Just scored a Rane PE-17 Parametric EQ in a trade.

[Flux Jetson]

After having used this thing for about 6 or 7 weeks now, I can say that it is easily the finest parametric EQ I've ever owned.

I've got my eye on another one, this time a Rane PE-15 (five bands, all of them arranged in series instead of the 3-parallel/in series with/2-parallel setup that the PE-17 uses).

Hopefully the PE-15 will deliver as well as the PE-17 has. The PE-17 is easily the best sub-thousand dollar EQ I've ever used.

[walterw]

+1 to rane being an awesome price score for high-quality used signal gear.

i have a 15 languishing in my closet, which i retired for a 17, both of which were/are used as the main EQ for my band's PA instead of the typical 31-band graphic (using one on bass seems like wildly excessive overkill, unless it was used to fight feedback on an upright).

the 17's high-pass slider comes in handy because my low end amp doesn't have a sub filter, but more to the point, using a para instead of a graphic as the FOH EQ forces you to tweak with your ears to fix specific issues instead of just shaping slider curves by eye.

in fact, the best thing about the unit may be how often i run it with most if not all 5 bands bypassed!

first i've heard of the 3 into 2 layout, and that confuses me a little; with 3 EQ bands in parallel, if i notched out a frequency on one band, wouldn't the same signal still slip through the other two unchanged, thus negating the effect?

[synaesthesia]

Congrats on the acquisition. I have one, it's good, and the hi pass is useful in saving your speakers and conserving your power amp's watts, allowing it to feed useful watts in the useful audible range. The low pass is also useful as a universal tone filter, tune around 3.2Khz or so and you get instant old school sound. Mine sits in a spare rack, I use a Klark Teknik DN405 mainly, similar but balanced in/outs only and a bit quieter in noise levels if quieter matters.

[Flux Jetson]

Originally Posted by walterw

....i have a 15 languishing in my closet, which i retired for a 17, both of which were/are used as the main EQ for my band's PA instead of the typical 31-band graphic (using one on bass seems like wildly excessive overkill, unless it was used to fight feedback on an upright).

I suppose we all have our own ideas about what ~good~ is, y'know?

Originally Posted by walterw

....the 17's high-pass slider comes in handy because my low end amp doesn't have a sub filter, but more to the point, using a para instead of a graphic as the FOH EQ forces you to tweak with your ears to fix specific issues instead of just shaping slider curves by eye.

Couldn't agree more!

Originally Posted by walterw

....in fact, the best thing about the unit may be how often i run it with most if not all 5 bands bypassed!

Even if all you use are the LPF and HPF, good state variable filters in rack format are not exactly in abundance. My modular synth filters offer those filter states, but a pair of them (one used in HPF state and the other used in LPF state) plus the power supply and preamp module would run in excess of $600 when the rack mount rails and power supply are included. So even if you only use the PE-17 as the LPF and/or HPF they're still wonderfully musical units. As your post points out.

Originally Posted by walterw

....first i've heard of the 3 into 2 layout, and that confuses me a little; with 3 EQ bands in parallel, if i notched out a frequency on one band, wouldn't the same signal still slip through the other two unchanged, thus negating the effect?

No, not at all. I suppose it all depends on how you use it. The filters aren't like resistors that are bound to laws of physics in that the sound will take the path of least resistance. You use the parallel filters as a mix, you mix-in (or out) each individual filter as needed. If they were all in series. if you set one up as a HPF and the other as a LPF, if the cutoffs were in the right places you'd have no sound, just silence. In parallel the two may be mixed together to form a notch filter (for example).

It also depends on how loud you have each band in their mix, bands one through three may be mixed at will without affecting each other. Same with bands four and five.

I use them like presets. I'll set up three different cuts/boosts on bands 1, 2, and 3 along with two other complementary eqs on bands 4 and 5, switching them in/out for totally different sounds/shapes. I can go from a deeply scooped modern tone to a 70's rock bass mid-hump tone just that fast by bypassing or engaging various bands using their individual bypass buttons. It's quite handy once you get the gist of the way it can be used.

