This discussion is locked
EdwardRyan

Q: Test signals

Just wondered why they missed out pink and white noise as test signals?

The manual mentions using i.R utility to also record hardware, but sinewaves and swept sines aren't really that good for testing hardware. Noise is much better for sampling hardware...

Macbook 2.2, Mac OS X (10.5.1)

Posted on Feb 19, 2008 4:21 AM

Close

Q: Test signals

  • All replies
  • Helpful answers

  • by John Buehler,

    John Buehler John Buehler Feb 19, 2008 7:06 AM in response to EdwardRyan
    Level 2 (385 points)
    Feb 19, 2008 7:06 AM in response to EdwardRyan
    White or pink noise can not be used to create an impulse. White or pink is only good for eq convolution, not reverb convolution. If one is trying to capture the eq curve of an outboard eq, white noise and the Match EQ plug is the tool for that job. But for reverb convolution, there can only be one point in time for every frequency. There is nothing wrong with using sweeps to capture a reverb or delay. The spiking method is also a choice if digital in AND out is possible, but I have not gotten good results that way.
  • by EdwardRyan,

    EdwardRyan EdwardRyan Feb 19, 2008 8:16 AM in response to John Buehler
    Level 1 (0 points)
    Feb 19, 2008 8:16 AM in response to John Buehler
    hi John.

    That is actually what I'm doing. I'm sampling an analogue filter unit. Have you got any more tips ?
  • by John Buehler,

    John Buehler John Buehler Feb 19, 2008 11:58 AM in response to EdwardRyan
    Level 2 (385 points)
    Feb 19, 2008 11:58 AM in response to EdwardRyan
    Use the match eq plug. First input the white noise into the eq to provide the template, then do a pass of white noise thru the filter, taking notes as to the settings of the filter. Once you have done that, you can generate the curve. The match eq plug allows you to apply the curve in percentage, from +200 to -100 percent. So if you do your filter pass with it set to it's extreme setting, you can then apply only 50 percent of the curve to the signal. I have done a lot of eq sweeps, and one that came out nice was the low end roll-off from a Trident Series 80-B, but I use it at a negative percentage, so it becomes a low end roll-up instead. It is this method that allows the "undoing" of a microphone's eq curve, by flattening it with a negative curve. Does that make sense?
  • by EdwardRyan,

    EdwardRyan EdwardRyan Feb 19, 2008 1:15 PM in response to John Buehler
    Level 1 (0 points)
    Feb 19, 2008 1:15 PM in response to John Buehler
    I kinda understand, but why do I want to use match eq? I'm not trying to recreate the sound of the filter using an eq model. i'm trying to sample it to build a convolution process that can represent most of the filters states. So I will need many samples, and a bit of data interpolation to do this..

    I thought I could pass pink noise through the filter and then inverse the noise. The resulting file should then just be the response of the filter in terms of freq, amplitude and phase.
  • by John Buehler,

    John Buehler John Buehler Feb 19, 2008 3:09 PM in response to EdwardRyan
    Level 2 (385 points)
    Feb 19, 2008 3:09 PM in response to EdwardRyan
    So it sounds like what you are trying to do is capture not only the spectral components of the filter, but the overtones and distortions that are added to the signal as well. This is most like how Guitar Amp Pro works.

    Space Designer and the IRU are for convolution in the time domain. Match EQ is for convolution in the spectral domain. A filter is a spectral process, not temporal, with the exception being distortion, harmonics and overtones. So I would do this..

    Do what I said about Match EQ, to capture the spectral components of the filter. Then, use a sweep with the IRU thru the filter to capture the overtones and distortions, and generate a Space Designer preset. The end result would be a Match EQ plug into a Space Designer plug. The Space Designer is set to 100 percent wet. Match EQ is set to whatever percentage your ears tell you is good. Now you have control over the amount of the filter spectrally, and the harmonics and distortions will be taken care of with Space Designer.

