rplace Posted January 10, 2021 Author Share Posted January 10, 2021 @Chris A, @Rudy81, @PrestonTom, @Ziggurat I'm back at it....but I think I forgot a few things. New bass drivers and back on my open baffle project. I think these are a pretty good first swipe. I know I got a valley and peak on the SPL plot to address but the phase is looking pretty good (I think). Also think the GD and Spectrogram are decent. Any advice for delay or anything else. This is using a no-named XO around 300. Quote Link to comment Share on other sites More sharing options...
Chris A Posted January 10, 2021 Share Posted January 10, 2021 I'd work on flattening the SPL response a bit more above 1.7 kHz. Your vertical scale resolution on your SPL graph is 10 dB/division. I prefer to use 2 dB or 1 dB per division. I know that REW tends to resize everything when you go to print using its screenshot command, so oftentimes I just use "Snipping Tool" to get a screen shot that I know it will capture correctly. Looking at the SPL response at 1 or 2 dB resolution usually will show you where you need to add PEQs--or use the EQ facility within REW to find PEQs that will flatten the response. I see that prolonged energy at ~3-4 kHz in your spectrogram. I assume that you're using a full-range driver of some sort that has a wizzer cone? Chris Quote Link to comment Share on other sites More sharing options...
rplace Posted January 10, 2021 Author Share Posted January 10, 2021 6 minutes ago, Chris A said: I'd work on flattening the SPL response a bit more above 1.7 kHz. Your vertical scale resolution on your SPL graph is 10 dB/division. I prefer to use 2 dB or 1 dB per division. I know that REW tends to resize everything when you go to print using its screenshot command, so oftentimes I just use "Snipping Tool" to get a screen shot that I know it will capture correctly. Looking at the SPL response at 1 or 2 dB resolution usually will show you where you need to add PEQs--or use the EQ facility within REW to find PEQs that will flatten the response. I see that prolonged energy at ~3-4 kHz in your spectrogram. I assume that you're using a full-range driver of some sort that has a wizzer cone? Chris I've been working on the flattening of SPL. Some good progress. I'll reread the scale division and try to replicate. Yes to Full range drivers and wizzer cone. These are an open baffle set up with Dual 15" woofers top and bottom and a full range 8" driver between them. I've not given up on the big Oris horns...I do like them. Just trying something different because I had the scrap lumber and some time on my hands. As usual, thanks! No delay suggestions from GD and Spectrogram. I've iterated a bunch and this seem so be the best I can come up with so far. However, I feel like I'm forgetting a step. Quote Link to comment Share on other sites More sharing options...
Chris A Posted January 10, 2021 Share Posted January 10, 2021 Group delay/phase look good. Quote Link to comment Share on other sites More sharing options...
rplace Posted January 10, 2021 Author Share Posted January 10, 2021 5 minutes ago, Chris A said: Group delay/phase look good. I though so. Here is the phase with some PEQ mods and 2dB divisions and a sniping tool used. Looks like I am +/- 4 if you use 70 as the mid-point. I'm not too worried about the stuff above 10K for the moment. I probably can't hear it at this point 😉 What next? Keep messing or relax and listen? Quote Link to comment Share on other sites More sharing options...
rplace Posted January 10, 2021 Author Share Posted January 10, 2021 I didn't have minimum phase applied for the first GD Quote Link to comment Share on other sites More sharing options...
Chris A Posted January 10, 2021 Share Posted January 10, 2021 That group delay looks good. Chris Quote Link to comment Share on other sites More sharing options...
