Need help: using a DAW EQ tool for balancing volume and analysis of EQ

  • Glad to be of service, 'Hookster!


    Yeah mate, if you can find a meter that offers a perceived-volume option, it'd be a simple matter of settling on a level you like, running a DI loop through the Kemper and knocking yourself out.


    There'd be no need to bother with fancy, schmancy filtering, white or pink noise or anything else.

    I agree with Monkey 100% that a perceived volume measurement would be ideal, but I don't quite agree about the input signal (guitar versus noise). The idea of using filtered noise is to capture every note and harmonic of the guitar in one shot, not just a single chord for example. Look at Flying Heel's images, you can see the individual notes. That makes it hard to judge EQ curves IMHO.


    For observing EQ curves, I'd use plain old white noise. IMHO this is the only way.


    For level matching, I'd use band-pass filtered noise that represents the whole frequency range of the guitar. Regular guitar loops might work OK here, but look at Flying Heel's images again and you can see all that missing spectral content that will be there as soon as he slides up one or two frets. Might as well test with everything there that is going to be there at some point.

    I hate emojis, but I hate being misunderstood more. :)

  • Thanks everyone. I have some 'progress' to report, but also, a true excedrin headache. :cursing:


    First, Reaper does include a pink noise generator. So, I have successfully been able to send pink noise to the looper, capture it, and play it back through different rigs and record the result into Reaper. A small win, but significant.


    Second, I found a free plug in, DP Meter, that allows you to look at the recorded signal in either ITU 1770 or EBU R128. I am not sure I am using it correctly, but here is a link:


    http://www.tb-software.com/TBProAudio/index.html


    I think I have a good, repeatable process/workflow for playing the pink noise through the various rigs and recording it, and then 'analyzing' it. The issue now is understanding the information I am being given and analyzing it. ?(


    This is the ITU 1770 analysis screen:


    [Blocked Image: http://i1368.photobucket.com/albums/ag198/flyingheelhook/Screen%20Shot%202017-03-19%20at%2011.44.28%20AM_zpskt2jswbo.png]


    This is the EBU R128.


    [Blocked Image: http://i1368.photobucket.com/albums/ag198/flyingheelhook/Screen%20Shot%202017-03-19%20at%2011.43.45%20AM_zpsejypkhgr.png]


    So, last night I ran a test using the ITU 1770 analysis and using the maximum peak of @ -18.9 +/- as my 'target'. I ran pink noise through 5 different rigs of increasing gain, one at a time, recording, analyzing, then tweaking to try and get each one close to that target. I save those amps and this morning tried to play through them and compare them to Crunch. Not entirely unexpected, but the only one that was close was the one with similar gain. The others varied quite a bit with the cleaner ones actually coming out much louder than I expected.


    This morning I attempted to tweak the cleanest to match Crunch using only my ear, playing at moderately loud volume. Both screen captures above show the comparison of the reference Crunch to the ear-balanced clean rig I am trying to get to match Crunch in volume. You can see there is quite a bit of disparity in all of the measured metrics even though to my ear, the volume match is pretty close.


    I am ok if I have to have a different number to calibrate to for different gain levels but which metric should I really be trying to measure and set my standards to? I am guessing RMS Integrated or Integrated Loudness?


    Thanks for you help. I need more aspirin... ||

  • Here's a graphical example about my suggestion to use noise to test EQ curves. This was done on a cheapo Behringer audio interface that I use on my electronics work bench. This wasn't run through an amp or sim, but rather just a straight loopback and a hardware EQ. The noise is a free white-noise loop I downloaded, and the guitar is an open E-chord that I was bashing quickly.


    I think the white noise is a better test signal for measuring the frequency spectrum of any device, be it a clean EQ or a distorted amplifier. You want to see the spectrum of the device under test, not the input signal.


    [Blocked Image: https://s27.postimg.org/oucdiwaub/Noise.png][Blocked Image: https://s15.postimg.org/67ui0e9i3/Noise_EQ.png]


    [Blocked Image: https://s4.postimg.org/iz1yqyie5/Guitar.png][Blocked Image: https://s13.postimg.org/hb2l2c3qv/Guiitar_EQ.png]

    I hate emojis, but I hate being misunderstood more. :)

  • @flyingheelhook, yes, use the integrated values. That is the perceived signal volume averaged over time.


    I'd also suggest trying white noise (which is a flat spectrum) and filtered white noise. Put shelving EQ's on the noise signal in the guitar frequency range, say 82 to 1200Hz to filter it.


    Then test your tests. See which method actually matches YOUR perception of volume when playing with your band. Take three patches for example and make identical copies of them. Level match one set by ear, one by pink noise, one by white noise, and one by filtered white noise. See which set sounds the most consistent at rehearsal and use that method moving forward.


    You are are on the right track IMHO.

    I hate emojis, but I hate being misunderstood more. :)

  • Thanks @Klappy!! I made a small win while I was offline - actually two. Reaper also has a white noise generator so now I can use both white and pink if I want, and I figured out how to run the noise through the Kemper and monitor the signal without having to record it so I can tweak on the fly. I am getting there!


    Thanks for your help, I will try to report on my progress later today if all goes well.

  • Now you've given me a headache, 'Hookster! LOL


    Well done on your progress so far, mate.


    I agree with Monkey 100% that a perceived volume measurement would be ideal, but I don't quite agree about the input signal (guitar versus noise). The idea of using filtered noise is to capture every note and harmonic of the guitar in one shot, not just a single chord for example.

    Fair point, Klapster.


    The counter argument will always be that one's playing guitar, not white noise, through one's Kemper, so I suppose an alternative would logically be to make the DI loop one that crams all the loudest ways one plays into it. It'd only require a second's worth of each "type" - dig-in power chord, dig-in lead note at most-resonant (loudest) point on the neck or whichever area one hears is the loudest (valid 'cause it's perceptual volume we're talking about), and hardest full (all strings), clean funk-style strum one can produce (just to cover the clean Rigs), for example.


    A very-short clip could be created of this stuff and, IMHO, quite-legitimately be employed in this levelling exercise.

  • and I need more...


    So, without going into copious detail on how I got there, suffice to say that even if I have two profiles of relatively similar gain (i.e. very clean), using a target RMS Integrated value to adjust the volumes does NOT necessarily mean that the perceived volumes will match. Reasonably close, but not close enough to save me any time using that method. *sigh* Anyhow, I'll keep digging. I have four different amps, each from different commercial profilers, each with a performance of 5 rigs of increasing gain. I used the RMS Integrated Value method (patent pending) to get them close enough that I can tweak them, but now I want to understand what their EQ curves tell me about how they sound the same or different. I did screen captures and I'm going to print them all out so I can look at them collectively and see what I can see. Stay tuned for the next episode of Teaching An Old Dog New Tricks...