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

Proper Gain Staging with High End Gear

Discussion in 'Recording' started by ChrisH, Feb 26, 2013.

  1. ChrisH

    ChrisH Active Member

    Hello everyone,

    Can't seem to get a straight forward answer about gain staging with high end gear.

    I'm using all transformer coupled preamps, API's and Dakings going into Apogee Symphony Converters.

    Up until then my recording peak levels into cubase was anywhere from -20db to -12db on a full scale.

    My Question, does that -12db to -20db "sweet spot" range still apply with high end pres and converters?

    If not, what range should I be tracking in?
     
  2. Kurt Foster

    Kurt Foster Distinguished Member

    yes. i assume you are using converters that run at +4dB (pro level). of course if you have your converters set for -10dB then you will not be driving your pres hard enough for optimal s/n.
     
  3. audiokid

    audiokid Staff

    To add to Kurt,

    DAW should be at -12 to -20 because this is the safe level so you aren't slamming your 2-bus when you get there, and most importantly! Others may disagree with me but I believe there is information that we can't see above that. I don't trust digital meters 100%.
     
  4. Kurt Foster

    Kurt Foster Distinguished Member

    and to add to what Chis just mentioned, what i found with DAWs is if you slam the levels when tracking, when you get to mixing you run into some real nasty sonic issues .. DAWs are a bottleneck in the first place and it's best to leave them some breathing room. too many bits spoil the soup.
     
  5. ChrisH

    ChrisH Active Member

    So at what levels should I track at?
     
  6. ChrisH

    ChrisH Active Member

    I wonder how I check what they are set at..
     
  7. Kurt Foster

    Kurt Foster Distinguished Member

    thumb
     
  8. Kurt Foster

    Kurt Foster Distinguished Member

    start at your function select and function adjust controls on the front panel.


    sio_fp_callouts-650.jpg

    i searched out the Symphony online and it does indeed have -10 and +4 ability for the A to D side. be sure you have the +4 selected for your mic pres ... some other devices like keyboards and semi-pro gear like tascam / fostex tape machines and mixers will want to see the -10 settings. no big deal to switch between levels as you track. it appears the outs are always @ +4dB.
     
  9. ChrisH

    ChrisH Active Member

    Thanks Kurt, we were probably looking at that at the same time. So should I track at somewhere between -12db and -20db with my setup when metering with my daw meter at unity?
     
  10. ChrisH

    ChrisH Active Member

    Also, on a sort of related side note, would it make sense that
    there would be less cymbal bleed in a Tom mic when tracked at -20db vs
    that same TomTom and mic position recorded at -6db??
     
  11. Kurt Foster

    Kurt Foster Distinguished Member

    i shoot for -16 with -20 being as low as i want to go and -12 at peaks ...

    no it's just the level you are recording at. relations of direct sound vs. background "bleed" is going to remain the same.
     
  12. Boswell

    Boswell Moderator Distinguished Member

    No, that is not the case. The ratio of the amplitudes of tom and cymbal in the mic output is determined by mic type and where you position it. This ratio remains exactly the same whether the mic is tracked peaking at -6dBFS or -20dBFS. You bring the tom up to the level you need at mixdown and the cymbal bleed is in there at the same fraction of that level.

    The gist of this thread is whether there is a difference in sonics between tracking at high levels and at lower levels, and also about whether a DAW mix sounds better if it is not having to cope with tracks at a high level. My view is that while individual tracks may show some sonic differences with tracking level due to A-D converter non-linearities, most DAWs will not behave differently with high-level tracks compared with low-level tracks, once the levels are brought to whatever is required for the mix. I remain to be convinced that there is a need for "mix headroom" in a correctly-coded DAW. Caveat: not all DAWs are necessarily correctly coded.
     
  13. Kurt Foster

    Kurt Foster Distinguished Member

    my experience has proven to me that on my DAW (Cubase) keeping the tracks at -12 or lower results in a more open mix ... i have been told that if i pack the tracks to near peak level, the function of bringing them down to levels the mix bus can deal with causes the cpu to do more math processing which can affect the sound. if you are going to pull the track levels down to less that -12 when you mix, what's the point of tracking them higher in the first place ???
     
  14. DonnyThompson

    DonnyThompson Distinguished Member

    This may indeed be the case with Cubase, Kurt.

    However, I haven't found this to be the case with Sonar, ( or at least in my experience I haven't found there to be an audible difference either way) so perhaps it is relative to the platform and its related method of processing being used. I think it's also relative to the quality of the pre amp and converters that are being used.

    I'm not suggesting that yours are inferior, btw. This audible difference you are hearing could simply be the way that Cubase processes the signal. If your method works for you, then that's all that counts.

