audio mix

[Shane Ross]

That I don’t know. I notice that MOST of the DVDs I rent do that…music ALWAYS overpowering the dialogue. I have always wondered about that? Is that because they are set up for Dolby 5.1 systems, and I am just playing it thru my TV Speakers?

Shane

“There’s no need to fear, UNDERDOG is here!”

Shane Ross
Alokut Productions
http://www.lfhd.net

[Gary Hughes]

It’s not a problem with the DVD or the playback system. It would be the same if your program were played back from betacam (or VHS or whatever), into your TV and into other speakers. It’s simply the fact that no two brands of speakers, or models in the same brand, are identical. Some speakers accentuate vocals, which makes your dialogue cut through the music and the music isn’t too loud. Other’s accentuate the music, which will make the music seem to overpower the voice a little more. Others are very flat or neutral and neither is accentuated.

Your goal as a sound mixer (and you just thought you were an editor), is to determine what your demographic will be playing back on and mix for that. Or, you can mix so that the dialogue is understandable on the speakers that normally make your music too loud, then worst case scenario is that others will hear the music as too quiet, but at least everyone will be able to understand the dialogue.

Personally, I mix the music to dialogue relationship on Sony MDR-7506 headphones or Bose Quiet Comfort 2 headphones, and I mix it so that the music is almost overpowering the dialogue. I’ve found that those particular headphones accentuate the music enough that it makes my mixes “safe” for most listening environments. I also “master” all of the dialogue to reduce the dynamic range, so it will cut through the music. I also always listen on computer speakers, TV speakers, and home stereo speakers. I also take into consideration the type of program and how much the music can hurt or help the mood. Then, basically, I find the happy medium and go with it. Just like editing, it’s an art.

Hope this helps,
Gary

[Don Greening]

A true “home theatre” system will have a centre channel, stereo left and right front, stereo left and right rear and a sub. With my system you can adjust/balance the rear, front and centre channels independantly, depending on how far the viewer sits from all these different speaker sources. If you’re close to the television then you may need more volume from the rear channels, etc. The short answer is that you can’t really customize your 5.1 sound design to suit every home theatre system. It’s up to the owner of each system to tailor the channel volume to suit their needs. It’s also been my experience that most people don’t take the time to properly configure their home systems to suit themselves, but end up just “living with it” as it were.

I would suggest mixing your tracks to be as equal in volume as possible. Of course, if you know for sure that your project will be viewed mostly on regular stereo speaker televisions then you may want to increase the centre/dialogue track to be a bit higher in volume than the rest of the tracks, so it jumps out ahead of everything else.

To be honest, any answer you get to this question will be very subjective because each person’s ears are as unique as the person using those ears. On my system the centre channel is preset substantially higher than the rest so the dialoque is crystal……and then everything else is a bonus……e.g. the house shakes when the thunder hits 🙂

– Don

[Gary Hughes]

How true you are, Don! And you’re absolutely right about the fact that it’s not possible to fit everyone’s situation, but the truth is, it’s not desirable for you to try. You should try to either fit your audio to a perfect system and leave the rest up to the homeowner, fit you audio to a THX certified system and leave the rest up to the homeowner, fit it to standard TV speakers and hope that everyone’s TV will make you shine, or you should try to find the happy medium with the dialogue volume, or best yet, hire an audio for video pro to mix it for you and trust them.

I always seem leave some important info out, for some reason. My reasoning is subjective, and it is for my type of project, which is 2 channel audio with dialogue as the focus (usually either marketing or instruction or informational). I would not even apply it to a feature film project, or a music video project, or even a surround sound project for that matter. For any “entertainment” projects, I would hire someone who mixes and masters audio full-time to do it for me. That used to be me, but now, I’m a video editor and have lost part of my touch, and certainly not kept up with the equipment of today when it comes to audio. But even if budget didn’t allow me to hire someone else, I still wouldn’t take my own advice in those types of projects. So, not only is a good mix subjective, and good EQ subjective, but also the projects themselves have variables.

IF, you are working on a dialogue based project with background music, my method will work. There are many other right answers too. If you’re not working on that type of project, just take my advice as a lesson for future experience and move on.

Good luck.
Gary

[Glenn Woiler]

garyh357 said: “I also “master” all of the dialogue to reduce the dynamic range, so it will cut through the music.”

Please explain to me what you mean here. Please… 🙂 I am a one man operation and can’t do everything top notch…. but I try. If I am hearing you right… you are mixing and cutting freaquencies a bit, –which does not let the voice get muddied in the music? Please take a few sentences to teach me this. Thanks… most of what I do is narrated and some has music behind it.

thanks,
glenn

[Gary Hughes]

Glenn,

I do not cut the frequencies unless that is needed for tonal quality. Actually, boosting just the right frequency a little can help the dialogue to cut through the music, but be careful if you don’t know what you’re doing. Changing the frequency has nothing to do with dynamic range.