    The other option is to try it with a sweep and the IRU/Space Designer alone, but you are going to have a heck of a time re-creating all the possible settings of the filter. Space Designer is more of an all-on, 100 percent wet deal. Which means you would need a different preset for every possible permutation of the filter. You can do it, but it could be difficult to use.

    One thing is for sure though, you can't use noise with the IRU, it must be a sweep. Anyone using noise to capture the behavior of an outboard piece is using a form of EQ convolution like Match EQ, not reverb convolution, like Space Designer.
  • by EdwardRyan,

    EdwardRyan EdwardRyan Feb 20, 2008 4:41 AM in response to John Buehler
    Level 1 (0 points)
    Feb 20, 2008 4:41 AM in response to John Buehler
    I don't want to build the filter as a plugin setting. I want to build it as a completely new AU unit. The method you describe is a good way of doing it if you only want certain presets, but is too restrictive for what I need to do.

    Tones and sweeps are not good for sampling hardware eq and filters, since the resulting file is input frequency dependent. When I try to model the filter afterwards the response would only be true at the frequencies tested. And with sweeps there are other problems, as the phase of the signal will change with the increase of the frequency. I'm just trying to capture the behaviour of the hardware, not it's response to tonal input.
    With noise the energy is equal per hz (white) or per octave(pink), instead of being just a tone or a series of tones. This is why I asked for a noise generator in the I.R utility.

    Any more ideas? It's good to have someoen to talk to about this stuff. This forum seems a bit neglected !
  • by John Buehler,

    John Buehler John Buehler Feb 20, 2008 7:58 AM in response to EdwardRyan
    Level 2 (385 points)
    Feb 20, 2008 7:58 AM in response to EdwardRyan
    So you are designing a completely new AU plug?

    I never said use a tone, that would be dumb indeed. I disagree about the phase of the signal with a sweep. The phase is intact every time I've done it. I have swept many eq's, there are no phase issues. I have seen the signal on phase meters, compared it to the outboard unit, and there is no difference in the phase of the signal.

    +I'm just trying to capture the behaviour of the hardware, not it's response to tonal input.+

    This is confusing to me, isn't the behaviour of the hardware the same thing as it's response to tonal input? It's a filter, what else does it do but respond to tonal input? If you don't want to emulate the filter with plug-ins that already exist, what do you want? Raw data?

    If you really want to try to turn white noise into an impulse, you will have to use Space Designer in Logic 7, where you can use any signal you want to create the impulse, but it won't work, and the effort is only useful if you plan on using Space Designer. It's not like you can derive any useful information about the hardware unit after the process, you just get an impulse and a preset. But as I've said, white noise won't make an impulse.

    The IRU is not for doing analysis like T.E.F. or Spectrafoo, which is what it sounds like you need. It makes Space Designer presets.

    So, can you clarify what it is that you are doing? Are you actually trying to make your own AU plug?
  • by EdwardRyan,

    EdwardRyan EdwardRyan Feb 22, 2008 5:33 AM in response to John Buehler
    Level 1 (0 points)
    Feb 22, 2008 5:33 AM in response to John Buehler
    I will be doing the AU plugin later. For now I'm working on building the VST version. We're taught the VST framework because it's the more commercially popular. Just for the fact that there are more PC's out there then mac's.
    Do you mean a TEF 25? I've never used one of those before, but have read a lot about them. Good piece of kit.. only big problems are the fact it needs software to do the analysis, and without the power supply it will only supply 1/2 the normal phantom power. Oh and it doesn't support ASIO, but they were working on that last point.
    I've used SMAART and a Nuetrik to do testing before. I just need something to convolve the results.

    I know the IRU wasn't designed to be an in depth analyser, but it's a shame it doesn't do a little bit more.
  • by John Buehler,

    John Buehler John Buehler Mar 3, 2008 9:11 AM in response to EdwardRyan
    Level 2 (385 points)
    Mar 3, 2008 9:11 AM in response to EdwardRyan
    It seems like Waves Q-clone would do the trick, but I haven't used it. But it continuously monitors the hardware for changes, so it would capture the full spectrum of adjustments to the filter.