, audio to be specific, is just 24bit 48Khz uncompressed LPCM mono tracks, two mono tracks for 2.0, six mono tracks for 5.1, 8 tracks for 7.1

Picture wise, it's 12bit X'Y'Z colour space, encoded in JPEG2000 sequence and wrapped in MXF.

This is actually interesting. There has been no decent way to present a surround mix in a theatre with out $$$$$. But this may be a big break for lower budgeted films.

You blink and things change ;~)

My reservations about trying to do a mix for the big screen in a little room though are founded in actual experience. You can if your room is calibrated and reasonable "in spec" do it IF you can do some test screenings to check the mix AND generally have some experience so you are in the ball park to start off with, for Stereo. Mixing surround in a small un calibrated room is just throwing dice and hoping you win the lotto (the metaphor mixing was deliberate, the odds against it coming out anything like you wanted it to are huge).

If some self destructive urge makes you want to do it anyway at least get or borrow a Lake technologies "TeatrePhone" HSM6240. You can probably find one used (they don't make them anymore (Dolby bought them and orphaned it)). It will give you a decent surround image in headphones. Doesn't work with speakers. It was designed to be used for premixing films in small rooms so you didn't take up a stage.

They were pricey in the day (like $2,000) but I have seen them new (NOS) on eBay for $200.

A side benefit if you need to mix on headphones for other reasons is that even if your only feeding them stereo they reduce the ear fatigue one gets with headphones a LOT. And there are two jacks so you can have the director on cans also and you can record the "encoded" LtRt so you can sent it to someone (again it will only sound like a surround mix with headphones).

[ryanjf]

Sorry about that Scott - you're right. So sample rate would be the key player in this? Audio frame rate has nothing to do with the speed - it's the sample rate. If the picture changes frame rate, the audio has to be SRC'd to match, correct?

Interesting about the frame rate. I did see that wiki article on it. I've only been accustomed to 24 FPS DCPs so far and found some articles on GS that talks about our topic:

http://www.gearslutz.com/board/post-...-festival.html
http://www.gearslutz.com/board/post-...final-mix.html

One key post in the last link above:

although 25fps is listed as legal speed by DCI (together with 48, 50, 60 etc), there's a big probability local theatre isn't updated to accept these formats, so it's safer to keep your DCPs running at 24fps.

Ok - say the pic frame rate is 23.976 FPS and the audio is straight 48khz, and a DCP is an expected deliverable... Following the practices in the links above, the pic would have to be pulled up from 23.976 to 24 FPS and the audio, as a result, would have to be SRC'd as well (0.1% up), correct?

[John Willett]

Thanks for the reply, for Digital Cinema I mean those cinemas with DCI compliant servers, projectors and speakers.

I'm a bit confused now to be honest, for web and TV delievery, dialogue usually peaks around -12dBFS, can I do the same for my Digital Cinema Package? so I just need to have one master mixdown, or do I need to decrease the volume a few dBs specificly for theatrical release so dialogue peaks around -20dBFS? Are speakers in a standard digital cinema adjustable by projectionist? so they don't have a fixed volume to avoid films that are too loud?

The AES-EBU standard is to set 0VU as -18dBFS
The American standard is -20dBFS
The old standard for 16-bit recorders was -12dBFS.

This is for recording, of course, to allow a decent safety headroom to avoid clipping.

When you edit in post then the whole thing may be manipulated and compressed and the loudest peak would then be set close to 0dBFS.

There is certainly now a new loudness metering standard for broadcasting which says that delivery is to be -23LUFS (LU = Loudness Units).

[Noiz2]

Ok - say the pic frame rate is 23.976 FPS and the audio is straight 48khz, and a DCP is an expected deliverable... Following the practices in the links above, the pic would have to be pulled up from 23.976 to 24 FPS and the audio, as a result, would have to be SRC'd as well (0.1% up), correct?

It's much better to talk about "speed" and frame rate as two separate things.

23.976 and 29.97 are US video speeds. IF you change speeds you need to change speeds for everybody or you loose sync.

