When is Clipping a penalty? - Page 3
Page 3 of 3 FirstFirst 123
Results 21 to 25 of 25
  1. #21
    Tech Guru exokinetic's Avatar
    Join Date
    Jan 2011
    Location
    The best place on earth.
    Posts
    561

    Default

    You may not be able to see the actual measure DB that the mixer is either cutting, adding to, or acting like a wire and sending through at 0db.


    But the LED's are your reference for measurment of output.

    So, as you do your testing and you find the theoretical single LED that you cannot have illuminate or you absolutely clip.

    Depending on how many actual LED's are on the "VU" meter (that the little row of LED's that light up based on output level) You probably want to put a little piece of tape, or use white out, or some kind of paint, and mark either the LED just beneath the TESTED "clip" LED, or a few, depending how much "resolution" the LED's are displaying.

    Imagine a VU meter with 3 LED's. On, Optimum, Clip.

    And a VU meter with 30 LED's. where about 15-20 represent sub 0db, one represents 0db, about 5-7 represent optimum and about 3-5 represent clip.

    The more LEDs on your readout, the more LED's of "headroom" you should leave between your new "Max" marker(I.E. Where you put the tape) and the "tested" clip LED.


    Hope that helps clarify some of it.
    I'm addicted to WoW.


    Please do not bother me about being a productive member of society.

  2. #22
    Tech Mentor
    Join Date
    Apr 2009
    Location
    WA, USA
    Posts
    148

    Default

    Depending on how many actual LED's are on the "VU" meter (that the little row of LED's that light up based on output level) You probably want to put a little piece of tape, or use white out, or some kind of paint, and mark either the LED just beneath the TESTED "clip" LED, or a few, depending how much "resolution" the LED's are displaying.

    Imagine a VU meter with 3 LED's. On, Optimum, Clip.

    And a VU meter with 30 LED's. where about 15-20 represent sub 0db, one represents 0db, about 5-7 represent optimum and about 3-5 represent clip.

    The more LEDs on your readout, the more LED's of "headroom" you should leave between your new "Max" marker(I.E. Where you put the tape) and the "tested" clip LED.
    Uh, I understand up to the point of marking the LED just beneath the tested clip. That part makes sense. Why should I mark further down then that? What is this "resolution" the LEDs are displaying that you refer to? What is the purpose or reason behind marking off more headroom on mixers with more LED's?

    My current mixer has 8 LEDs at sub 0 db, 1 at 0 db, and 3 in the red. I haven't tested yet, but I am guessing the clip is gonna happen at the 1st or 2nd red. This is on the master gain as unfortunately there are no levels on the channel gains.

  3. #23
    Tech Guru exokinetic's Avatar
    Join Date
    Jan 2011
    Location
    The best place on earth.
    Posts
    561

    Default

    Thanks for the infor on your VU!

    That makes it easier to explain.

    Given that orientation, I would highly advise using the "0db" as you "peak" LED, and try to never let it go above that.

    If you find that that LAST LED on your meter is actually where it starts clipping the signal...

    (i.e. not just audible distortion in normal musical material, at soft clipping levels it is more of a "muddiness" where some frequency ranges feel like they get "lost", so its difficult to really hear if you don't know what to listen for, AND if you are really used to listening to clipping subs)

    ...then I would put your marker on the first red LED above the 0db and use that for your "peak". But something tells me the mixer manufacturer already did that testing for you, they have to in the design process of the mixer.


    So it is almost 100% safe to say the best level to ensure that you completely avoid clipping out of your mixer is to keep it at or below the 0db




    Now it seems your mixer only has one meter for the master output, and not for the channels?

    This can cause some sneaky things to happen, and the combined level of both channels is being measured on the master output, but this is only based on how much extra the master is adding.

    If the two channels are sending a really hot (loud) signal to the master bus, then you don't have to turn the master volume up as high to hit 0db.


    But if those channels are turned up so hot that they themselves are clipping (the channel gain is usually the worst culprit for DJ's that like to "take it to 11" for the terrible sound of clipping).


    SO, even though your mixer is reading 0db, your channel bus is clipping the signal.

    Your mixer has as many channel buses as it has channels, unless the manufacturer used a lower fidelity single board for the "mix" bus, containing all channels, and sometimes even the mic too!

    If any bus in your mixer clips, the master out, no matter the level will be deliver a clipped signal to anything after it.


    So your most beneficial testing is going to be to find how far you car turn your channel gain knobs at a given input volume (i.e. a louder track might clip at a certain knob level, but a quieter recording may not) with ought them clipping.

    It would be really easy, especially for a DJ with little regard for "red" lights, to completely clip the signal to oblivion at the channel if there are no meter to even tell him hes hitting the red.
    I'm addicted to WoW.


    Please do not bother me about being a productive member of society.

