How to avoid a cluttered mix?

M

Mitch.Kear

Well-known member
I'm looking for some advice on an upcoming job I'm doing for a pilot episode of a cartoon. In the past I have had trouble with a cluttered mix, not muddy in the low end just lacking the natural real life feeling in the mix with Foley, sfx, dialogue, ect. I pan, adjust levels, eq/comp, use reverb and still...

Hoping for some tips from you post guys. If you came across this issue in the early years learning, or even still today. As well as steps you take to get over this issue.

I mix in protools and my audio is always clean. I've only been working at mixing for about three years, but as I've said before this is where I want to learn and grow in and figure cartoon post may not have as many issues and problems, being a great place to get experience.

Thanks,

Mitch
 
I've mixed a few short films and a feature. I also have 20 years as an audio engineer for music stuff - part time. Get good monitors that rem=present the bass well. I LOVE my LSR 4328s!

Just remember the rule - DON'T MOLEST THE DIALOGUE.

Make the dialogue sound great, and the rest falls into place. I like to put some compression on the vox to bump up the quieter stuff, and cut down the peaks. If there is a really quiet talker, I edit to bring up their volume, rather than relying on compression. Compression should be used sparingly to gel the peaks and valleys. As for other stuff, if you think its cutting into the dialogue, try cutting the level, but if that kills the vibe, try making a hole for the dialog with EQ - scoop out a hole in the freq range of the human voice. Keep the voice right in the center, and spread Foley wide - leaving less clutter in the center. Mix the music/effects quieter, because at the end, when you master the final track it pushes that stuff up. I like to use an ultramaximiser as the last thing in the chain by either making a print of the mix, then mastering that separately, or mixing everything to the point of being done, or close to done, then slapping an ultramaximiser on the master bus and maybe a little EQ. OR there is also Ozone4 from iZotope, that has a lot of mastering effects, but be careful not to kill your mix going overboard. Movies like to have a wider dynamic range than say a TV ad, or a song. You can try using a multi-band compressor but you should practice with that. You can really over-squash a mix with that, or you can use it to preserve the voice.

I'm not one for trying to make the audience "experience" more of the action, by jacking up the volume at those times. To an extent it's OK, but it's just a bad trip for me as a viewer, and I grab the remote to turn down, and I find myself constantly riding the remote volume while watching some movies with action.

Mainly just don't molest the dialogue. Get that straight and consistent first, then bring in the music, effects, Foley. Spread out what you can to make room in the center, and try to preserve the dynamic range without killing the ears of the audience when a fight breaks out.
 
Thanks Chad.

I actually have an "edit bay" set up with multiple monitors. I learned with music too and know the tools. Since its all v/o the dialogue I record will be perfect. I learned not to molest the dialogue track the hard way and learned subtlety is key in that aspect. The issue is the final mix once I have everything in place. All my tracks are clean and lined up.. Now trying to fit it properly in the "3D" audio field.

Curious though, aren't you afraid using a maximizer will kill your dynamics? I also didn't realize post guys eq'd the dialogue to create room for Foley/sfx.. Thought it was only music, and I thought wrong haha.

Thanks for the advice.
 
Dynamic range for TV should be around? 10 - 15db? Or are my suspicions correct thinking that in the past I hadn't left enough room resulting in this kind of cluttered mix.
 
Mitch I don't EQ the dialogue to make room for Foley! I EQ the Foley to make room the the dialog. And as for maximizing, yes you can kill your dynamics if it's misused, but I don't use it as I would on a song - attempting to get a high constant level. I just use it to boost the volume of everything a bit, cut some peaks, and generally gel the whole mix. I keep it mellow. If I play the DVD next to another Hollywood DVD and I feel the overall audio is too quiet, I can boost it a bit without risking any clips. And personally I put the limiting ceiling at -1db. For some reason, stuff louder than that can sometimes clip in FCP!?? and that 1db won't be missed really.

For VO stuff over music, I just treat it like a song. I put a channel strip on the VO, EQ, and Compression. Squish to taste, then hand edit the music to raise and fall where the voice isn't and is. Then I slap izotope Ozone 4 over the whole thing, with a preset I made that consists of EQ, Maximiser, Multi-band compression - (great for taking care of a tubby mix) and maybe a little tube saturation to give a bit of sizzle. I mix VO stuff more like a song. I use Cubase, and live it. I like that I can cut up a clip, then every piece of the clip can be individually volume adjusted up24db, or down 60db. That way I can level out the track without committing to writing a level adjustment. If I want to raise the whole track volume, you can't easily do it once you've written fader rides. With the clip adjustments I can even out a track, then use the volume fader to lay it over everything else with more ease.
 