This used to be a problem in the US because we typically did sound post to video tape and so were working on a pulled down version of the film. So you would set up everything to run slightly slower while working and change that setting back to normal when you went to the mix stage and played back against picture.

An alternate route was to do that for dialog and not worry about it for FX and then just switch the opposite way on the stage so all the FX played back sightly slow but now in sync to film. It's a very small pitch shift so not really noticeable with FX but very with dialog.

With file based workflows and computer video playback we are not stuck trying to output to monitors that only can see video speed so there is no reason not to be working at the same frame rate as picture.

This is all about work flow. If the folks were aiming at big screen playback at 24 and they shot at 23.976 they made a mistake.

But on your end IF they have sent you 32.976 files to work with then when they assemble the film (at 23.976) all will sync. When they convert it up to 24fps everything including your mix should speed up.

As with all of this stuff you (they) NEED to plan for the finish BEFORE you start or big nightmares can result.

[Ed Kishel]

ok- thanks for the input so far, makes me realize how much I don't know about audio.
Re: my original question- below is what I have planned so far to calibrate my DAW...

1) play back the -20db 1k ref tone in audio editing app, keep it playing...
2) then open up the sound card's mixer and make sure it's meter is reading -20db on the channels running out to the monitors, adjust in the mixer if needed (or is it in the app?)
3) place a SPL meter at the mixing position and set it to c weight and slow response
4) go back to the the audio editing app and stop the 1k tone playback. Now start the full range -20db pink noise playback (but this plays back at approx -12db, why not -20 ?)
5) set windows volume to (?)
6) turn on left monitor only and adjust it's volume until the spl meter reads 82db
7) turn on right monitor only and adjust it's volume until the spl meter reads 82db

Does this make sense? 2, 4, and 5 I'm not too clear on. By the way- this is stereo editing only- no surround for me yet.

[Noiz2]

No need for tone just use Dolby Pink. The tone doesn't tell you anything in an acoustic environment, it was used to calibrate analog equipment.

Not sure about the Windows internal chain so? You do want to set the levels one speaker at a time. 82dB may be a bit hot for your room. I set mine at 80dB. It really depends on the room size and how far from the speakers you are. My room got quite a bit larger but my speaker position stayed the same. It's possible I may creep the level up a bit but? I know this setup and I know how it translates so probably it stays at 80dB.

In theory you can work at any level provided you know that level and know how it translates to the stage you are delivering to. The catch is that not all sounds change at the same rate. So if you mix at a lower level somethings that are going to play back on the big screen the same will seem quieter than others. You have to know your set up. Big stages are all calibrated the same so that a mixer can come in and theoretically be able to mix exactly the same as they would anyplace else. If it's your space you are the only one that needs to know how it translates.

[Steve House]

[QUOTE=Ed Kishel;1986152488]

go back to the the audio editing app and stop the 1k tone playback. Now start the full range -20db pink noise playback (but this plays back at approx -12db, why not -20 ?)

[/QUOTE]

The reason it's -12 and not -20 is due to the way energy is distributed in the waveform and how a meter reacts to it. A peak-reading meter indicates at -12 when the AVERAGE signal level is -20 when the signal-source is a real-world waveform such as speech or music; pink-noise mimics such a waveform. The same meter indicates -20 when the average signal level is -20 when it's fed a waveform of a pure sinewave tone. So if you have two signals, one a sinewave and the other pink noise, that have exactly the same AVERAGE level such as -20, and you measure them in turn on a peak-reading meter, the pink noise will read about 8dB higher on the meter's scale.

[Ed Kishel]

so you play back the pink noise as normal (with meters @ -12db, don't adjust volume to make it -20db). And then check the monitors with the SPL meter.

[Steve House]

so you play back the pink noise as normal (with meters @ -12db, don't adjust volume to make it -20db). And then check the monitors with the SPL meter.

Yep, that's the drill

[Noiz2]

Yep, that's the drill

Well except that if you are playing proper pink noise your meters will read -20.

If yourc DAW doesn't have a calibration mode then you need to set all of the levels at unity (in the software), which normally is "0".