  4. #24
    Tech Guru mostapha's Avatar
    Join Date
    May 2010
    Location
    Atlanta
    Posts
    4,748

    Default

    What equipment, specifically, are you using? If I knew that, I could probably help more specifically.

    Most DJ mixers have some kind of metering on the channels, though it may not be obvious. My xone:62 has several with different levels of detail…at least, you can trick it into giving you several. In addition to the 4-LED meters on each channel, the "master" meter will read the level of the cue signal instead of the master out if anything is routed to the headphone circuit. It's almost annoying behavior that I don't actually like, except that thanks to Traktor, I don't have anything cued when I'm actually mixing. Well…not right now. I probably will have to when I switch to TP2, which means I'll be looking for a different mixer.

    As another example, the Rane Empath has no meters on the channels, but the one meter on the thing can switch (via a button) b/t the cue signal and the master, which you can use to set your channel gains. It just takes a bit more time. Most battle mixers have a switch to change between channel meters and a master meter.

    Where the gain makeup (from a quiet signal out of Traktor, for example) should happen depends on the quality of different parts of your mixer/system…and mostly you have to figure it out by experimentation unless it has good metering. On some mixers, the EQs, faders, etc. are quiet enough (quiet as in low-noise) that you can run the channel gains a bit cold and turn the master up. The danger is in compressing your dynamic range, though it's not a huge risk if you're using 16-bit files and good 24-bit DAC. On others, the faders/eqs/etc. are noisy enough that it's better to run the channel gains a bit hot and run the master lower, just making sure not to clip the channels.

    On modern digital mixers, like the DJM-800, it's really hard to clip inside the mixer (i.e., clipping a channel) because of the same types of tricks that Traktor and every other piece of software use. Again, it's really easy to clip the final output unless they've done some serious work to try to prevent it. Even then, weird things might be happening to your signal that you don't know about and can't control. For example, it might be using something as lousy as Traktor's internal Limiter to keep you from clipping, but I don't know that because I haven't done any testing…every time I've played on a DJM-800 it's been at a party or when I was trying to sell them when I worked at GC.

    In general, though, modern mixers make it really easy and quick to figure out these issues. Back in the days of the Urei and Bozak rotaries, the mixers didn't have any metering…you were just supposed to use your ears, and maybe you bought (or the club installed) a single meter for the master output that wouldn't read anything else, because it was plugged in after the mixer.

    The biggest complication on modern mixers is that a lot of the meters I've seen are just labeled wrong: 0dB might mean 0dBU or 0dBV, which can be very different things. A&H Xones have been criticized for having really wimpy outputs. A 0dB signal at unity (flat EQ, 0dB gain, 0dB channel fader, CF off, 0dB master) will not read 1V on the output like it should if 0dB means 0dBV. They appear to have labeled their meter wrong, presumably to try and trick DJs into not running the mixer too hot. Unfortunately, that backfired and just lead everyone to slam the master level into it's "clip" LED, which apparently still gives you some headroom. I've personally watched a lot of big-name and local DJs run it so hot that the clip light stayed on and the LEDs below it barely turned off except during breakdowns. I was wearing earplugs that rolled off highs every time I've seen it, so IHNFC if they were getting into the "soft clipping" range or not.

    (offenders off the top of my head: Armin van Buuren, Steve Porter, Sasha, PvD, Oakenfold, Derrick Carter, Ferry Corsten, and well…everyone else I've seen play on a xone)

    What seems to work best for my xone:62 is to have the channel faders never peak above 0dB with the master fader all the way up (I have an older UK-made one that has a fader instead of a knob). A track at unity will peak at about +3dB (clip light is at +6dB, IIRC), which sounds fine to my ears and still needs a wide-open gain on my sound card for recording without coming close to clipping my sound card inputs regardless of whether I record with Traktor or Ableton.

    As an example, my last mix CD (available on mixcloud).



    It was recorded that way (mostly peaking around +3dB on the xone) with my FS Open's input gain all the way up and Traktor's internal recording set at Unity. It doesn't look like I lost any dynamic range, the peaks aren't clipped (I'd have to zoom in to show you), and it still needed like 8 or 10dB (IIRC) of boost to get the loudest peaks up to about -.3dB.

    Either the recording out on that mixer is really quiet, or 3dB doesn't mean 3dB on the master. I haven't taken an oscilloscope or volt meter to the outputs to figure out which.

    It sounds fine, though, cuz having to add about 10dB to get it to 0, means it was really about a 22-bit effective resolution signal (instead of 24) based on 16-bit files, which is fine.

  5. #25
    Tech Wizard
    Join Date
    Feb 2011
    Posts
    37

    Default

    Clipping anywhere in the signal chain is just going to be transfered allong the chain.

    Inside the mixer, soundcard or whatever your signal source is there is a small amplifier that amplifies the signal which goes to the amps.

    SO if your source is clipping your are sending a clipped signal to your amp which will be clipping your speakers.

Page 3 of 3 FirstFirst 123

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •