Dealing with clashing frequencies

[Dave Boulden]

I'm always up for trying out different techniques for mixing. One of the more recent things I've tried is scooping a dip in instrumental tracks by EQing the group busses to leave a "hole" somewhere around the 1kHz-4kHz range (dependent upon the harmonic content of the particular vocalist) for the main vocal to sit in. I decided to try this as I was mixing a song that had been tracked over in the US and it was, to say the least, rather busy musically.

Normally, I would try and address the coverage of frequency ranges at the arrangement stage to avoid frequency clashes at the mixing stage. Does anyone else scoop group tracks to make a vocal sit nicely in a mix? Do you use it only when needed or as a matter of course?

[TrevCircleStudios]

I'd recommend against this approach as, because no given vocal is a single blob of frequencies inevitably you'll be left with a hole around that area. The key is to sit the vocal in the mix where you want it (perhaps using a little compression) and then EQ conflicting frequencies out. That means using a fairly surgical EQ with a narrow Q and sweeping the spectrum with a high level of boost to find the worst offenders. When you find them reverse the boost of those frequencies into cuts. Voila, your vocal will now sit more happily in the mix.

Hope this is useful!
Trev

[Dave Boulden]

So better to do tight surgical spot EQ cuts on the actual vocal track where it clashes? I nearly always use UAD Cambridge for EQ... unless all that's required is a simple low bass roll off when I tend to just use the standard Cubase Type I LPF.

If so, is it better to automate the EQ and only cut where an instrumnental part clashes with it, or apply as a static surgical EQ cut for the length of the song?

[TrevCircleStudios]

Generally speaking once you've isolated the clashing frequencies (there are onlu usually 3 or 4) you can leave it for the song. I'd only tend to automate a sweetening type EQ (in UAD terms the Pultec or Neve say) if I wanted a very different sound between, say, verses and choruses.

[Dave Boulden]

Thanks Trev

[TrevCircleStudios]

no worries.

[TrevCircleStudios]

It just occurred to me that it might not have been clear from my post. Often it will be the other clashing instrument that you cut rather than the vocal. You just gotta try it and see which one works best.

Another thought is that if say you cut the instrument (though actually whichever you cut) you might try a very small boost of the vocal at the same frequency to help it out (though as always that's only if the song needs that) .

Oh yeah, and finally - as a general rule (and rules are of course to be broken if that's what the song needs) - when you are cutting you should use a narrow Q but when boosting you might first try something a bit gentler.

[FrankGIK]

Originally Posted by TrevCircleStudios

It just occurred to me that it might not have been clear from my post. Often it will be the other clashing instrument that you cut rather than the vocal. You just gotta try it and see which one works best.

Another thought is that if say you cut the instrument (though actually whichever you cut) you might try a very small boost of the vocal at the same frequency to help it out (though as always that's only if the song needs that) .

Oh yeah, and finally - as a general rule (and rules are of course to be broken if that's what the song needs) - when you are cutting you should use a narrow Q but when boosting you might first try something a bit gentler.

All good advice. Vocals typically conflict with guitars, keys, synths...the stuff that's heavy in the 1-4Khz area...you'll find the same conflicts over and over again, and you'll learn how to make some general seperation judgement calls based on the style of music. For instance, metal is easy...typically guitars are scooped in the midrange anyhow, so there's your vocal. Singer-song writer kind of stuff can be tricker...hard to get a nice, woody acoustic sound without conflicting with the vocal, but in that case the ACG is a little more backgroundy so panning or tracking a stereo acoustic line would be your friend.

Lots of different solutions.

Frank

[TrevCircleStudios]

As Frank says, there's a lot more than one way to skin a cat. Another to play around with is ducking using a sidechained compressor. All good fun!

[edit... and here's a more detailed discussion of the subject: Side Chaining ]

[TrevCircleStudios]

It occurred to me as I submitted the last post that some people might not know what I mean by ducking. Being inherently lazy I couldnt contemplate typing out the process. I just did a quick google search and copied this straight from another forum:

If you're making anything with 4 on the floor (dance music) sidechain compression between kick and bass is almost imperative to get that glue and extra 2 dB headroom.

This example is based on Logic Pro but the concept is very similar in most other sequencers that have external sidechaining (key in) .

First you still need to equalize and compress both the kick and bass separately as usual. The sidechain compressor is an extra compressor you add at the end.

10 Step Guide
1. Output the kick sample to a bus (e.g. Bus 1). The reason being that Logic can't sidechain directly to an audio instrument, only to audio channels or busses. So this is a workaround.

2. Insert a Logic compressor at the very end of the chain on the bass instrument. Remember this is an extra compressor for the sole purpose of sidechain compression. Be sure to use the default settings of the compressor (very importantly: no Autogain, no limiter etc.)

