Hi guys, new reg here, I am hoping someone will have some knowledge I can use on this one! Googling has produced some confusing results! Basically, I am building a guitar amp using a ~45W class D amp from ebay, combined with a 6j1 tube pre which I will be using as a double boost since it is a stereo unit.
I plan to send the signal through one side of the 6j1 first, then into an EQ/tone stack selectable with a switch (more on that in a minute) and then back into the tube pre on the other channel and out to the digi power amp to a 100W 8 inch PA speaker I purchased (the daichi IS8000). My question relates to the EQ, as its purpose is basically to change the frequency response of the speaker to closer match a greenback. I have looked closely at the charts for both speakers, and it seems that I can get roughly close to it with this setup on a 10 band fixed EQ:
60hz cut 100hz range (20-120hz)(.6q) 10db cut
600hz boost 600hz range (300-900hz)(1q) 3db boost
2250hz boost 1k range (1750-2750hz)(2.25q) 8db boost
2250hz boost 400hz range (2050-2450hz)(5.625q) 5db boost
4k boost 2k range (3k-5k)(2q) 13db boost
7k boost 1k range (6.5k-7.5k)(7q) 6db boost
10k boost 4k range (8k-12k)(2.5q) 8db boost
10500hz boost 2k range (9500-11500hz)(5.25q) 3db boost
16k boost 2k range (15k-17k)(8q) 15db boost
I have looked around for a customisable EQ circuit and it seems that they are basically just a bunch of band pass filters with boost or cut via the slider pots. What I want to do is fix the EQ with resistors so that I get a fixed EQ at those frequencies and Q values to closer match the sound of that elusive greenback.
I have no idea if this will work or work well, but I am willing to give it a go! I have found some calculators regarding the frequency/resistor/cap relationship but I am having some trouble figuring out the parts to use to get the Q values and frequencies I am after. Any advice would be greatly appreciated. Also, would I be better off going with an op amp active circuit? Or would the fact that I have the pre amp mean that I am better off going with a passive circuit, boosting the whole thing by the biggest boost value and then cutting back each frequency by X until I get the desired result? This is my first time dabbling in this stuff, though I have some electronics experience, this is a bit more complicated than most of the stuff I have built in the past!
Thanks in advance.
I plan to send the signal through one side of the 6j1 first, then into an EQ/tone stack selectable with a switch (more on that in a minute) and then back into the tube pre on the other channel and out to the digi power amp to a 100W 8 inch PA speaker I purchased (the daichi IS8000). My question relates to the EQ, as its purpose is basically to change the frequency response of the speaker to closer match a greenback. I have looked closely at the charts for both speakers, and it seems that I can get roughly close to it with this setup on a 10 band fixed EQ:
60hz cut 100hz range (20-120hz)(.6q) 10db cut
600hz boost 600hz range (300-900hz)(1q) 3db boost
2250hz boost 1k range (1750-2750hz)(2.25q) 8db boost
2250hz boost 400hz range (2050-2450hz)(5.625q) 5db boost
4k boost 2k range (3k-5k)(2q) 13db boost
7k boost 1k range (6.5k-7.5k)(7q) 6db boost
10k boost 4k range (8k-12k)(2.5q) 8db boost
10500hz boost 2k range (9500-11500hz)(5.25q) 3db boost
16k boost 2k range (15k-17k)(8q) 15db boost
I have looked around for a customisable EQ circuit and it seems that they are basically just a bunch of band pass filters with boost or cut via the slider pots. What I want to do is fix the EQ with resistors so that I get a fixed EQ at those frequencies and Q values to closer match the sound of that elusive greenback.
I have no idea if this will work or work well, but I am willing to give it a go! I have found some calculators regarding the frequency/resistor/cap relationship but I am having some trouble figuring out the parts to use to get the Q values and frequencies I am after. Any advice would be greatly appreciated. Also, would I be better off going with an op amp active circuit? Or would the fact that I have the pre amp mean that I am better off going with a passive circuit, boosting the whole thing by the biggest boost value and then cutting back each frequency by X until I get the desired result? This is my first time dabbling in this stuff, though I have some electronics experience, this is a bit more complicated than most of the stuff I have built in the past!
Thanks in advance.
Kind of low SPL of 90 dB, don't expect to rock your sox off. Using IC's would probably be easier, I am not sure just eqing a response will get the greenback sound.
Something is very odd here - there is no guitar signal to speak of at 10 kHz or higher. Heck, often there is nothing above 6 kHz on-axis with the speaker, and even lower off-axis, which is the usual listening position.
This is certainly how graphic EQs are constructed. Graphic EQs are easy to use, but complex to construct. The more bands you want, the worse it gets.