The Dude Posted January 11, 2021 Share Posted January 11, 2021 On 11/25/2019 at 1:17 PM, Chris A said: Note that the spectrogram "spike" is slender and almost vertical. This implies that the driver has very good internal damping or has very low moving mass (in this case, it has both). Chris, are you saying that the straighter the line, the better and what about the areas of red? I've recently picked up a set of the BMS 4590 coaxial, not sure how they stack up the BMS 4592. I've read through 3-4 threads trying to wrap my head around this, I figured this should be the year i finally learn this stuff. I have tried your guides, but seem to get lost along the way. I was getting some nasty dips around that 6k range right before the crossover. I stumbled upon your recommendations for the 4592 from another thread and thought they would give me a good starting point. I still had that nasty dip around 6k, but then decided to invert the polarity on the mid and seemed to take care of that dip. Only problem now is that the mid section from 330ish to 9900ish is about 9ish db down. If I bump this back up, i get the hiss and it sounds like a tin can. Any suggestions? If anyone wants to me to start a new thread, please say and I have no problem. One thing I will note about the Hypex fushion software is that it doesn't allow for delay below 11' or 3.8m. Quote Link to comment Share on other sites More sharing options...
Wirrunna Posted January 11, 2021 Share Posted January 11, 2021 2 hours ago, The Dude said: Only problem now is that the mid section from 330ish to 9900ish is about 9ish db down. If I bump this back up, i get the hiss and it sounds like a tin can. Any suggestions? Dude, My miniDSP hissed at me and I spent a bit of Google time and finally came across a reply by Bill Waslo of Liberty Instruments and he suggested an L-Pad. That disappeared the hiss, just had to adjust the gain to compensate. As you can't hear the hiss when it is 9db down, a 9db attenuator should be fine. https://sound-au.com/articles/l-pad-calc.htm#s1 is a calculator I use, I reckon a pair of 4.7ohm 5 watt non inductive resistors should work (calculates to 8.26db). However, you say "it sounds like a tin can" - maybe an SPL plot would explain the problem more. Quote Link to comment Share on other sites More sharing options...
The Dude Posted January 11, 2021 Share Posted January 11, 2021 Chris after re reading some of the post, I have a better understanding of the line/shift and somewhat what the colors represent. Please don't waste your time re explaining things you already mentioned. I just need to read more. It does help reading this stuff while in the process of perfecting the sound.Sent from my SM-G986U using Tapatalk Quote Link to comment Share on other sites More sharing options...
Chris A Posted January 11, 2021 Share Posted January 11, 2021 Looks like John (@Wirrunna) has also provided good information that I didn't have. Let me know if there is something that's still too obscure, etc. or questions that still remain after reading everything. I think this is a good thread for those just coming up to speed using DSP crossovers--and I think it was needed. Many thanks to Rich (@rplace) for leading the way on this one. Chris 1 Quote Link to comment Share on other sites More sharing options...
rplace Posted January 11, 2021 Author Share Posted January 11, 2021 2 hours ago, The Dude said: It does help reading this stuff while in the process of perfecting the sound. I think this is key. While you are measuring and adjusting keep the "recipe" in front of you. I tried to digest it and understand it in its entirety before I ever got started. No luck. This is like a lab class....hands on learning. And you won't get right the first time. What took me literally days now takes a couple of hours. I found it helpful to realize analyzing and adjusting for one screen impacts the others. Quote Link to comment Share on other sites More sharing options...
The Dude Posted January 11, 2021 Share Posted January 11, 2021 11 hours ago, Wirrunna said: suggested an L-Pad. That disappeared the hiss, just had to adjust the gain to compensate. So one could bump up the gain back to level everything out, which would make the hiss present. Then add an L pad or resistor, which wouldn't lower the db in that region, but elimanate the hiss? Am I understanding that correctly? Quote Link to comment Share on other sites More sharing options...
The Dude Posted January 11, 2021 Share Posted January 11, 2021 11 hours ago, Wirrunna said: However, you say "it sounds like a tin can" - maybe an SPL plot would explain the problem more. I think it was that the mid was that much hotter so it sounded like someone talking through a cheerleader's megaphone. Maybe this is all of the honk that people talk about with horns, maybe its personal preference? But yes, I should have a copy of the SPL plot on my personal laptop that I can share later. Quote Link to comment Share on other sites More sharing options...