Here's Rane's own explanation of the parallel/series configuration of the PE-17 including a wonderfully useful block diagram ....

http://www.rane.com/pdf/old/pe17dat2.pdf

So the PE-17 has two shelving EQs (LPF and HPF) and FIVE fully parametric bands that have 100% overlap, kinda like having five of the exact same peaking parametric EQs that all have the exact same freq ranges (ten-hz to twenty-khz) that are arranged in 3/para x 2/series.

The PE-15 has five parametric bands, with bands one and five being switchable from peaking to shelving, all five of which are arranged in series. So it may be set up as a "bass - mid - treble" EQ with fully adjustable freqs on the "bass and treble" bands, and three fully parametric "midrange" bands with deeply overlapping freq ranges. Their overlap is so deep that most midrange frequencies may be accessed by three bands.

The PE-15 may be easier for the casual filter user to use and understand. For filter ~freqs~ (ahem) such as myself the PE-17 is a nice treat.

I'm hoping to collect at least two PE-15s and one more PE-17 for my system. Keeping in mind that my setup is used for bass, guitar, and modular synth.

If you're into techy reading, Rane's Technical Library is a phenominal resource.

http://www.rane.com/library.html

My ~dream~ filter would essentially be a modular synth format design that mimmicks the individual Rane PE-17 parametric band archetecture. That way I could arrange them in any parallel or series or series/parallel manner I please, or place individual filters at any point in the signal chain. One of the filters in my modular synth has a band pass state available, but it has a fixed rolloff of 6db (single pole). It's also a bit noisey.



I normalized the filter's outputs to a mixer to provide patchcordless mixing of all four states to create a filter of any type by mixing the states together. Any of the normalized connections may be instantly overridden by inserting a patchcord into any given state's output, through the use of normally closed switching jacks I installed on the filter itself which are hardwired to the filter mixer's input channels.



So that's close, but not quite.

A nice 2nd choice might be a PE-17 with individual I/O jacks for each band, using normally closed switching jacks so when the jacks are not used the unit would default to Rane's orginal parallel/series/parallel arrangement.

[Flux Jetson]

Originally Posted by synaesthesia

Congrats on the acquisition. I have one, it's good, and the hi pass is useful in saving your speakers and conserving your power amp's watts, allowing it to feed useful watts in the useful audible range. The low pass is also useful as a universal tone filter, tune around 3.2Khz or so and you get instant old school sound. Mine sits in a spare rack, I use a Klark Teknik DN405 mainly, similar but balanced in/outs only and a bit quieter in noise levels if quieter matters.

Cool, I'll have to look in to the Klark Teknik DN405 .. noise ALWAYS matters!

[Passinwind]

Originally Posted by Flux Jetson

Cool, I'll have to look in to the Klark Teknik DN405 .. noise ALWAYS matters!

I narrowly missed snagging one of those K-Ts a couple of years ago. Another nice option is an old Orban 642 like this one:



This is a two X 4 band format, switchable to one X 8. Each parametric band is on its own daughter card and there is plenty of room on the back panel for hacking an I/O matrix. There's, ahem, one for sale in the Effects Classifieds at the moment.

[walterw]

i'm still not getting the "parallel" thing;

i use mine almost entirely for cutting rather than boosting (as befits FOH EQ use), so how can i knock out a feedback frequency with one band if it's always still passing through at least one other parallel band unmolested? wouldn't i always have to dial up say the last two bands to notch out the exact same frequency if i wanted to have any effect on it?

does this mean that the last two bands are more effective individually for cutting, since they're only being bypassed by one other band instead of two? if i knock something down the full -15dB with one band and it's still passing through at 0dB on the other (parallel) band, does that mean i'm only really cutting it by half that?

[Flux Jetson]

Originally Posted by walterw

i'm still not getting the "parallel" thing;

i use mine almost entirely for cutting rather than boosting (as befits FOH EQ use), so how can i knock out a feedback frequency with one band if it's always still passing through at least one other parallel band unmolested? wouldn't i always have to dial up say the last two bands to notch out the exact same frequency if i wanted to have any effect on it?

does this mean that the last two bands are more effective individually for cutting, since they're only being bypassed by one other band instead of two? if i knock something down the full -15dB with one band and it's still passing through at 0dB on the other (parallel) band, does that mean i'm only really cutting it by half that?

