Bass pedal

[tallis]

Hey,

I'm a recording engineer and a bit of a newbie at audio electronics. Dabbled a bit but not much experience, but handy enough with a soldering iron.

I wondered if you guys would be willing to help me with a design for a bass pedal I want to build.  It's based on how I tend to go about mixing distorted bass guitar, which is a dual-band approach.  The bottom band is clean and compressed, the top band is just the top of a distorted, amped signal.

So effectively, the pedal would split the signal in two...

Signal 1 goes through an FX loop, where you can add distiortion, or anything else you like, which then comes back in and goes through a high pass filter (it's important the the HPF is after the distortion, so the distortion has to work on the full band).

Signal 2 goes through a low pass filter (filter comes first this time), and then through a compressor.

Signal path mapped out in the (hopefully) attached image.

A good old friend of mine has been helping with the research.  So far:

- This for the splitter; I'm wondering whether it's worth having variable input impedance or whether that's just going to be too much faf.
http://www.muzique.com/lab/splitter.htm

- Craig Anderton compressor (there might be an easier one to implement, but it needs to sound pretty smooth).

What I'm really struggling with is coming up with something for the filters. I'm thinking second-order RC filters around 150Hz, but I'm not sure how exactly to match them to prevent a bump or a hole at the stated frequency (just like a crossover network).  One solution, which might also add some flexibility, would be to have the two filters on pots and tune them by ear, but not sure how to implement this, and it does seem a bit Heath Robinson.

Anyway, it would be amazing to hear your thoughts.  I'll keep you informed on any progress.

Peace,
Tom.

[11-90-an]

Hello and welcome to the forum.. 

Is the compressor built in?

Maybe use one of those simple first order RC LPFs/HPFs...? (ya know, the one resistor one capacitor ones...)

[tallis]

Thanks for the welcome - glad to be here!  And thanks for the reponse.

Yes I'm hoping the compressor will be built-in.

...and yeah that's what I was thinking with the filters - nice and simple.  But as I understand it (might be wrong) if I just just set both of them to 150Hz exactly, there would be a bump at that frequency because both signals are producing that frequency.  So I'd have to move them slightly apart, say 145Hz and 155Hz.  But I assume there's a way of working that out with a bit more precision?

T

[marcelomd]

Check this out: http://www8.plala.or.jp/KandR/cir_bassthru.html. Almost what you want, has all the basic blocks, except the compressor, which is easily added.

[tallis]

That's really useful thank you.  I think I spotted something similar to that.

So this has a LPF on the non-effect-loop signal path.  Is there a way to "match" a HPF on the other signal path?

[Ben N]

YOu can do that with a state variable filter (SVF) with HP and LP outs. (https://www.analog.com/media/en/training-seminars/tutorials/MT-223.pdf)

[tallis]

Amazing, thanks for that Ben.

You might have to abide my ignorance a bit here....

So would that actually do the signal splitting for me as well as effectively providing the crossover function accurately?

If so, I don't think I'd be able to get the signal through the FX loop before it gets filtered right? (Every chance I've understood this wrong).

[marcelomd]

That's really useful thank you.  I think I spotted something similar to that.

So this has a LPF on the non-effect-loop signal path.  Is there a way to "match" a HPF on the other signal path?

If you want fixed frequencies, just use the same values.
If you want to sweep the frequencies, just use a state variable filter, as noted. I like this project: https://sound-au.com/project148.htm

[tallis]

Man yeah, that looks really cool.  Thanks!

So that has a single input that divides into two bands based on the position of the pot.  Is there a way to give it two different inputs, one of which ends up as the top band, and one of which ends up as the bottom band?

Tom.

[marcelomd]

Man yeah, that looks really cool.  Thanks!

So that has a single input that divides into two bands based on the position of the pot.  Is there a way to give it two different inputs, one of which ends up as the top band, and one of which ends up as the bottom band?

Tom.

If I understand correctly, isn't it a simple matter of applying a separate low/high pass filter to each input, then?

[tallis]

...well! This is something I haven't decided yet. I like the idea of a pot to tune the crossover point, but if I can get a flat response without the extra faf, I can live with a fixed value.

The text in your link mentions that Butterworth curves produce a 3dB bump at the x-over frequency, whereas his recommended Linkwitz-Riley curves give you a flat response within 0.5dB. Are standard rc filters Butterworth? In which case it might be worth the extra hassle after all...?

[Fancy Lime]

Hi Tom and welcome to the Nuthouse!

I have gotten into effects diy because I wanted a decent bass distortion and have designed and built my fair share of them since. Therefore, please permit me a few questions and musings. Slightly OT.

