1. Register NOW and become part of this fantastic knowledge base forum! This message will go away once you have registered.

de-essing school

Discussion in 'Recording' started by Hack, Jul 16, 2003.

  1. Hack

    Hack Active Member

    I have always had problems de-essing. I cant seem to get it to work right. People tell me to de-ess around 1.6 - 3k. BUt the freqs i hear jumping out are higher like 6 - 8k. I have a avalon 737, which should de-ess perfectly, according to the manual. By time I get the comp to kick in on the esses its kicking in on all kinds of stuff. And the ratio between ess compression and program(for lack of better term) compression isnt enough to make it worth doing. So I eq the track. What do I get?... you guessed it, a dull vocal track. I hear so much hi end in vocals these days and yet the esses are perfectly squashed. Is there something that I can listen for when tracking to aviod this problem? I usuall place the vocal mic just above the singers mouth and aim down at the mouth. I do this so I dont have to use a pop filter. Maybe this is the wrong mic placement. Suggestions???
     
  2. Jbuntz

    Jbuntz Guest

    1.6-3k is low for a de-esser. Try 4k or higher. Trust your ears. Depends on the mic, singer de-esser (different algorithms between de-essers), etc.
     
  3. RecorderMan

    RecorderMan Distinguished Member

    Try finding a used dbx263 deEsser. It's the cheaper half rack version of a dbx902...but is jsut as good. A VERY good deal.
     
  4. Alécio Costa - Brazil

    Alécio Costa - Brazil Well-Known Member

    if you are into a DAW, let deesing for the mixing stage. if ya are a PT TDM user, insert a digi deesser along with a PAZ (Waves RTA).

    Verify where are the peaks happening mostly and work with the de-esser setting. For example, my voice tracked with an At4050 thru my console mic pre/dbx analog comp is around 8khz. But I have faced situations up to 12khz. :p:
     
  5. Hack

    Hack Active Member

    I usually record vocals thru an at4060. I have used, of all things, a peavey de esser which actually worked. But its not mine, and peavey? I'll look into the DBX263.
     
  6. mixopenta

    mixopenta Guest

    I also find it very hard to obtain decent de-essing. In some ways one also has to take the mastering in account, as the hi-end (around 16-20k) always is boosted slightly.

    I'm using the dbx263 but I'm not sure I'm getting the settings right. If I set it to 4k, it sounds like frequencies above is slipping through, and if I set it to 8k it's the frequecies below that jumps out.

    The 263 has a normal/broadband switch, but I really can't decide in which mode it should be. I guess it should be "normal", but if someone could shed a light on that I'd be very happy.

    I usually record vocals with C414 in cardioid, homemade popfilter (stocking on a pianostring) eq flat, and through dbx 163 compressor and Mackie console to analog tape. I'm not de-essing when tracking. The compressor is usually set to work around 4-8 dB gain reduction.

    For this particular mix (I'm almost done now) I ended up making a compromise by reducing the de-essing not to make it too audible, and roll off some high end. And as Hack said, the vocals becomes a bit dull sounding...
     
  7. Alécio Costa - Brazil

    Alécio Costa - Brazil Well-Known Member

    try WAves or Digi de-esser. A thresold around -23db for very strong sibilant singers shall do the job.

    One trick: USe PAZ to evaluate the frequency content of your vocal track without and with Reverb sends engaged. It can alo be that the high frequency damping of your reverb patches are not properly set. Some LPF/HPF might help tailoring the overal sound.

    :p:
     
  8. falkon2

    falkon2 Well-Known Member

    A combination of a sonogram and a spectogram will tell your eyes better than what your ears can describe, the frequency ranges and threshold you need to set respectively.
     
  9. Alécio Costa - Brazil

    Alécio Costa - Brazil Well-Known Member

    It is very helpful but nothing compares to nice trained ears for final evaluation.
     
  10. RecorderMan

    RecorderMan Distinguished Member

    Let's take this in order:


    I usually record vocals with C414 in cardioid, homemade popfilter (stocking on a pianostring) eq flat, and through dbx 163 compressor and Mackie console to analog tape. I'm not de-essing when tracking. The compressor is usually set to work around 4-8 dB gain reduction.
    [/QUOTE]

    First, you are going into the Mackie (for the Pre) , and then into the 163x on the insert....right?

    Second, If you (or whom ever is singing) is very sibilant you cab try differrent mic and/or mic them from more of the side, rather than directly in front. This helps to varying degrees. Try backing off the compressor a little hit it at round 4db .

    The 263 has a normal/broadband switch, but I really can't decide in which mode it should be. I guess it should be "normal", but if someone could shed a light on that I'd be very happy.
    [/QUOTE]

    The "normal/broadband" switch tells the 263x weather to help work on only the high frequencies by using a hi-pass filter before the detector, while in broadband mode the whole signal goes to the detector. If there's too much low-end stuff like p-pop's ,then "normal" mode can work better. It's one of those things where after you've got it set up, switch in/out and pick which sounds better...if you can't hear a diff. pick one. On the Big Brother DBX 902 deEsser, usually it's left in the equivalent of broadband mode, and it works just fine.

    I'm using the dbx263 but I'm not sure I'm getting the settings right. If I set it to 4k, it sounds like frequencies above is slipping through, and if I set it to 8k it's the frequecies below that jumps out.[/QUOTE]

    Contrary to the oscilloscope crowd (just a little humor there) DeEssing DOES work best in the 1.8K to 2K to 4K range. On a DBX902 it's around 1.8K for males and 2.5K for females....the 4K & 8K are more like the "first harmonics" of the Ess fundamental...regardless of what the analyzer looks like.
    Start by hitting the DeEsser really hard...like 10b of 'Gain Reduction'. Then sweep around the 1.5K to 3K region untill you get the maximum reduction in 'esses. Then...back off to maybe 4-6db of gain reduction. BUT do this by ear...back off untill you hear an 'S' ..but not a sharp/sibilant 'S'.