Chadfish said:
Mitch I don't EQ the dialogue to make room for Foley! I EQ the Foley to make room the the dialog.
Click to expand...
I said it backward but you typically would still kick up the Vox at the same freq a tad. Also, that's odd about having to set at -1db I set at -.1 in ozone and no issue with clipping in fcp. I find I try to apply music muinh to much to post. Best tip I've gotten was automating every aspect of the sound. I think I will actually have to wait and see if I have any issues then report back here for advice.
 
I used to master with a peak limit od -0.3, which is my norm for music, but oddly I was getting some clipping in FCP, so at -1 I feel I'm safe. I think it was after I exported the file, and brought it back in for something else, then it would. Not enough to hear, but I would get red indicators. It seemed that the render file process added .1 - .3 db of volume.

The TV stations I make stuff for turn it way down anyway, so it's not too important.
 
Last edited:
I also would use EQ in ambience, foley to make "space" for dialogues. EQ would be around... 10kHz, with quite wide Q value (the "range" of the equalizer at that point).

Sometimes low cut in ambiences also helps. And don't worry to play with it and exaggerate to hear the effect.
 
Chadfish said:
I used to master with a peak limit od -0.3, which is my norm for music, but oddly I was getting some clipping in FCP, so at -1 I feel I'm safe. I think it was after I exported the file, and brought it back in for something else, then it would. Not enough to hear, but I would get red indicators. It seemed that the render file process added .1 - .3 db of volume.

The TV stations I make stuff for turn it way down anyway, so it's not too important.
Click to expand...

I'm surprised you can get away with submitting deliverables that are that hot. Most network qc technical specifications still call for "never exceed" levels of -10dBFS or -12dBFS for SD deliverables, -3dBFS for HD. And then there's this whole new dialnorm specification now legislatively mandated where average dialog levels must sit at -24dBFS +/- 2dB LeqA averaged over the entire segment, with a peak dialog 'never exceed' limit of -17dBFS, and other program element's levels mixed with that as the reference point. (PBS Redbook TOS for the example)

Older analog SD broadcast specs typically would read like this one for CTV/Discovery Channel Canada "For (SD) broadcast “full mix” tracks, transient audio peaks must not exceed +8 dBm above reference tone when measured on an audio meter using the “True Peak” ballistic set (0 ms rise, 200 ms fall). For secondary audio tracks, such as music and effects tracks, and audio elements, peak levels must not rise above +10 dBm. When mastering to a digital format and/or using an Absolute Scale or Peak meter, where “0” is at the top of the scale and (0VU) reference tone is at -20 dBfs, audio for the “full mix” tracks should peak at no more than -12 dBfs."

For HD materials, they say "Transmission limiters clip at +8 dB (over 0VU). For broadcast stereo tracks, transient audio peaks must not exceed +8 dB above reference tone when measured on an audio meter using the "True-peak" ballistic set (0 ms rise, 200 ms fall). ... For 5.1 surround mixes, audio peaks may rise as high as +21 dBm (-3 dBfs). When mastering to a digital format and/or using an Absolute Scale or Peak meter, where "0" is at the top of the scale and reference tone is at -20 dBfs, broadcast stereo tracks should peak at no more than -12 dBfs... Maximum permissible audio peaks in a 5.1 or 5.0 soundtrack shall be -3dBFS (+21dBm). Although the Max dynamic range (max. peaks) for 5.1 channel mixes is considerably higher than for Stereo-only LPCM mixes, it is understood that many 5.1 mixes will have a dynamics structure which more closely resembles a -12dBFS stereo mix in order to facilitate the simple creation of an Lt/Rt fold-down mix. ... Regardless of the gain structure of a 5.x channel surround mix, it is crucial that the supplied DIALNORM value accurately reflects the Leq (A) of program dialogue."


Given those numbers, it's no wonder your submissions get backed down at the broadcaster's studio. Local stations tend to be less rigid than network QC but I'd still expect them to be happiest with on-spec material.
 
Last edited:
Doc Bernard said:
Don't know if you have this, but I have (and love) this book. And I highly recommend it.
Click to expand...
I do not own this book but I seems like a good read. Thanks Doc!Steve, I admit I'm can't quite figure out what these levels are for. Care to expand, sum up or simplify? Thank you.
 
Mitch.Kear said:
The "network qc technical specs". First paragraph. You can't exceed -10db.. On the overall mix?
Click to expand...

If reference tone in the program slate is set to -20dBFS on your editing workstation's meter and you play through the entire program with the mix complete - dialog, music, sound FX, etc all in place - at no point in the program will the meter ever jump up above -10 dBFS, even momentarily. In the station, the broadcast engineer will set the playback level so your recorded tone reads 0VU on the meters on his playback equipment, confident that no peak in the program will ever go over the +12dB or so of headroom his audio chain has available before it distorts. In addition, the 'dialnorm' numbers means that when you mute everything except the dialog and use a long-term averaging meter (not the peak reading meter in your NLE) that has specific averaging characteriistics, the average level of the dialog portion only must be -22dBFS with a tolerance of 2dB on either side of that and some specs I've seen include a provision that at no point in the show can momentary peaks in the dialog exceed -17dBFS.
 
Last edited:
Mitch.Kear said:
Wow, I never knew that, I just mixed by ear. Thank you this is something I read more into.
Click to expand...
Mixing by ear is fine ... you have dialog and music together and you would use your ear to set the level of the music with respect to the dialog. No problem doing that, but you need to take into account the limits imposed by the standards as well. You might want that sudden gunshot to be super loud compared to the whispered dialog preceeding it but if it's so loud it drives the broadcast transmitter into clipping the audience won't hear it as a gunshot at all, they'lll hear the hash of distortion and overload.
 
Film is mixed by ear in a calibrated room. That last part is critical. Video has to hit specs that are handed down to you. I was about to express the same amazement as Steve since anything I submitted at those levels would have been spit back by QA.

As to peaks in general. Unless you are using an overshoot meter (if your not sure the answer is no) then anything over -3 may be clipping on the analog side of the D/A converter. The converter interprets an analog wave based on data points. Because of that two data points that are close to zero create a wave that crests over zero. There is a a nice paper on this by Bob Katz.

So despite the loudness wars nobody should be hitting any higher than -3 or they will be distorting their sound.


As to your original post.

You actually seem to be having two problems. One that you feel the mix is cluttered and the second that it doesn't sound "real".

They probably interact but it's really two separate problems.

Generally not feeling real I would tend to point to not enough layers. Your mind figures everything out and is bored so it starts picking things apart. When you have enough depth to the sound scape it starts to feel as loaded with info as a real world situation would be and everything comes to life.

It's kind of a classic low budget problem because it takes time to create a full world. On a film with a budget, and a decent supervisor, you start covering everything that is seen that moves or needs to make a sound. Then add everything you don't see but know is there and might make a sound. Then you start adding the things that are just there for texture. After that you can add more little bits if you have the time.

So two folks having coffee at a cafe.

You have all the obvious stuff you see. Then you have whoever else is in the room. They talk and drink coffee also. The machines, waiters taking orders, cash registers all the stuff you know has to be going on. Then there is out side. Maybe busses go by? a truck, birds, pedestrians,etc.

But then maybe some folks live over the cafe? Maybe there is a plane, siren, dog, folks arguing in the street. Just lots of life.

If your doing a David Lynch kind of film some of these things may be very stylized or very POV. Chairs scrape, insects walk along????

Anyway when you have a tone of "reality' going on it will seem much more real. It's not about reverb it's about complexity.


Uncluttering is probably about being more selective about all that reality. The birds are great but they need to go away after a bit, the argument needs to keep moving down the street etc.

OR your bunching up in the same frequency range. Clear out some sonic space for things, especially around the dialog. If the birds and something else are in the same space move one. A lot of sounds (crickets come to mind) only need to be in your face when they are being established. After that they can just about disappear and you will still think you hear them. In fact if you don't bring them way down your audience will hunt you down and kill you...

You probably hardly ever need to use reverb except as an effect. Generally try everything dry first ,then add with a very light hand.

That is extra true if your not mixing on a film dub stage and your film is going to get played on a big screen. Reverb will sound many times heavier in a big room.
 
I am aware of the various standards for audio levels depending on the station. Locally each station has a different standard, so they just set the levels - they are used to doing it. So I mix for web / dvd levels which is all the way up.
 
You must log in or register to reply here.
Back
Top