3. Set the compressor's external sidechain (the Logic compressor also has an internal sidechain via the expanded options - ignore that in this case) to Bus 1. You set the external sidechain input in the upper right corner.

4. Set the attack to 0 ms. You want that bass ducked immediately to gain headroom. The attack of the kick will act as all the punch you need in the mix for that split second when the bass is ducked. This is partly a psycho-acoustical phenomenon.

5. Set the knee to hard (0.0). You want the peak of the signal to be the trigger, nothing else.

6. Set mode to "Peak". For the same reason above. RMS mode is sometimes used for delay ducking, which is slightly different from kick/bass ducking.

7. Set the gain to 0.0 dB. You don't want to change volume otherwise, so set it to unity.

8. Set a ratio between 4 and 10. How much depends on how much ducking you want. A ratio of 10 will almost remove any bass attack which might be just what you want.

9. Set the release time depending on how fast you want the bass to stop ducking, usually fairly quick, e.g. 20 - 50 ms. Longer times will make it pump more, which could be a desired effect - so experiment. Auto release is optional but primarily recommended when used in combination with more varied rhythms and note lengths than a standard kick/bass pattern.

10. Set the threshold according to level and trigger point for the ducking, use your ears and look at the meter. About 3 dB of gain reduction is usually more than enough. Remember that in this scenario you just need the very top of the signal to react, so don't pull that threshold down too far, you'll gain nothing extra.

Not only can it make your track sound slightly faster and improve perceived timing, it'll probably make it a good 2 dB louder without any noticable side effects.

The closest you get to a magic increase in loudness potential in dance music.

Other great uses for sidechaining might include:

Ducking an acoustic guitar by sidechain compressing it to the vox track
Ducking background or ambient noise by sidechaining the special fx track
Ducking the rhythm guitar track by sidechaining the guitar solo track

[TrevCircleStudios]

Dave it occurs to me that this thread is starting to become a useful resource for people who have issues with clashing frequency issues in a mix. Is it worth changing the title to "dealing with clashing frequencies in a mix" or something similar so it is a bit more searchable?

[modz1]

Originally Posted by TrevCircleStudios

Dave it occurs to me that this thread is starting to become a useful resource for people who have issues with clashing frequency issues in a mix. Is it worth changing the title to "dealing with clashing frequencies in a mix" or something similar so it is a bit more searchable?

I'm not 100% sure (will look into that tomorrow) that you can edit your own title as OP Dave.. cool if so..
but with your permission, I can take care of Trev's suggestion if that's not the case..?

[Dave Boulden]

Originally Posted by modz1

I'm not 100% sure you can edit your own title as OP Dave.. cool if so..
but with your permission, I can take care of Trev's suggestion if that's not the case..?

Please do... I'm all for making useful information easier to find!

[EdRyan]

It's a good thread.

I've always found that panning things slightly apart off centre helps me out as well. Not too far mind you, as if your mix is ever played in mono it may fall apart and have more issues then you solve.

I was sent one track that collapsed so badly in mono because of over zealous panning and mix techniques. Things simply went missing, and they did something funny with the kick because that sounded al weak as well. Either way I mixed it out after a minute and binned it.

Mono still gets used an awful lot.... particularly in nightclubs. For some bizarre reason, nightclub engineers seem to love mono-bridging sound systems. I know you get more more bang for buck, but it still baffles me as to why they force mono listening.. stero should be a minimum. Considering the size of most venues and how many speakers they have, you could do some amazing multi-channel entertainment. But no, as ever.... let's just have one, or if you're lucky you can have two and listen to the track as it was intended to be heard...

[FrankGIK]

Originally Posted by EdRyan

It's a good thread.

I've always found that panning things slightly apart off centre helps me out as well. Not too far mind you, as if your mix is ever played in mono it may fall apart and have more issues then you solve.

I was sent one track that collapsed so badly in mono because of over zealous panning and mix techniques. Things simply went missing, and they did something funny with the kick because that sounded al weak as well. Either way I mixed it out after a minute and binned it.

Mono still gets used an awful lot.... particularly in nightclubs. For some bizarre reason, nightclub engineers seem to love mono-bridging sound systems. I know you get more more bang for buck, but it still baffles me as to why they force mono listening.. stero should be a minimum. Considering the size of most venues and how many speakers they have, you could do some amazing multi-channel entertainment. But no, as ever.... let's just have one, or if you're lucky you can have two and listen to the track as it was intended to be heard...

All good points Ed. The simple solution is to switch to mono and back during mixing. You can still pan and EQ pretty radically while still maintaining mono compatibility, but you have to check frequently. Your issue with the kick drum is another thing to discuss...I'd be willing to bet the kick was interacting destructively with the OH's...yet another thing checking a mix in mono would show.

Mono is still used in clubs because the speaker systems are distributed throughout a large-ish building...now way to ensure good stereo coverage, so each area gets a mono mix+sub.

Frank