    For this particular mix (I'm almost done now) I ended up making a compromise by reducing the de-essing not to make it too audible, and roll off some high end. And as Hack said, the vocals becomes a bit dull sounding... [/QUOTE]

    DeEss first in your chain, then EQ. If you DeEss right, yuo should actually be able to ADD a shelving EQ , anywhere from 5K up to 12K (depends on source & mix) a few DB for air.

    -------------------
    Now for the DAW. I also use the waves DeEsser (they're great!) when I'm tracking/OD'ing in my daw...if I need to do that...hopefully I've found an acoustic way around, but sometimes that's just how they sing.
    But, if I'm mixing w/ a daw (even if I'm mixing on an analog board and the daw is my tape machine) THE BEST WAY TO DE_ESS is to not use a DeEsser at all, and instead zoom into all the spots where there's a 'S' to fix and do a volume ride (anwhere from 4-6-10-? db) on just that part of the beginning wave-form where the huge 'S' build-up is...THIS WAY IS THE BOMB.


    -----------------
    UNRELATED DeEsser Trick
    ----------------
    As an unrelated bonus...here's something you can use as outboard analog deEsddr for if you are trackng drums.
    For Snare. Weather you use 1 mic on top, or two (top & bottom), etc. bus the snare to two tracks. Track #2 will be our unprocessed snare for when they go to the sidestick. On bus# 1 go through a DeEsser first, then a compressor. Hit the DeEsser hard @2.5k with 10db of gain reduction, then kiss the compressor @4:1 with a 1/4 to 1/2 db of gain reduction. This works great on loud rock drums with heavy hat. Your OH's still get the "air" of the snare..this will give you more shell. And you switch to or ride up the unprocessed track for side-stick sections.
    ------------------
     
  11. mixopenta

    mixopenta Guest

    Thanks Recorderman, Great tips. I will try them out right away, as I'm scheduled for doing vocaltracks this weekend. :)

    RO rocks! :p:
     
  12. Alécio Costa - Brazil

    Alécio Costa - Brazil Well-Known Member

    I would de-ess post eq.
    Comp > Eq > de-ess. :s:
     
  13. Kurt Foster

    Kurt Foster Distinguished Member

    ACB,
    Like RM says, I would de ess first, then comp and eq, or eq and comp. The de esser will work better on a un compressed signal and then you will have a cleaner track to process with eq and dynamics.
     
  14. realdynamix

    realdynamix Well-Known Member

    :) A diamond in the rough, thanks RM!

    --Rick
     
  15. RecorderMan

    RecorderMan Distinguished Member

    anyyway that works for you works....period.

    I myself generally chain in this way.
    DeEss (if I'm using one)-limiters (High threshold and high ratio, gain reduction of about 1db)-compressor (low ration; 2.5:1 to 3.5: to 1 at about 4-5db if it's a 160, if it's an 1176 then 4:1)-then EQ. Just one scenario. As Andy Johns has stated in the past...one good reason to compress first is that compressors by thier nature knind of reinforce and build up the fundamental, rollin off the top as they do. So If you're trying to add top with the EQ your fighting yourself (not to mention the old adage that if you tweak your qe your affecting yuor compression). Where I do stray from this is either if & when & use a high-pass filter (first in line) or just some subtractive EQ (usually on a badly recorded tack).

    But...again...whatever get's you through the night.

    How's the weather down south...today's the firsr break in about three weeks from a heat wave we've been havin' in ol' LA.
     
  16. falkon2

    falkon2 Well-Known Member

    Wow... great tips.
     
  17. Guest

    In this particular "Los Angeles vs. Brazil" controversy, I vote with Los Angeles.

    However, futbol would be a different story. Gots to go with "the beautiful game..."
     
  18. Kurt Foster

    Kurt Foster Distinguished Member

    Harrumph! IMO the bestht de essther on the planet! You can make sthomeomne sthound like they have a sthpeech impediment with it! :D
     
  19. paul lani

    paul lani Guest

    i believe that the main difference between the 1/2 rack
    263 & the 902 is that the 902 has a bit more headrom. they are great. i sometimes de-ss pre vocal
    processing and post - because if i am doing really
    heavy compression , it usually lets that first transient still
    spike thru. sometimes with the right compressor like an
    la 2a the attack on those helps take care of it.
    speaking of 1/2 rack dbx stuff, i know that
    back in the day tom alge used to carry around the compressor unit of the same model ( not the d-sser )
    for his lead vocals. what i would also do on an ssl is
    on the old e's you could side chain the eq to dynamics.
    or better yet , id sometimes use an extra fader to
    key the neighboring vocal fader so that if an ssss was
    really bad id push up the trigger fader. but what it
    really did for me was allow me to not key the F's &
    the th's. you could set this up with a daw too.
    now i just use an waves ren d-sser.
    s
     
  20. Heh, heh,
    Speaking as theplug-insForum Guy, I have to say that if you want you use bypass automation on a plug-in de-esser to have it activate only for "esSSes" and keep it off for the rest. I find the Waves de-esser is so good it doesn't even need that trick. It is one of the bestplug-insIMHO.
    Kuul trick for the drums R-Man. David
     

Share This Page