Mixing Levels: K-System vs Peak for cues??

[SladeKennedy]

Hi Taxi Folks -

I'm fairly new to this world of cues and really dig it. In my effort to improve and get more forwards I'm wondering at what volume everyone mixes/masters? I know that most folks mix to -3 to -6 dB before mastering, I tend to think this is standard. However, I like more dynamics in my recordings and have been recording at K-14 and mastering at K-12. There is a noticeable difference in the end as the recording is not as compressed.

My assumption is that the music editor would adjust the level to fit the use, but was wondering if one maybe preferred over the other. I will be putting together a couple of tracks that I mix and master at Peak and then at K-System just for discussion if folks are interested.

Thanks,
Slade Kennedy

[kova]

I think it's a matter of personal preference and the genre of music. Orchestral music, for example, should have more dynamics (and thus, less compression) than, let's say, a hard rock piece.

For me personally: when I mix, I try to keep a peak of -4 to -5 dB. If it's an orchestral piece, the peaks tend to be lower naturally so I'll adjust the volume accordingly to hit that mark. In rock pieces, I usually hit that mark way too easily so I need to adjust accordingly.

When I "master" (and really, all I do when I master is run it through a volume maximizing plugin), I shoot for an RMS of about -9 to -10 dB. I eyeball the meters mainly to make sure that I'm not killing my dynamics but when I'm trying to determine acceptable volume level, I always have a reference track at hand to see if I'm matching levels and so it becomes a balancing game between reaching those levels and not squashing my mix. Again, for pure orchestral pieces, I usually don't have a problem with getting the volume up; if I do, it means that I did something wrong at the mix level. For rock pieces, it becomes a matter of taste as to the amount of limiting.

[shorttonpro]

Your target audience should help determine the final level.

Some libraries require your tracks at a specific RMS level which keeps their sound more unified. And like Kova said, everything is genre dependent.

However, if you're pitching a Pop/Rock/Dance song/cue, the supervisor is listening to hundreds of songs, many probably from commercially released artist albums. If you track comes immediately after a loud track, yours could sound minuscule and unfinished in comparison as the human ear naturally believes louder=better. Yes, it is simply a matter of turning up the volume, but the smallest annoyances may give the supervisor a reason to pass. After he turns up your track, when he moves onto the next, he'll be blasted again and have to turn it down.

On the opposite end of the spectrum, editors like to be able to see edit points and dynamics in the actual waveform. So when a song looks like a brickwall, it makes their job more difficult (and might not sound so good). So with cues you don't have to necessarily compete with label tracks, but be in the same ballpark.

So yes, probably a good idea to keep your mixes lower and dynamic, and when mastering, make your best judgement call based on genre, who will be listening to the track, and what you're trying to accomplish.

[andygabrys]

I am all for discussion

For anyone not familiar with it, here is the basis of the K-system from the inventor himself http://www.digido.com/how-to-make-bette ... art-2.html

So there are a lot of things tied up in what the K system does and it becomes more complicated when you consider we are producing music in isolation which is then incorporated into a final production which will include SFX and dialogue as well as music, and under current standards the final product is subject to loudness standards in Europe and North America. Note that there are no standards for how loud music by itself should / could be.

For me, the biggest question is how do you incorporate the K system into your work while still producing music that is acceptably impressive and competes well with existing examples of production music and music that is actually placed in TV and film productions. As Kova mentioned above, strict A/B comparisons against genre specific ref tracks will lead to productions that are most similar. Using the K system will lead to more dynamic pieces certainly, but tuning your monitoring level to a setting where reference tracks sound "normal" will mean your tracks sound less impressive than the refs.

Why?

I have found that most commercially mastered material is always mastered hotter than K-12 (approximately -12 dB between peak and average level) for anything that isn't orchestral. Pop and pop rock usually has a peak to RMS difference of -6dB to -9dB. Some productions in genres that can tolerate more distortion without sounding audibly crunchy might be down as low as -3dB to -5dB.

Does it matter?

I suppose it depends on who your client is and what they are looking to hear.

Since most production music is pulled down in volume to not compete with dialogue when it runs on TV, it's almost certain that we don't need to master tracks any hotter than that -9dB peak to average difference (which is where I usually end up) or even close to that. As Bob Katz says, k-14 or even k-20 would do in a lot of circumstances.

But some publishers want to hear tracks the FEEL like a certain reference or a la track. In this case, even if the track is pulled down to run under dialogue, the track still needs to be produced to be super loud like the refs.

One thing that I have found (and is mentioned in the article above) is that music is listened to in a variety of different environments (home, car, phone) with varying amounts of background noise. So to get something mixed where you have a "really flat balance" and you can "hear everything" most people will monitor at an extremely low level to get fine balances. On my system that is barely audible. Has little to do with the K-system at that point. It's just a way of hearing everything at super low subliminal levels, which in some ways is relative to hearing your piece turned down really low under dialogue.