If what you're thinking was true, then mixers would not work. And we know that mixers work, right? Your last sentence is getting you closer to full grasp. If you want full effect, bypass the bands you're not using. Otherwise the band you're using becomes a percentage of effectiveness - how much percentage depends on how you have the other engaged parallel bands set. How loud they are, what freq they're dialed in at, how wide/narrow their bandwidths .. aka "Q" .. are set. Just like a mix. If you only have one mixer channel used then the entirety of the sound coming out of the mixer's output(s) is made up of just that one mixer channel. If you have other mixer channels going into the mix, then the output is made up of all of the channels, with their relative levels mixed into the total output.

If you ran one source ... let's say a bass .... through one channel of a mixer, and tweaked it at 80hz, then the mixer's output would be nothing but that bass with the 80hz tweak. Now let's run that same bass into THREE channels of the mixer, with only one channel that hasw the 80hz tweak. The output of the mixer would depend on the levels of all three channels, how much emphasis the 80hz tweaked channel has fully depends on how much of it is mixed in relative to the other two channels.

This is a greatly oversimplified metaphor of how the Rane PE-17 works, but it might help you wrap your head around it better. Basically if you don't want other parallel channels to have any influence on the sound you need to bypass them. Parallel processing allows you to create complex "mixes" of a single EQ'd signal. It permits you to create very exotic/complex filter shapes that are otherwise unattainable with series filters. Series filters still definitely have their value, but parallel filters are able to create complex very useful, very precise, and very musical EQ curves.

For instance, if you used one band to create a wide scoop in the midrange (like a wide ~smile~ shape), and another parallel band to create a narrow/tall "hump" in the middle of that wide scoop, you'd create a very defined "W" shaped EQ curve. The wide scoop would create the two outter "arms" that reach upward of the "W", and the sharp boost in the middle of that scoop created by the other band would create the center point of the "W". You may even move the center spike towards lower or higher freqs, creating totally different sounds by doinf so. If you used series filters to do the same thing, it would take FIVE of them to create that same shape (three boosts, and two scoops ... one boost for each upward peak - total of three, as well as needing to adjust their slopes, and one notch for each of the two dips in the "W", again with the ability to adjust their slopes as well. What a headache, not to mention added noise from all of that processing!).

If you tried to do the same "W" trick with two series filters, if you had the smile first, that is all you would get. Why? Because once you scoop out all of those middle freqs to create the smile, there's nothing left to "boost" in the middle when you send that smile signal to the next filter in line. It's already been hacked out of the signal, so there's nothing for the next filter to boost in the center of the smile to create the "W" shape we're after. In parallel, you simply create a smile with one band, and a peak with the other band, then mix them together .. creating the "W". Exactly the same way a mixer works.

The same thing with creating defined notch curves. Use one parallel band to create a HPF, and another band to create a LPF. Now sweep the freqs of the two bands to fine tune the notch. That way you can create a notch that has a different low end slope than the high end slope. That is impossible to do with series filters. Why? Think about it ..... if you used one series filter as a HPF to cut everything below 3khz, and another series filter as a LPF to cut everthing above 1khz, the result would be silence (the HPF cut off everything below 3khz, and the LPF cut off everthing above 1khz .... everything has been cut out!). There would be no sound at all. But if you do the same thing with the same filters in parallel, the result would be a ~scoop~ that spans between 1khz and 3khz. The HPF leaves everything above 3khz, and the LPF leaves everything below 1khz, the 1khz to 3khz region is missing, creating a scoop or notch filter curve when you mix the two processed signals back together after they've been filtered. I know this seems weird but it is how it works. It may take some thinkin' time.

You may want to actually get out a pencil and paper (remember that stuff?) and actually draw some of these things out to get your head wrapped around them. I do that a LOT when something like this evades me (which is more often than I'd care to admit).

