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

Need help to design a Line level coloration unit

Discussion in 'Electronic Repair Modifications DIY' started by JakeAC5253, Feb 23, 2014.

  1. JakeAC5253

    JakeAC5253 Active Member

    I want to make a line level unit that I can mount within a rack space and use it to process recorded audio when it needs a little something or other. I'm going to have the unit split the signal into three frequency bands bass, mids, highs. Then I might process each section differently and then they will get combined at some later point in the circuit to be run into some kind of harmonic saturator to add some umph which I haven't decided on yet. I need some advice.

    -Since the input is line level, you shouldn't need an input buffer, but considering that there will be three passive taps from the same signal, I think it would be better to have a buffer just so you don't cause too much loading on the input signal, right?

    -What sort of resistor values should I use for the signal splits so each filter doesn't load down the next and cause a mess of unintentional cutoff frequencies and other problems? Maybe I need a different topology altogether? Active bandpasses?

    -Any practicality in using an input transformer in the unit, like those classic mic preamp designs that sounded great and used input transformers?

    -How should the signals be combined? It's easy enough to combine two signals and have them be variable a-la Big Muff tone stack, but the presence of the third signal complicates the design greatly.

    -Any suggestions for a mild but flavorful harmonic saturator circuit? Vintage and/or tape sounding would be desired. I'm thinking two color gain stages cascaded and installed in a blend circuit.
  2. DonnyThompson

    DonnyThompson Distinguished Member

    I'm certainly not the guy to ask when it comes to resistors, buffers and splits, as I'm no electronics wiz..

    Well, perhaps if you could be a bit more detailed as to what you want to accomplish? You say you plan on splitting the signal into three bands, which I'm assuming would implement some kind of crossover circuitry, but at that point, what do you want to be able to do with the signal(s) from there? Is this something that will be more of a forensic-type of EQ? Or something more along the lines of signal saturation? When you mentioned the Big Muff as a reference, that is more of a gain thing (although I guess if we got right down to it, everything relies on gain processing of some kind or another and at some point, right? LOL)

    When you say "coloration", that's a pretty wide subject area, Jake.

    The term "color" is generally accepted by most engineers to describe tone/EQ. But, it can also encompass anything and everything from Gain Reduction to Spatial Enhancement / Panning to more obvious effect-based processing like Reverb/Delay/Chorus, etc...

    Is this processing that you would like to aim at vocals... or, more of an instrumentation thing? Is it something that would target processing on a track level, or would it be designed more for 2-bus/master output strapping? Mono or Stereo?

    Are you looking at Solid State? Tube? Class A? EQ? Gain Reduction/Expansion? etc., there are so many possibilities...

    Just curious...
  3. Boswell

    Boswell Moderator Distinguished Member

    It shouldn't be too difficult to design and build a 1U rack-mount that does the frequency splitting, has insert jacks on each split for you plug in whatever processing you want in that frequency band and then a summing amp on the output. This would be similar to the operation of a multi-band compressor but without the actual dynamic processing in the box.

    Several questions though before you can get down to details:

    1. Is this unit just mono, or would you want to be able to process stereo signals with it?

    2. Can you fix the frequency bands at design time, or do they need to be variable and set by front-panel knobs?

    3. Are the frequency bands non-overlapping, so that, for instance, the frequency of the upper edge of the lowest band is always the same as the lower edge of the middle band? (I'm referring here to the -3dB points, so the slopes of each band would intersect)

    4. Do the inputs and outputs have to be +4dBu balanced?

    5. Do you need metering?
  4. JakeAC5253

    JakeAC5253 Active Member

    Anything I want really. I guess I'll just play around with what I hear and add circuitry until it sounds good enough. I'm open for suggestions for fun circuitry to add to any three of the bands.

    That could be one of its uses. Dial the bass signal up and it gets warmer, dial the highs up and it gets clearer. The individual bands will have some user-controllable processing going on that will change the sound of those ranges, maybe saturation or something like a hard/soft knob.

    I just meant the Muff tone stack topology. It splits the distorted signal into lows and highs and combines them at either end of the tone knob which allows you to blend their relative balance levels. Same concept with my unit, except that my unit has the middle control as well, which complicates matters.

    Yeah, I mean color as in like the difference between clean digital and a tape machine pushed into the red. One is transparent, the other is adding something special.

    I'm thinking more source related, so mono input, and can possibly be used on clean/acoustic guitars, vocals, or bass or anything that needs a little something.

    Solid state hopefully as it would simplify the design greatly, but whatever works. I've got a respectable stock of JFET amps that I bought when they were going out a couple years ago.

    Yes, exactly the idea.

    Mono preferably, I can see limited uses for needing it to be stereo, and that would just make the design 3x more complicated.

    As of now, I will fix the bands during the design phase, and then just use the level balance knobs to tailor the tone balance.

    Well I assume you are asking if I need one knob to control the cutoff point for two bands at once, so considering they will be hard designed, I don't think so. I would like the bands to be relatively isolated, but considering the end result of this endeavor is good sound, and not a perfect spec sheet, some overlap certainly does not bother me. I still do need to decide how steep the cutoffs will be. I'm familiar with how -6dB/oct filters work in and out, but I vaguely know how -12dB/oct filters work (I know they require a gain stage). Any slope greater than that I don't know about, and how to accomplish what I want efficiently and without signal loading is currently unknown to me. I looked up some BMT circuits and found that it looks quite a bit more complicated and deliberate than I had guessed. I'll definitely need help with that portion.

    I'm not sure how to answer this effectively. I'll just be taking the output from my audio interface via TRS and plugging it into the unit and taking it back to the interface via TRS.

    Eh, not really. It would be a nice feature, but when I imagine metering, I see the price of the unit going up and up and I don't really see much of a use for it.
  5. paulears

    paulears Well-Known Member

    Isn't the real issue going to be recombining them? The processing you are going to do could well introduce artefacts and maybe even polarity swaps, so when you recombine them very odd things might happen.
  6. Boswell

    Boswell Moderator Distinguished Member

    OK, so where are you going to go from here? I'm talking about after you have fixed the gas pump display, of course.
  7. JakeAC5253

    JakeAC5253 Active Member

    Good point. I think it should be easy enough to perform a unity gain polarity flip though. That I'm not worried about really.

    I don't really see how that relates at all. After I design and make this audio unit, I'll use it, and after I fix the fuel dispenser... I'll fix more of them. :cautious:
  8. Boswell

    Boswell Moderator Distinguished Member

    Sorry, I was being a bit light-hearted.

    I don't think you should worry overly about the re-combination of processed frequency bands. The filtering needed to split the bands in the first place will generate a complex phase-frequency relationship, and then this will be convolved with that of the signals going through any external processing. The final summation cannot generate false harmonics, as it is an addition process, not a multiplication or modulation.

    You could think of it as sending signals from multiple microphones each covering a different frequency band into an analog mixer, or as a 3-way loudspeaker crossover network in reverse. It would be straighforward enough to design the frequency splitters in your box to be Linkwitz-Riley types to achieve smooth amplitude and phase characteristics through the crossover ranges, but given that you will be doing unknown things externally to the splits in both phase and amplitude, it scarcely seems worth the extra effort and components to do this.

    It's a great idea for a project, and I would get down to drawing out your specification for the splitter box and then we can assist further as necessary with the design and implementation.

Share This Page