I think you may be able to get close to what you want with something simpler - one or two of the usual RC shelving filters, combined with a second-order high pass filter at the bass end, and a fourth-order low pass filter at the treble end.
My suggestion is to start by working out the actual frequency response you want from your correction filter. You can simply subtract the frequency response of your 8" speaker from your target Greenback to get the difference curve. Use the published frequency response curves of each speaker, and do this one frequency at a time. For example, if the Greenback shows 98 dB at 100 Hz and your Daichi shows 90 dB at the same frequency, the difference is (98 - 90), or +8 dB; your correction filter needs to provide this.
Do this over the entire frequency band relevant to guitar - say between 80 Hz and 5 kHz or so. Don't bother with frequencies far above 5 kHz, they just don't matter for guitar. I suggest using at least ten to twenty dfferent frequencies, so that you get a reasonably detailed correction filter frequency response from your subtraction process. I also suggest making a simple spread-sheet to do the subtraction for you.
When done, plot the result using the spread-sheet, and, in principle, you have the frequency response you will need to generate with your correcting filter.
("In principle" because in real life speakers have different frequency responses at different listening positions, different frequency responses in different rooms, different frequency responses if you simply move the speaker to a different location in the same room, and so on.)
Alternatively, if you are sure your current EQ is good enough, simply measure and post the frequency response of the actual graphic EQ you're using. You can do this by hand with a signal generator and 'scope (very tedious), or using your computer's sound-card and some audio measurement software like Room EQ Wizard (REW).
Once you get this far, post the frequency response of the correcting filter here.
The next step is to try to create a reasonably close match to that using RC filters. If the corrections turn out to be relatively small and smooth, this may not be hard. If the corrections turn out to be huge and jagged, there may not be a practical way to DIY it at all. (Digital filters will easily do the job, but you'd either have to know quite a lot about digital signal processing, or buy an expensive digital speaker emulation filter and feed it the proper impulse response.)
As Printer2 says, there are no guarantees that your Daichi can be coaxed into sounding like a Greenback, but as long as this is a labour of love, you might as well give it a try. 🙂
-Gnobuddy
Well having tested the 10 inch 200w version in a home made cab with my ZT Lunchbox as an external cab, I can say it is certainly louder than the 6.5 inch speaker in the ZT which was selected for the amp! I am fairly impressed with the sound for the money actually, I'm not even sure it's worth messing with EQ when I am this happy with the end result!
Thanks for such a detailed answer! I have now tested the 10 inch daichi in a small homemade (re-purposed hollow wooden foot stool from the local good will for $10) and I am extremely happy with the sound this thing puts out as a stand alone cab! I am not sure it's worth messing with the EQ when I am this happy, so I will likely just try to design an amp that sounds good with it!
Also, the way you describe to get the frequency response numbers, is pretty much exactly how I did it!
What I actually did was superimpose the images of each in photoshop with the top one at 50% brightness and then stretched and skewed one image to make the scales match, then it was a simple (took over an hour or two lol) matter of counting on the scale, and came up with a 10 band EQ setting to make them (kind of) match. If you notice, there is a double boost in there at the same frequency but with a different Q as there was a bump with an even bigger bump that I had to try to figure out!
And lastly, while I agree that there SHOULDN'T be any guitar signal to speak of above about 6k, the evidence I am seeing looking at celestion EQ response curves is that all of the best reviewed speakers that they sell for guitar have a distinct spike between 10 and 16k-ish! I am wondering if that awesome tone that they have come to be known for might be something to do with emphasising a few resonant frequencies that tubes/guitars/amps mat tend to bring out of a sound?
It just seems like a huge coincidence otherwise! Thoughts?
There's no image attached, so unfortunately, I can't see your correction curve.
My thoughts (which, of course, you are free to disagree with! ): Back in my late teens I was obsessed with Hi-Fi, and I tried the experiment of switching a 10 kHz, first-order RC filter in and out while listening to music. While it was quite easy to hear the change (particularly in high-hats and cymbals) if you focused on it, I found none of my friends ever noticed if I switched the filter in before we started listening to a piece of music.
Evidently that top octave from 10kHz to 20 kHz was nowhere near as important as I had thought. Even young people with good ears never noticed if I removed it entirely.
Fast-forward to the present time, and a poll taken a year or so ago showed the majority of us on this forum are in our fifties. Most of us can't even hear above 15 kHz any more, and those who either played loud live music, or went to lots of loud concerts, probably can't hear anything much above 10 kHz.
And then there is the question of the guitar speaker itself. By 16 kHz the response is usually at least 20 dB down from the 1 kHz baseline, and often it's down 30 or 40 dB. I don't think this is really going to be audible - it's too far down in the weeds, even if you have good ears.
