peaking level for film and Dolby E

[hectorC]

Hello,

This is my first post to the forum. I've been reading and searching and I can't find a definitive answer to my question.

I'm preparing the audio for a short film that will be presented at a festival (this is in Canada). The festival is asking for Dolby E on HDCAM tape. I went to a studio to have the Dolby E encoded and the engineer there told me that my tracks where peaking to high (at around -2 dbFS) and he recommended to not go over -8 dbFS. I've been reading here that for film there shouldn't be a problem as long as it's not clipping so that's why I kept it like that. Back in my studio I added a limiter to my mix, set at -8 dbFS, but now I'm getting a waveform that seems a bit to square in the loudest parts (the film has no dialog and is just music/sound effects). In one of those loud section (the climax of the music, which lasts around a minute) I'm hovering constantly between -15 to -12 dbRMS. The medium-quiet parts hover around -25 to -20 dbRMS.

My question is, does all this make sense? Will my levels translate properly to a movie theatre? Unfortunately I have a small studio and I had to calibrate down to 67 dbSPL for -20 dbFS pink noise since anything closer to the standard 85 dbSPL was too hot. I came to that number after I had to use a voice sample recording (one averaging at -24LKFS) I found here in the forum, for doing a perceptual calibration.

Thanks!

[kk@jamsync.com]

Is there an overwhelming reason why you can't just pull down the volume? Dolby E and Dolby Digital will both fail (distort and even mute) if you try to combine several tracks that are too hot. Compressing the tracks makes it harder for the algorithm to know what is signal and what is stuff the lossy compression can throw away. If everything is tending towards being a square wave, how can it know?

You need to be cal'd no less than 75 in a tiny room. The reason that you thought 85 was too hot is that you were assuming everything should be the same level as the pink noise. It shouldn't be! I cal with earplugs because I can't stand 85 dB SPL of pink. If you cal too low, you will run your mix levels too hot!

This is what you are experiencing.

[danijel]

Originally Posted by kk@jamsync.com

You need to be cal'd no less than 75 in a tiny room.

Hector, I think you're probably cal'd to around 75 if that ATSC -24dBFS LKFS test clip sounds good when you play it back straight. How did you measure to come up with 67? If you are not sure about pink noise and SPL measurement, I suggest you stick to the 'perceptual' calibration.

Originally Posted by hectorC

Will my levels translate properly to a movie theatre?

No one can answer that. Unless a Dolby person has calibrated the playback before the festival, you're going to be either too hot or too soft, it never happens to be 'just right'
If this is a festival with low budget documentaries etc that are mixed TV-style, than you will be all right with your current level (according to the -24 clip). If this is primarily a feature film festival, you'll be too hot by at least 5dB.
As for the peaks, I don't know because I never did Dolby E, but this has always been a problem for people who transfer to tapes - they know the TV mixes don't go past -x dB, so they either call me to check if I really want to peak past that, and I even had one lower the whole mix by 10dB, without asking.

[hectorC]

Thanks all for your replies. I didn't just bring the mix down since it was already sounding too quiet when I listened to it at the studio where I went to do the Dolby E. The guy there suggested the same when he saw the peaks but he later heard the whole mix and told me it was too quiet. I think my problem is that I usually do audio for different purposes where the dynamic range can be wider so my soft sections are maybe too soft for theatre standards. What is a safe dynamic range for theatre? What would be the dbRMS suggested for the quiet moments (e. g. an orchestra pianissimo with no dialog)?

Originally Posted by kk@jamsync.com

If everything is tending towards being a square wave, how can it know?

You need to be cal'd no less than 75 in a tiny room. The reason that you thought 85 was too hot is that you were assuming everything should be the same level as the pink noise. It shouldn't be! I cal with earplugs because I can't stand 85 dB SPL of pink. If you cal too low, you will run your mix levels too hot!
This is what you are experiencing.

When I did the calibration I didn't expect that everything should be as hot as the -20 dbFS pink (I know that's not the purpose of calibration, not sure how did you get that). I did the calibration then played my mix and it was then that I found everything too hot so I remixed until it sounded ok but then the Dolby E studio guy told me it was too quiet (as I just explained). Also I said it was only in the climax (1 minute long section) where it started to look square, not everything.

I played that sample voice recording (the -24dBFS LKFS test clip) to perceptually adjust the level of my system so the voice sounded comfortable and natural. Then keeping that level I played the pink again and measured with the SPL meter and found it was 67 dbSPL. Is this procedure wrong?

Cheers.

[kk@jamsync.com]

Originally Posted by hectorC

I played that sample voice recording (the -24dBFS LKFS test clip) to perceptually adjust the level of my system so the voice sounded comfortable and natural. Then keeping that level I played the pink again and measured with the SPL meter and found it was 67 dbSPL. Is this procedure wrong?