Are you sure you want or need a split path, log compressed, hog distorted, arrangement for the sound you want? That is a technique that is popular with sound engineers and is pretty easy to implement in a DAW or even with the analog gear sitting around in a well equipped studio. But doing this with diy analog electronics is a major undertaking if you want to have a really good result. You can get a very similar and in my opinion often better result from a simple single path distortion with some careful pre and post clipping tone filtering. Such a device would be much easier to develop and build. If you are fairly new to electronics diy, I would suggest exploring this road first before jumping in at the deep end. It may or may not get you the sound you seek but even if it doesn't, you can still use it for the distortion section of the device you proposed in the original post.

When mixing high and low passed paths for bass, I find first order filters provide a much better blending of the path s than second order filters. The result is more natural and more cohesive to me. Second order blending often sounds a bit disjointed to me, like to separate instruments playing in unison. While this can easily be tweaked in a DAW to sound more cohesive, making all the relevant parameters variable can be impracticable in analog electronics. I suggest to try first order mixing first, if you do go the mixing road.

Have you considered building only a parallel looper with adjustable hog and log, respectively? Having that as a standalone box in which to plug a distortion and compressor would be more flexible and easier to develop and debug than a massive all in one box.

Hope this helps a little although it does no answer any of the questions you actually asked,
Andy

[marcelomd]

...well! This is something I haven't decided yet. I like the idea of a pot to tune the crossover point, but if I can get a flat response without the extra faf, I can live with a fixed value.

The text in your link mentions that Butterworth curves produce a 3dB bump at the x-over frequency, whereas his recommended Linkwitz-Riley curves give you a flat response within 0.5dB. Are standard rc filters Butterworth? In which case it might be worth the extra hassle after all...?

Iiiiii... really don't know =)
I remember the difference between filter types being how the coefficients were calculated.

My suggestion, building on top of Andy's is to make the simplest thing that could possibly work and build from there. Buffer, 1st order, low pass, high pass, mixer, etc.

If you have the means, try to do it in software first, and see if you like it. Check https://neuraldsp.com/products/parallax for ideas.

A few years ago I was trying to emulate a biamped rig with a Boss LS2, GEB7, GE7 and LMB3 plus a few distortion pedals. It sounded horrible. A simple clean blend design, like the Darkglass B3K was much better. The B3K, in particular, sounds good even without the clean signal.

[tallis]

Hi, thanks for the replies, guys. And certainly thanks for sharing your experience. You're right that I'm probably spoilt with the ease of routing and manipulating signal paths inside the DAW. But I've got a lot more clarity now over what I'm actually aiming to achieve, so thanks!

OK I'll start again with the axiom that simpler is better.

My starting point is that whenever I record and mix distorted bass, I always take a clean DI as well. I've recorded loads of weird and wonderful flavours of distortion, but it always ends up sounding better to me when I use the mids/highs from a miked up cab, but replace the bottom end with a heavily compressed bottom end from the DI. You get the benefit of the richly harmonic and chaotic mid/highs with a really solid and consistent bottom end.

I've never had satisfactory result from just distorting the upper band. Has to be distorted full-range first and then filtered.

So maybe all I want is a routing pedal with filters. Two FX loops, one pre-HPF, and one post-LPF.

Thanks for the tip on using first order filters. I'd be interested to know if there a way to match high-pass and low-pass RC filters to avoid this 3dB hump when you give them the same values. Is there a mathematical approach, or will I just need to get a bunch of components and experiment? I guess I could mock up the filters in the DAW and see what response I get.

The state variable filter seems the most elegant option, but I'm not sure there's a way to get two separate inputs to feed the two filters. I wondered about using a stacked pot controling two state variable filters separately?

Again thanks so much. I'm gaining a lot of insight here!

[DIY Bass]

The Zorg Glorious Basstar doesn't have a compressor, but it is a 3 band bass fuzz in which each band can be clean or very dirty.  Can get it as a pedal, kit or pcb.  If you wanted a compressor you could probably work out where in the schematic to add a compressor to the bottom end.

https://www.zorgeffects.com/index.php/en/glorious-basstar-detail

[tallis]

Oh amazing!  Thanks.

[marcelomd]

Check the VFE Triumvirate. It's also a 3 band distortion. I knew I was forgetting something.

[Fancy Lime]

Hi, thanks for the replies, guys. And certainly thanks for sharing your experience. You're right that I'm probably spoilt with the ease of routing and manipulating signal paths inside the DAW. But I've got a lot more clarity now over what I'm actually aiming to achieve, so thanks!

OK I'll start again with the axiom that simpler is better.

My starting point is that whenever I record and mix distorted bass, I always take a clean DI as well. I've recorded loads of weird and wonderful flavours of distortion, but it always ends up sounding better to me when I use the mids/highs from a miked up cab, but replace the bottom end with a heavily compressed bottom end from the DI. You get the benefit of the richly harmonic and chaotic mid/highs with a really solid and consistent bottom end.