And what of the guitar itself? Do those dense steel strings put out anything at 16 kHz, some thirty times higher than the highest open-string frequency, 329 Hz for the high E? Does the (relatively) soft wood of an acoustic guitar vibrate at such high frequencies? Does an electric guitar pickup work?
The last one is easy. We know for sure a conventional electric guitar pickup doesn't - usually it rolls off steeply (usually at a 12 dB/ octave or 18 dB/octave) above maybe 5 kHz.
Some years ago, I tried an experiment with a basic Danelectro Fish-n-Chips graphic EQ. I zeroed out all the lower frequency bands, maxed out the highest (6.4k) band, and played guitar through it. I wanted to hear what the guitar was was actually putting out at frequencies mostly above 4 kHz (each slider is nominally an octave wide.)
What I heard (and this is subjective, of course) was a nasty harsh sound that hurt my ears and wasn't at all musical (to my tastes.) You may know that the overtones of a steel string gradually go more and more out of tune - they're not really harmonics by the time you get to relatively high frequencies. They're harshly out-of-tune. That's what I was hearing.
Playing with the EQ a bit more, with all the other frequency bands back at centre, I could adjust the amount of "bite" or "edge" in the sound by fiddling with the 6.4 kHz and 3.2 kHz bands.
To my ears, one of the most common mistakes amateur guitarists make is having too much high-frequency treble in their sound. This applies to both electric and plugged-in acoustic. Often it takes a lovely performance and makes it literally painful because of "ice-pick" tone. The louder the music, the more you need to cut the high treble from the guitars, otherwise it sounds excruciating rather than pleasurable.
Speaking for myself, then, I am quite convinced that there's nothing I want to hear at 16 kHz from a guitar. In fact, there's almost nothing I want to hear above 5 kHz from a guitar.
Since you have a fairly sophisticated EQ handy, why not try your own listening experiments?
-Gnobuddy
All good points and it is good to know you have done some tests yourself to verify! I have done a few tests of my own hearing and indeed above 16k I am deaf already at 38, and 12-16k is something I have to concentrate to hear, my one and only argument, and it is a weak one but one I have heard others mention, is that the higher frequency harmonics might somehow make the lower frequencies resonate in a certain way that creates the signature sound of the speaker?
I am not disagreeing with you at all to be clear, Everything you say there makes complete sense, and this could very well just be one of those "magic sound goo" myths, I just don't have 10 grand to test all the best high end speakers to find out for myself! Thanks for putting in the leg work!
Edit: oh and to answer that top question, I was referring to the numbers, not the graph. I will post links to the graphs of both speakers though. I might even have the superposition I made.
I found the images I used to get the numbers, I saved the superposition over the daichi graph though so there are only 2 but you get the idea from the pics.
Celestion Greenback V Daichi i8000 charts - Album on Imgur
I am not disagreeing with you at all to be clear, Everything you say there makes complete sense, and this could very well just be one of those "magic sound goo" myths, I just don't have 10 grand to test all the best high end speakers to find out for myself! Thanks for putting in the leg work!
Edit: oh and to answer that top question, I was referring to the numbers, not the graph. I will post links to the graphs of both speakers though. I might even have the superposition I made.
I found the images I used to get the numbers, I saved the superposition over the daichi graph though so there are only 2 but you get the idea from the pics.
Celestion Greenback V Daichi i8000 charts - Album on Imgur
There was only one guitarist I know who really made good use of the high frequencies in his rig, and that was John Cippolino, of Quicksilver. Chipper was unique in so many ways, and his tone was very special..partly because he used finger picks with an SG! Here's a snap of his rig: (see my avatar)
Attachments
Ah, what you need is an MLSSA tester to get a spectral decay plot.
I'm guessing that there's a more modern version of this toy but it's quite useful for spotting non-linear behaviour in speaker drivers.
You have a way cheaper solution, 1000 times simpler and more accurate: just buy a real Greenback, period.
Only some $150 to begin with, substract the price you paid for your 8" wonder and we are talking around $100.
Only some $150 to begin with, substract the price you paid for your 8" wonder and we are talking around $100.
Yes. And also one of Celestion's 12" budget models has a curve very similar to that of the Greenback, just 3dB less sensitive thro'out the range. Go to their website and see if you can identify which one. (Listening to speaker comparisons on YouTube led me to checking out the published freq. curves.)
I hadn't heard of either Cippolino or Quicksliver, but Google turned this up: YouTube
I realize that we don't all have the same preferences. That said, for me, Cippolino's guitar tone has all the sonic beauty of the deafening metallic shrieking brakes on a subway train. 😱
I wonder if Cippolino had damaged his ears to the point where he was largely deaf at higher frequencies? That might explain the shrieking treble he dialed in.
-Gnobuddy
- Status
- Not open for further replies.