Cheers.

Yes, I think that is incorrect. You mix to the cal...not cal to what you hear. If you are cal'd too low, your mixes will run hot. Just like, if you cal a small room at 85 dB SPL, you'll tend to run your mixes light, partially due to compression by the room itself. If you're playing -20 dBFS Pink and getting 67 dBSPL, you're either cal'd too light or something is wrong with your amplification system. The fact that your mixes are too hot tells me the 67 dB SPL figure is probably correct and you should cal with pink noise at 75 or 79; then mix.

I'm not going to argue this point, because the whole point of calibration is that dub stages and small rooms have different but related specifications for calibration. You can fairly accurately predict what a calibrated room in the real world will sound like. It isn't perfect, but it's a lot better than everyone deciding what sounds good in her or his own studio with a random sample and then hoping for the best. It also is not rocket science. Anyone who can use a multimeter, an SPL meter and an omni microphone can do it.

[kk@jamsync.com]

Originally Posted by danijel

I even had one lower the whole mix by 10dB, without asking.

One of the reasons I just "do it" without asking these days is that I had a project delivered to me with all the levels completely slammed. Thinking the artist's engineer had sent the archival files by mistake, I called the engineer. He replied that "those are the masters from [insert famous Nashville mastering engineer's name here] and you obviously are just trying to make more money." So, I shut up, brought down the levels and the DVD went out and was a big country hit for the artist. It wouldn't have been if I hadn't brought down the levels. The artist's engineer was the typical good old boy who "doesn't believe in digital; everything should be analog".

After you do enough of these things, you just do what you know will be right and stop asking people because that particular pursuit leads nowhere.

Let me re-iterate, Dolby E and Dolby Digital will both distort or mute if you try to encode multichannel that is too hot. It ain't pretty and yes, there is an art to mixing and compressing so that the peak-to-average levels work without making the whole thing a mess...but that's really the mixer's job and then the mastering engineer's job and finally (when it's really too late) the compressionist's job. Careful compression and limiting of mix elements with an eye on peaks and averages on the 5.1 bus is really important (assuming you have a properly cal'd room...and around we go.)

[hectorC]

Originally Posted by kk@jamsync.com

Yes, I think that is incorrect. You mix to the cal...not cal to what you hear.

Thanks K. K. Proffitt for your comments, I honestly appreciate them, but I think you might not be fully reading or following my posts (I say it in a respectful way). I never said I calibrate to what I hear. Danijel was asking how I came to that 67 dB figure and I was just describing how I did it: I first calibrated to 85dB using pink noise (a proper one for calibration) then found that my mixes where resulting too quiet when played at another studio (where they'll do the Dolby E encoding) and also I found that the voice -24dBFS LKFS test clip (found here in the forums) sounded too loud so I turned my system down until that test clip sounded comfortable. I then measured the pink noise again to find out what the difference was and I measured 67 dbSPL. Is my SPL meter telling me lies maybe (it's a cheap Galaxy Check Mate SPL meter)?

I've read lots of information about how to calibrate, I was just looking for feedback from people with more experience than me regarding appropriate dBFS peaks and RMS levels for audio done for the described context (HDCAM, Dolby E, film festival theatre presentation), since it looks like my setup is not reliable enough for just mixing by ear (due to the calibration discrepancies mentioned). We also don't have the budget to hire someone else to do the mastering so we have to work with what we have.

What are those peaks and averages that I should be looking at in the 5.1 bus? What numbers?

cheers

[kk@jamsync.com]

Originally Posted by hectorC

I found that the voice -24dBFS LKFS test clip (found here in the forums) sounded too loud so I turned my system down until that test clip sounded comfortable. I then measured the pink noise again to find out what the difference was and I measured 67 dbSPL. Is my SPL meter telling me lies maybe (it's a cheap Galaxy Check Mate SPL meter)?

I'm reading your posts. I understand exactly what you're saying. If -20 dBFS pink measures 67 dB SPL, then that's your calibration level. This is the reason your mixes are too hot and the person you mentioned in the initial post requested that you pull down your levels. Yes, you shouldn't go above certain peaks, but more than that the *average* levels have to make sense as well. Slapping a limiter on tracks that have average levels that are too hot will make problems for encoding. If the mix sounded too quiet after pulling down levels, then your average levels are wrong, but maybe they're only wrong in certain frequency areas, say around 2k to 4k. Without doing a frequency sweep of your room and calculating room modes, you won't know...but that's something you do *before* you cal. You really can't break the laws of physics; a small room will have problems with room modes that will have to be straightened out with traps and diffusion (and really cannot be perfected the way a large room can). When you said you used a vocal sample, lights went off in my head...your room apparently supports that range so much that you turned your master fader down, but what about the other frequencies, and what about playback in a room that registers fairly flat in the audio spectrum? There are a lot of things to consider. Without having a fairly good playback space, getting the mix right even with calibration at the "right" level may not work.