Dynamic range is how much or little difference there is between the peak volume and the average volume of the audio, or in this case, dialogue. Here’s an extreme example of dynamic range. If you are at a shooting range and you record a guy whispering, right before he shoots his glock, you’ll have a high dynamic range recorded. If you are shooting a sad scene in a movie and someone goes from a whisper to a quiet talk, you have a low dynamic range.

There’s a lot of options and methods for how to mix audio, but with the type of video you are doing, you’re better off with a low dynamic range in the voice. Voice naturally has a lot of dynamic range, and if you don’t treat it, and if you mix your music kind of loud, at some point, the voice is going to get quieter and be drowned out by the music. So by treating it, it’ll stay high and above the music the whole time. Treating it is, however, not an exact science. It’s more of an art. To treat it, you’ll use some sort of compressor and/or limiter. They go by many names. Mastering limiter, compressor/limiter, dynamics compressor, and then there are actual model names. Elemental audio calls theirs Neodymium and Finis. All too confusing for many, but the bottom line is they are simply compressors or limiters or a combination of both. Not to make light of one brand over the other. Many are just not that good and others are really nice, some are easy to understand and others are really hard to figure out.

In FCP, (or at least in mine), in the “Effects” tab in the browser, in the “Audio Filters” folder, and in the “Apple” folder, there are a couple of filters there. AUDynamicsProcessor and AUPeakLimiter. And in the the same location, but in the “Final Cut Pro” folder, there’s Compressor/Limiter. If you don’t have Sound Track Pro or Peak, you can use those, but I don’t, so I don’t know what the best settings are. I use Sound Track Pro and I just bought Peak Pro 5, so I’m going to start using it. Telling you what settings to use is pointless because you have to tweak it for the content you’re using. I will give you some pointers to get you started though. I’ll start with “what not to do” because I think it’s almost more important than “what TO do”.

What to avoid: Like most things in the creative field, less is more. If you use too much compression or too much limiting, you’ll start to hear it “pump”. It’ll sound like the volume slide is being ridden by a paranoid person trying to keep the audio level perfectly the same. It could even start to sound distorted. The better the tool, the less of this you’ll hear. Some compressor/limiter plugins cost upwards of $1,000.00 because of how much work has been put into making them sound like they aren’t doing anything. That may sound pointless, but actually that is the whole point, to try to squeeze the volume peaks and averages closer together without making it sound unnatural.

Gotchas: Noisy background or lots of ambient reverb in your raw audio can really be a problem. If you compress a clip with background noise, you’re actually making the background noise louder while turning the voice down. You better use some sort of noise reduction or elemination tool, like the one that comes with Sound Track Pro or Sound Soap, BEFORE you compress that clip. Get rid of the noise, then you have a clean clip to compress. As for reverb in your clips, there are compressor limiters that do a good job of compressing the louder audio while reducing the quieter audio. Those will work for reverb filled clips. Usually it’s just the way you set the one you already have. One really really nice one is the Elemental Audio Neodymium compressor. As someone with an audio background, and completely understand the way compressors work, I find it very refreshing and visual and easy to understand. I’m not sure that same impression would come across to a complete compression newbie though. They have a demo that periodically mutes the clip so you can’t actually get work done with it. https://www.elementalaudio.com/products/neodynium/index.html This is why I like theirs, it is set up to be up to 4 compressors in one, and each can be set for compression or expansion, which is the noise reduction part. You can set one zone for compressing everything from the quietest “S”s in your audio, up to the loudest peaks, and set another zone from just below the “S”s down to reduce the volume. This works great for compressing noisy audio, audio with reverb, or audio with bleed-over from other people on stage or off-set, including nearby UPS trucks. Their Finis limiter is nice too. It’s a “brick-wall” limiter, so that if you set it to limit to -6dB, absolutely nothing will go over -6dB. The two plugins work in Sound Track Pro or Peak and are $159 each or $209 for the set. Don’t rush out to get them if you haven’t played with what you already have though.

Your Goal: To squeeze down the peaks so that they are closer to the average sound level which in turn allows you to turn back up your audio without going above your original peak. Everything compressed will sound louder and more even. When you play your timeline, with only the dialogue track audible (solo it or mute the others), you’ll notice that you may have peaks much higher than where the average level is staying on the meter. For instance, let’s say your peaks are averaging -6dB while occasionally going almost up to 0dB while the bulk of your audio is down around -16 or -20. After compressing and limiting, that same audio could be peaking no higher than -6 while the bulk of it is averaging -9 to -12. Those transient (really quick) peaks that were above -6 weren’t really noticable, but would be a problem in a broadcast situation. However, -9 or -12 is going to sound considerably louder than -16 or -20, therefore sounding more even in volume.

The Method: Common to all compressors are a few terms to become familiar with. Some of them call these something else, and some do them automatically so you don’t have to (or get to) mess with them, so don’t panic if you don’t see these exact terms in your compressor.