[walterw]

so the key is bypassing everything i'm not using? a bypassed band is passing no sound, as opposed to an unaffected sound? (mixers i am familiar with, so your analogy is well-taken).

i bypass unused bands on principle, but i guess it's critical with parallel bands like this.

i should check it again at my next gig (it stays in the other guy's van with the rest of the PA rack) to see what happens when i bypass all the bands, but not the overall bypass; i should get no sound, right?

now that i think about it, i've fairly often run it with the last two bypassed; shouldn't that be giving me silence? or is there some sort of switching in there that "straight-wires" it if all the bands in one parallel bank are bypassed?

[Flux Jetson]

No, they're not ~mutes~, the bypass switches do exactly that .... bypass the EQ circuitry returning that band to "flat".

Nearly any/every EQ made is a parallel circuit. When you turn the tone control of your passive bass towards a darker sound, you're routing more of the bass's signal through the parallel tone circuit which rolls off the high end. When you turn the tone control back towards brighter sounds, the control routes more of the sound around the tone circuitry, effectively bypassing it.

My guess is that Rane divided up the signal in thirds. However each band seems to also be buffered because they do not change the overall volume of the signal (unless you have it tweaked to severely reduce the level).

All you really need to know is what your ears are telling you. I know that my Rane PE-17 can create some VERY powerful sounds, even while using just one band along with the two shelving EQs that do their chores before the signal is routed to the parametric bands. Whether each band is only technically affecting only "one third" of signal or not, it sounds as tweaked as I want it to. I never want for "more" of anything. However, if I did wnat more (let's say) 800hz boost, it's simple enough to set up one of the series bands to the same settings which "add" their ability to the parallel group before it. Said another way, if I boosted 800hz up 12db on band 1, and I wanted even more 800hz boost, simply set up band four (one of th etwo bands that are in SERIES with bands 1,2, and 3) at the 800hz freq and add more boost to that freq. If you were to set the second series band to 12db of boost at 800hz, it would create a total increase of 24db of boost at 800hz. The series bands can "stack" settings on top of the group of parallel bands that are routed before them.

So if it seems like only some of your signal would cancel out whatever you've set up on one of the parallel bands, it doesn't .. not by a LONG SHOT. Sometimes you just gotta accept the fact that certain things work. Until you get more information into your head you won't EVER be able to sort out the ~whys~. To fully answer your inquiry, I'd need not only a full-on schematic, but operating levels within the layouts as well. I can't tell you how much of the signal is run through each parallel band "by percentage".

But I CAN tell you that it works FAMOUSLY and I'm never wanting for more of anything.

Without Rane's IP and theory of operation of the PE-17 I can't tell you how much signal is routed through the parallel bends in terms of percentage. But I am able to testify how well the danged thing works!

I also need to correct the hell out of myself on one little detail .... the PE-15 is not .. NOT .. fully series ..... it is fully PARALLEL. I was thinking parallel but saying (writing) series. Screwball, I.

[walterw]

Originally Posted by Flux Jetson

I also need to correct the hell out of myself on one little detail .... the PE-15 is not .. NOT .. fully series ..... it is fully PARALLEL. I was thinking parallel but saying (writing) series. Screwball, I.

OK, 2+2 is coming out to 7 for me here, so i am obviously missing a key detail.

with 5 bands in parallel, if it worked like i'm imagining it would, then turning a certain frequency range in one band down would have almost no effect at all!

it sounds like you're using this thing to boost for extreme and specific tone-shaping, so even if these bands worked like i'm imagining in my confused state, they would still have a strong effect; it's the cutting that i can't see how it would work if it was being mixed with other, uncut versions of the original signal.

[Flux Jetson]

Haahaa! Ok, either I'm a bad teacher, or your a bad student, or a combination of both of us being too unintelligent to communicate with each other! Haahaa!

Hmm .. let me try it this way since the mixer analogy seems to sortof work here.

OK, imagine having one signal split into (let's say) five mixer channels. Like taking your bass's signal into a five-way active/buffered splitter so that all five "splits" have exactly equal levels and impedances. Now send each of the five equal splits to separate mixer inputs.

Now, on the mixer, set all five channels with their faders at midway points ... we'll call that configuration set up "zero cut - zero boost" (aka "flat"). Now tune each one of the five channels to different EQ settings. You either make each one of the five channels louder (boosting freq), or quieter (cutting freq). The mixer's mono output is a sum of the entire mix of all five channels.