    Now, I don't think Chris has much to worry about in relation to quality of converters, in his case, using the Apogee.

    LOL... certainly not nearly as much too be concerned with as someone who is using a Realtek or Soundblaster. ;)

    fwiw
    -d.
     
  15. Boswell

    Boswell Moderator Distinguished Member

    Well, this is where it gets interesting.

    Firstly, there are several factors involved in the tracking level, just considering the sonic quality of a single track for the moment. I mentioned one of them in the earlier post: any non-linearities present in the ADC conversion range. A level that minimises these commonly gets called a "sweet spot", implying that levels a long way below this can be noisy and grainy, and levels that take the peaks up near FS can exhibit distortions due to slew rate in the ADC drivers and other non-linearities in the conversion process. So most of us track at some way below what could technically be accommodated in the voltage range in order to stay in what we hear to be a sweet spot.

    Secondly, there are the unexpected sudden peaks that are the bane of all live recordists, but can also occur in standard studio tracking. These are genuine sounds that produce peaks above what we would anticipate in the type of program material for the recording being made. Tracking near FS may be able to capture the whole performance, but it may not if there are these unexpected peaks, and it's sensible to leave headroom for them to happen if they do.

    Thirdly, there are the rogue peaks, some of which are not detected by the level monitoring systems. I don't have a genuine engineering explanation for these, as they should not happen. When I have been asked to include peak measurement in pre-amplifiers that I have designed under contract, I go to great lengths to make sure that the peak detectors do detect the real peak of the input. However, there are some problems, as the crest factors (ratio of peak to r.m.s.) can be extraordinarily large, even for seemingly mundane instruments such as a piano. Despite all these, I see, and others have reported seeing, recorded peaks that hit the 0dBFS limit without the peak detectors flagging an error.

    Now all these (and there are more) are concerned only with getting the microphone sounds captured, amplified and digitised on to their own track. The prudent option for tracking level is that which sounds the best for that track, taking into account the sort of things I have mentioned.

    When it comes to mixing, there is no processing time difference due to the track levels, although a track at digital zero may process a little faster than one not at digital zero. Equally, the reduction in level needed from track level to mix level is just a multiply, and, provided it's not a 16-bit mix process, there will be no sonic difference due to the level of the source track. Where the differences can come in is at the summation stage, and this is what I hinted at by mentioning correct coding techniques in the previous post. Any DAW that uses floating-point for its summation process will not show any sort of overload at the mix stage unless there are shortcomings in the design to get back to the master mix levels. Fixed-point DAWs have more of a challenge here, and it could be where differences are heard between the manufacturers' offerings.
     
  16. ChrisH

    ChrisH Active Member

    Heres why I ask about my "More Gain = More Bleed" theory.
    In relation to the preamp stage in a tube powered guitar amplifier..
    Lets say you have a guitarist "Bob" strumming some chords on an electric guitar with the preamp gain set to 2 giving it the dynamic range of 20db-80db, then his buddy "Fred" comes over and starts tapping their finger on the body of the guitar, but Bob doesn't notice because his strumming is being amplified at 80db and the tapping is at 20db. Bob then turns the gain up to 9 still with a peak of level of 80db but now the dynamic range gap is shortened to 70db-80db. Again Fred comes over and taps on the guitar body but this time Bob noticed it because his strumming is still at 80db but the tapping is now at 70db.
    Now there's much less distortion going on in an API preamp than a guitar amp but it makes sense to me that the dynamic range would still change on a preamp with the gain turned up, thus closing the gap between that fat tom sound you worked so hard on and that honky cymbal in the background.
     
  17. Kurt Foster

    Kurt Foster Distinguished Member

    what Boswell said ... the song remains the same ... da da, da da da da daaa, da da, da da da da daaa

     
  18. ChrisH

    ChrisH Active Member

    Okay, so I have my Analog Level set to +4dbu, and digital reference to -16dbfs, how ever on the digital reference setting you can set it at -20,-18,-16,-14,-12,-10.
    What would you guys set it to for drums and basic rock tracking? I'm not sure how to utilize this setting. Does changing it change the digital reference level change the converter input level?
     
  19. Kurt Foster

    Kurt Foster Distinguished Member

    i would set it at -20.

    i'm pretty sure this takes a +4dB VU signal to the digital recorder at -20 digital reference.

    it's really just an output volume control for calibrating to the input of the daw. that's a pretty cool feature.
     
  20. ChrisH

    ChrisH Active Member

    Okay I'll set it at +4dbu and -20, now do i still track at -12 to -20 when metering with my daw meters? I use cubase 6 btw.
    I wish I could rap my brain around how that feature would be useful.
     

Share This Page