• Threshold sets the level at which the compressor starts working. If the threshold is set to -6 and your audio peaks at -3, it’s barely going to do anything. It’ll only compress the top 3dB in that case.
• Ratio sets how hard the compressor will work. A low ratio is barely going to compress. A very high one will turn the compressor into a limiter or even a distortion box. Since you’re just starting out, try settings between 2:1 and 5:1. I bet you’ll hear it working at 5:1 and you’ll probably settle on somewhere inbetween.
• Output is a volume slider to compensate for how much volume was dropped by the compressor. A good setting here will put your final peaks roughly where you want them to end up on your timeline.
• Autogain is an automatic volume compensation and really takes the place of using the output slide. However, it can sometimes allow your peaks to go too high, so you might try turning it off and manually adjusting the output gain.
• Input is a setting that is only occasionaly available. Most compressors assume that your input is already set.
• Attack is how fast the compressor starts to work once the threshold is reached. Faster is generally better for dialogue. Slow works great for slap bass guitar because it doesn’t compress the initial attack of the thumb slapping against the string, so you get a quick peak then suddon drop in volume for the rest of the note.
• Release is how fast the compressor lets go once the threshold is crossed on the way back down. If there’s an auto setting, use it. If not, set it for around 100ms till you are comfortable with everything else.
• Knee. Setting the knee to soft, which is generally a higher number, will soften the blow of the compressor and allow it to work a little more without noticing it as much. For dialogue, I use a soft knee.
• Compression meter vs Output meter. The compression meter shows how much the compressor is working. If your meters are vertical, it will start lighting at the top and working downward toward the bottom. It’s like turning a typical peak meter upside down. You’ll only see it if you’re actually compressing. The output meter is the same as the FCP peak meter. It shows the output level. Some compressors will have an input meter as well. It’s nice to compare the input to output so you can see exactly how much you’re changing it.

To get you started, set the ratio to 4:1 and start turning down the threshold till you see the compression meter working. If it’s working too much, staying near the bottom most of the time, your dialogue will start to sound like the Monster Truck Rally Commercial DJ. If it’s rarely moving, your not doing much. Find a happy medium. Then set the output so that the peaks are where you want them. Then compare by toggling between bypass and active. It’s helpful to have your peaks coming in already where you want them to begin with. And again, do this outside of FCP if you can. You’ll get a nice gui with the same plugin and it’ll be much easier and make more sense.

Then, you’ll probalby want to use a limiter, which is often built into the same plugin. It’s not much different from a compressor as far as settings go, but without the ratio or knee settings. Start by setting the limit or output to the exact level that you want for your peaks. Then start turning down the threshold adjustment (or up the input) till you just see it working occasionally. All it needs to do it cap off the extreme of the peaks. There are nicer mastering limiters out there, but just start with this excercise and save that for later. Once you get used to everything, start playing with input or threshold to make it work a little harder and see what it does.

Once you’ve done this to every dialogue clip in your program, your show will sound consistently even in volume from beginning to end, and you’ll be able to set your music a little louder and the dialogue can still cut through, and if you send your programs out to broadcast, they’ll be much happier with your audio.

Setting compression will never be a “cookie-cutter” experience, but if you find settings you like, save them as a preset. You’ll find you won’t have to tweak much, especially if the next clip you work on is for the same talent. You’ll probably only have to adjust the threshold between clips. If you always compress, setting a limiter is much more of a “cookie-cutter” experience though. Because most of the work is done by the compressor, the limiters settings will rarely need to change.

That’s my 2 cents on dynamics. Again, it’s subjective and I doubt in a group of 10 audio pros, you’d find 3 that agree on any one thing. This is just one method, that works for me. Hope it helps you.

Gary

[Tom Wolsky]

Worth a lot more than .02. Thank you for the detailed write-up.

All the best,

Tom

Author: “Final Cut Pro 5 Editing Essentials” and “Final Cut Express 2 Editing Workshop” Class on Demand “Complete Training for FCP5” DVD

[Don Greening]

[Tom Wolsky] “Worth a lot more than .02. Thank you for the detailed write-up.”

Indeed. Thanks, Gary. Your detailed explanation basically mirrors what I’ve been doing in ProTools for quite some time now. Just before xmas I was tasked with transferring 28 vinyl record tracks to 2 CDs for a local church group, whose chior had recorded everything back in the late 70’s to mid 80’s. I used ProTools for the recording, then I brought everything into Soundtrack Pro for noise, scratches and pops elimination. This was my first time using Soundtrack Pro and found it to be fairly proficient in doing that job. Then back to Protools for compression/limiting and final output. In 3 days 🙂 They wanted everything done in time for xmas so they could sell copies, mostly because there were a lot of xmas carols they’d recorded. Oh yeah, and I did the CD jacket and insert art at the same time. I think I charged around 200 bucks for everything, since it was for the local church and being it was xmas and all. They got a deal and and a half, alright.

– Don

[Glenn Woiler]

They got more than a deal! You gave ’em a BIG Christmas Present. What you did can take a lot of time.

[Gary Hughes]

Thanks for the kind words. I really appreciate it.

Gary