I have found through experience that (my) poorly balanced mixes sound really bad when turned down low under dialogue. You hear certain things sticking out, and you get little character of how the piece would sound at a higher pure music monitoring level.

So....this is just what I have found to work for me. I have calibrated my monitors to the 77dB standard with pink noise as referenced in Bobs article. I have spent a lot of time contemplating k-14 and k-12 and beyond and referencing against ref tracks.

In the end I work this way because it's not necessarily relative (strict application of the K system) in mixing specifically production music. Not to say that it doesn't make sense in general or somebody else might find a better way to work with it.

[Cwadroon]

I've been looking into this topic myself. (R128 and all that jazz) Was curious if there is an applicable standard for the music itself.
Thanks Andy, for clarifying! appreciate the info.

[Paulie]

From the article:

he next step is to examine a simple relationship between the 0 AVG level and the sound pressure level. For typical pop productions, our monitor gain has been adjusted to -6 dB (below the standard reference, which yields 77dB SPL with -20 dBFS pink noise).

So, to achieve this monitor setting does this mean I should generate pink noise at -20dB and adult the volume on my monitors until I read 77dB in a dB meter?

[WeWillWriteUaSong]

I'll start off by saying I do mostly hip-hop/pop/r&b . I have been mixing cues the same way I mix final instrumentals (beats) for clients. I have my speakers calibrated for k14 metering. While tracking/mixing I keep my levels on channels @ 0dbvu = -18dbfs - and keep my master hovering around zero on the k14 meter. When my mix is done and I go to master - i am careful to get the loudness i want - which if mixed properly - will master around -6 to -7 rms while only hitting the limiter around -2db gain reduction. I know thats loud....but as said earlier - i tend to master to the references-which in urban music is jacked up to the max (carefully). Even though it doesnt make audio sense to do this (its much nicer to leave headroom for vocals or scene dialog), that seems to be what libraries/clients are looking for. The stuff I have signed to libs has all been mixed/mastered this way and has been fine. I never got any loudness complaints.

Now when I am recording a client with a track I produced...I do not use that jacked up master version. But if someone asks for a track and I give them the mixed k14 version, they do not understand why it is so much lower than "youtube" beats. So you have to play the game. Like everyone said, depending on your genre.

[andygabrys]

From the article:

he next step is to examine a simple relationship between the 0 AVG level and the sound pressure level. For typical pop productions, our monitor gain has been adjusted to -6 dB (below the standard reference, which yields 77dB SPL with -20 dBFS pink noise).

So, to achieve this monitor setting does this mean I should generate pink noise at -20dB and adult the volume on my monitors until I read 77dB in a dB meter?

Paul - here is a more step by step version (quick read on it and it looked right in my view) https://www.gearslutz.com/board/masteri ... orial.html

Note that the author in this thread recommends setting your 0 level on your monitoring to equate to K-20 and then -6 would be K-14 and -8 would be K-12 but it doesn't really matter what you set it to. When I did it I happened to chose K-14 as -21 on my apogee interface as the volume knob has detents allowing precise preproduction. It doesn't matter as long as when you run the test tones it equates to 77 (for K-14) and 83 db if you do K-20.

Note also in the article where the mention was made of small rooms (which is really what 95% of us are working with) there is the issue of added reflections making a calibrated K-14 at 77db actually screaming loud. I agree. Its actually uncomfortable to be close to near fields (mine are each about 1.5 meters away from my head) when they are outputting that kind of level. Small rooms usually have really funky bass responses as well so listening at a really healthy level can also lead you astray.

So like I alluded to above - you might find that you do more of your mixing and mastering at quieter levels and crank it up to the K-14 (or whatever you choose) for short periods and ref to see how it sounds against other things that squate on the same scale (like the honor roll on Bob's site). As I, and Marcus and several others remarked, its not necessarily so useful if you already have references that are screaming loud - as that is your bar, not some k-14 or whatever it is.

[Paulie]

Awesome... I will check out the article. My room is about 12x11, 8-foot ceiling with lots of sound reading.

PC

[mojobone]

It's kinda hard to have this discussion with so many loose terms being thrown around; for instance, decibels SPL are not the same as decibels Full Scale and the relationship between the two isn't consistent and they shouldn't be conflated. These matters are thoroughly discussed in Bob Katz' Mastering Audio, which I highly recommend. In order to cut through some of the chaff, we should be talking about crest factor, which is the relationship between peak and RMS average. (RMS average corresponds most closely to what the ear perceives as loudness) You actually can mix without calibrating your monitor system, it's just that you'd never know where your levels actually are, but it's not the most important thing.

Crest factor is the important thing; match that to your references, leave a little dBFS headroom for the editors and you should be good to go. It's helpful to have a meter that reads peak and RMS average side by side, it's helpful to mix at very low levels to hear what sticks out, and it's helpful to listen at the recommended SPL to make sure the balance doesn't change too much when you turn up the wick and you haven't overdone the highs and lows. It's helpful, but not necessarily critical to measure SPL at the mix position, and the decibel SPL meter in your phone's audio measurement app is probably good enough.