Wirrunna Posted January 12, 2021 Share Posted January 12, 2021 10 hours ago, The Dude said: So one could bump up the gain back to level everything out, which would make the hiss present. Then add an L pad or resistor, which wouldn't lower the db in that region, but elimanate the hiss? Am I understanding that correctly? Not quite. In my case the hiss was coming from the miniDSP. The signal chain is Raspberry Pi player >>--> NAD pre-amp >>-> miniDSP output 1 >>-> NAD amp >>-> K-Horn bass bin miniDSP output 2>>-> N-Core NC122MP amp >>-> B&C DCX464 Mid miniDSP output 3>>-> N-Core NC122MP amp >>-> B&C DCX464 Hi Turning up the NAD volume control didn't affect the hiss level but sure did affect the signal from the RPi player. So, I added a 12db L Pad across both the Mid and Hi sections of the DCX464 - hiss all but gone -sticking your ear in the mouth of the Eliptrac you can still just hear it, but I can't from a metre away. Now, when I did a REW measurement the spl from the DCX was down by 12db which required an adjustment of gain in the miniDSP which boosted the signal but not the hiss. So I was assuming that the DSP section in the Fusion amp is adding the hiss just like the miniDSP does. If you have a spare L Pad speaker attenuator you could try it out. 1 Quote Link to comment Share on other sites More sharing options...
The Dude Posted January 12, 2021 Share Posted January 12, 2021 I believe this highlighted area is the area where I had increased from -10.5 to -8. It got everything gain matched, oddly this is across 2 crossover points. Quote Link to comment Share on other sites More sharing options...
Wirrunna Posted January 13, 2021 Share Posted January 13, 2021 Dude, I just looked up the specs on the BMS 4590, that driver at 118db is even more sensitive than the 111db B&C driver I use. I'm sure that the hiss you hear is the same circumstances as my setup. You are cutting the signal level for the super sensitive mid and hi so it is down near the noise floor. After putting the L-Pad in my system I have the mid at +5db and the Hi at +2db. As for the cheerleader's megaphone sound, was it before you applied -10.5db ? Quote Link to comment Share on other sites More sharing options...
The Dude Posted January 14, 2021 Share Posted January 14, 2021 15 hours ago, Wirrunna said: As for the cheerleader's megaphone sound, was it before you applied -10.5db ? minimum hiss and no megaphone sound at -10.5db If I bump it to -8, I that highlighted frequency range is flatter, but there is a noticable hiss and that is when things sound as a megaphone. I actually a slight on both channels, one with they hypex and the other channel with the Ashley. But I realized I was running everything hot, so I bumped the input gain down and made some adjustments on my minidsp and I am happy to report barely any hiss. I have another amp coming in tomorrow so hopefully this weekend I can get the left side dialed in. There is lots and I mean lots of room for improvement, but all in all I am liking what I am hearing. I am going to dable here and there, but once i get the other driver and amp running, I am going to focus on sound panels before really trying to dial these in. Quote Link to comment Share on other sites More sharing options...
The Dude Posted January 17, 2021 Share Posted January 17, 2021 I feel like an optometrist, camera 1 or camera 2? Checking to see what everyone else's thoughts are, is there other measurements I should look at? Here are the 3 spectrogram from beginning to where I am now. As you can see, I have a decent shift either around the 600hz with a 1st order crossover at 450 hp and lp, and the area now at the 400hz section with 2nd order 420 hp and 470 lp. No matter what I try, I cant seem to find the equation to get rid of that shift or get back to that nicer straight line in the first screen shot. Any ideas to try? Quote Link to comment Share on other sites More sharing options...
The Dude Posted January 17, 2021 Share Posted January 17, 2021 Those last measurements were from the left, I started with what I had programmed for the right, but they did not look like this (which is the right channel). Again, I have a shift around 800 this time. I would need to see what is set different. But as one could see how it may be important to know that if the room isn't treated correctly (which is a guess of mine), that one would need to eq, delay, etc.. each speaker independently. So I guess without any more room treatments, it may be hard to get them equal? Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.