You have complete free will to mix and cal any way you want...it's a big universe. However, people who tend to do this for a living pretty much stick to large and small room calibration levels. If you want to get a handle on theatrical dynamic range, perhaps you should check some of the threads about people's preferences for film sound around here, buy the DVDs and listen to them in your room. If they sound wrong in your room, then your stuff will sound wrong in theaters and on TV. If you don't know how to test for room modes, etc., it's a good idea to hire someone who knows how to "tune" a room, preferably someone who has worked in surround-capable rooms.

[hectorC]

Originally Posted by kk@jamsync.com

I'm reading your posts. I understand exactly what you're saying. If -20 dBFS pink measures 67 dB SPL, then that's your calibration level. This is the reason your mixes are too hot and the person you mentioned in the initial post requested that you pull down your levels.

I'll put it in this way, maybe it's a bit more clear (chronologically ordered events):

1. Calibrate system at 85dBSPL using pink noise.
2. Do the mastering by ear.
3. Play at second studio for Dolby E encoding and the maximum peak is found to be too high (at -2 dBFS) but the general level is perceived as generally low (low RMS). A max peak of -8 dbFS is suggested.
4. Bring down the monitoring level (which seems to be the reason why the final master sounded too quiet).
5. Remaster by following the meters. I apply limiter at -8dbFS, bring up the general level and also bring up a bit more the quiet sections which now hover around -25 to -20 dbRMS. Loudest sections (three different one minute sections out of the total 9:00 minute length of the short film) hovering constantly between -15 to -12 dbRMS.
6. Play the -24LKFS test file and adjust levels so it sounds comfortable.
7. Measure pink noise and get a reading of 67dbSPL
8. Replay mix and seems to sound adequate.
9. Get all confused and worried about if this will sound OK at the theatre.
10. Post at Gearslutz looking for help!

Cheers,

Hector

[kk@jamsync.com]

Originally Posted by hectorC

I'll put it in this way, maybe it's a bit more clear (chronologically ordered events):

1. Calibrate system at 85dBSPL using pink noise.
2. Do the mastering by ear.
3. Play at second studio for Dolby E encoding and the maximum peak is found to be too high (at -2 dBFS) but the general level is perceived as generally low (low RMS). A max peak of -8 dbFS is suggested.
4. Bring down the monitoring level (which seems to be the reason why the final master sounded too quiet).
5. Remaster by following the meters. I apply limiter at -8dbFS, bring up the general level and also bring up a bit more the quiet sections which now hover around -25 to -20 dbRMS. Loudest sections (three different one minute sections out of the total 9:00 minute length of the short film) hovering constantly between -15 to -12 dbRMS.
6. Play the -24LKFS test file and adjust levels so it sounds comfortable.
7. Measure pink noise and get a reading of 67dbSPL
8. Replay mix and seems to sound adequate.
9. Get all confused and worried about if this will sound OK at the theatre.
10. Post at Gearslutz looking for help!

Cheers,

Hector

I'm with you until #6 because if you tweaked the mix at 85 dB SPL and it was way too loud, you were "mixing by eye". I'm not sure why you used the -24LKFS test file to adjust the level rather than your own mix. You're psychoacoustically influencing your decision by playing the test file.

Why don't you take a day off, sleep, recal at 79 or 75 and listen to your mix?
They won't be playing the test file at the theater. That said, there are some films that require higher average levels than others (action films come to mind). I think you need to get away from it for awhile and come back for a re-listen when you have cal'd at a lower level (but not 67....

[danijel]

Originally Posted by hectorC

6. Play the -24LKFS test file and adjust levels so it sounds comfortable.
7. Measure pink noise and get a reading of 67dbSPL

I'm betting my favorite mittens that you're either:
- not playing the -24LKFS file straight to your monitors (it should be on a mono channel, routed directly to your hardware output, with master fader at 0, if any),
OR
- your SPL meter is not working.

Check your PM, I'll send you some clips from movies, so you can compare

[hectorC]

Originally Posted by danijel

Check your PM, I'll send you some clips from movies, so you can compare

Great, thanks to you both for all your advice. I did a recalibration at 75 dB and tested the clips and it sounded fine to me. I'm redoing the levels now. Those clips gave me a good idea about dynamic range too. I could have used my own DVDs but I wasn't sure if using DVD standard would work for a movie theatre, but it seems it should be fine.

Will see how it goes the day the movie is played!

Cheers,

Hector