I've never had satisfactory result from just distorting the upper band. Has to be distorted full-range first and then filtered.

So maybe all I want is a routing pedal with filters. Two FX loops, one pre-HPF, and one post-LPF.

You are comparing apples and oranges here, to  some degree. Or rather, you are trying to recreate an orange that tastes like an apple, when you could just have an apple instead. OK, I realize that this was not helpful. Let me try and explain. Your experience comes from miking cabs for the distorted part. The low frequencies from any cab through any microphone in any room always suffer from all sorts of internal and external self interference, which creates booming of certain frequencies and "holes in the spectrum" at other frequencies. A really good engineer with really good equipment in a really good room can mitigate this but never fully avoid it. Physics is physics, no matter what. This is, in my opinion, one of the most important reasons why adding clean compressed DI-bass and cutting bass from the cab is often a big improvement. However, if you do you processing electronically and nothing ever gets miked and cannot resonate in a cab or room, you don't have that problem. You can therefore avoid some of the hoops you have to jump through with the miked cab situation.

Thanks for the tip on using first order filters. I'd be interested to know if there a way to match high-pass and low-pass RC filters to avoid this 3dB hump when you give them the same values. Is there a mathematical approach, or will I just need to get a bunch of components and experiment? I guess I could mock up the filters in the DAW and see what response I get.

The state variable filter seems the most elegant option, but I'm not sure there's a way to get two separate inputs to feed the two filters. I wondered about using a stacked pot controling two state variable filters separately?

Again thanks so much. I'm gaining a lot of insight here!

Choosing the frequencies right to avoid "the hump" is not the big issue here. The big issue when mixing two signals from the same original source that then went through different paths, is phase. The phase of the signal gets shifted around a lot in things like compressors and distortions and also in the crossover filters. The problem is that the phase shift is frequency dependent. Trying to reconcile the phases of both paths is pretty much hopeless and depending on how the phases are at the crossover frequency, you may get constructive or destructive interference, creating a hump or a dip, respectively. Phase at this frequency may also change with the action of the compressor or the gain setting of the distortion. The same thing is not as much of a problem for the miked cab situation because phase from the mike is an absolute mess anyway. The phase shifts across the frequency spectrum there are likely so numerous that the dips and humps in the combined spectrum are too small to matter. But when doing the same thing with electronics, they do because the phase is still twisted around across the spectrum but only a few times, so the humps and dips by interference are much more prominent. TL;DR: If you want to do the whole dual path thing, tuning the crossover by ear is the only way. I would have separate pots for the low pass and high pass frequency. That way, you can fine tune the crossover and take care of any phase related dips or humps.

The reason why I think you can probably get away with simply clipping all frequencies if you get the frequency shaping right has to do with what clipping really is. Clipping is a compression of the wave form. On the side of dynamics, that leads to the effect we also know as a compressor. Well, duh. On the side of frequency spectrum, it leads to overtones. By filtering out all the treble after the clipping, you can keep the compression effect but discard the extra overtones again. But you want some overtones because you want a distorted sound, right? But it is not supposed to sound muddy, right? The trick to that, in my experience, is to filter out the right overtones only. Mud lives in the lower mids. Cut those a bit and see what happens. I suggest the following experiment: In your DAW, run a bass track through a nice fat distortion plugin, maybe a Rat or Big Muff type of thing. Or a simple DOD 250. After that, use a single parametric EQ band to cut the lower mids. Play around with all parameters and see if you can get the sound you are looking for. It's not the same as a dual path with a miked cab but it may give you an indication if this much simpler approach might be worth persueing in the analog realm. I suggest you also simulate your whole dual path approach with clipping and compressor within the DAW for comparison. This should make it easier to decide, which way to go forward. I'm curious about the findings.

If I were to design a "studio bass overdrive", here is how I would start:
1. Input section with (variable) treble boost and variable gain
2. Clipping stage (can be incorporated in input stage or come thereafter) with gradual transition into clipping (soft knee) and frequency selective clipping (less clipping of lower frequencies)
3. Fully parametric singe band mid EQ, implemented with a state variable filter
4. Baxandall 2-band tone stack
5. Second order low pass filter at 5-7kHz

Cheers,
Andy

[11-90-an]

Bah! Ya'll giving such complicated stuff! 

One of my favorite fuzzes/distortion is the bazz fuzz:
http://www.home-wrecker.com/bazz.html

Feel free to tweak about everything about it 

[Ben N]

Now we have a lime lecturing us on apples and oranges. 

You are comparing apples and oranges here, to  some degree. Or rather, you are trying to recreate an orange that tastes like an apple, when you could just have an apple instead.