Think of a regular song that has five channels. If you reduce just the drums (all on one channel), that creates a pretty drastic change. If you boost just the guitar, that also has a pretty drastic change in the overall mix of the song. Same with the bass channel, boost or cut it and the entire mix sounds a LOT different.

The PE-17 is like having two of these splitter/mixer things, one running into the next. One of those splitter/mixers that has 3 channels, and a second splitter/mixer that has two channels. The 3-channel splitter/mixer's output is sent into the 2-channel spliter/mixer's input. So you start with one signal, split it three ways - (FX) - mix it down to one signal again. Then split that one signal again, but this time into two channels - (FX) - mix it down again to one output. Start with one signal, end with one signal

The PE-15 is one of these splitter/mixers that splits the original signal into five separate and equal channels - (FX) - remix all five down to one.

That's about as simple as I can break it down. It's more complicated than all of that, but that is about the most simple visual image I can draw.

So, was that of any help?

[walterw]

your analogy is crystal-clear (and appreciated ) but is what's making me confused!

Originally Posted by Flux Jetson

OK, imagine having one signal split into (let's say) five mixer channels. Like taking your bass's signal into a five-way active/buffered splitter so that all five "splits" have exactly equal levels and impedances. Now send each of the five equal splits to separate mixer inputs.

Now, on the mixer, set all five channels with their faders at midway points ... we'll call that configuration sets up "zero cut - zero boost" (aka "flat").

right, that's what i'm understanding; what i'm not understanding is that if i really did that, with my signal split into five duplicate channels on a mixing board all run flat, and i started hearing feedback at 400Hz, i would be helpless to fix it!

if i grabbed the mid control one one of those channels, spun it to 400Hz and dropped it by 10dB, nothing would happen, i'd still be getting that feedback and now would be getting dirty looks from the rest of the band

[Passinwind]

Walter, you might want to take a look at the Symetrix 551 PEQ manual, which shows the different interactions of series or parallel configs.

[walterw]

thanks, and i just did, but it doesn't really address my question; they showed overlapping boosts, and how that differs with adjacent bands vs non-adjacent bands; i assume the adjacent ones are in series, but they don't specify.

they also saw no need to warn about the problem i'm imagining, so there's gotta be a reason it's not a problem!

could it be that these bands (using the all-parallel PE-15 as an example) actually cut way deeper than -15dB, and that the -15dB spec is just the result of the other 4 bands passing duplicate flat signal in parallel, reducing the overall cut?

[Passinwind]

Originally Posted by walterw

what i'm not understanding is that if i really did that, with my signal split into five duplicate channels on a mixing board all run flat, and i started hearing feedback at 400Hz, i would be helpless to fix it!

if i grabbed the mid control one one of those channels, spun it to 400Hz and dropped it by 10dB, nothing would happen, i'd still be getting that feedback and now would be getting dirty looks from the rest of the band

Do you agree that if you just turned one channel fader down all the way the feedback would stop, if it were just on the verge of ringing initially? If you turned down 400Hz by 10dB on three or four channels would you agree that the feedback would likely stop? Haven't you actually seen that happen, only with separate mics on each of those channels? It's just a question of granularity, but at some point just turning down 400 on one channel will actually kill feedback in your scenario, and at some other point it no longer will.

[Passinwind]

Originally Posted by walterw

could it be that these bands (using the all-parallel PE-15 as an example) actually cut way deeper than -15dB, and that the -15dB spec is just the result of the other 4 bands passing duplicate flat signal in parallel, reducing the overall cut?

Yep, that's what I was getting at in my last post. If you have decent recording software you can make up some composite curves fairly easily.

[walterw]

clarity has been reached; thank you gentlemen.

[Flux Jetson]

Thanks for jumping in Grampa .... I was at my ends trying to find the right words. There is a reason why I'm not a teacher or an instructor! Sometimes I just do not possess the ability to paint the proper image.

[walterw]

flux, your "images" were perfectly clear

i was thinking that the actual circuitry in each band by itself only truly had -15dB of cut, which would end up being squat when summed with all the other parallel bands that weren't cutting at that spot. if the actual cut is way deeper, and just ends up at -15dB after the summing, well that makes sense to me.