Had a little success getting this worked out. User jtalden on HomeTheaterShack created a spreadsheet application to help create text files, but it only does HP or LP based on a single frequency. To get around this I created a HP first and copied those values, then entered my LP values and looked for the amplitude where they "leveled out" and copied just those frequencies to merge with the HP portion of the text file. It's a bit of a PITA but works very well if you have a large bandwidth. If you're only working with a narrow range it's difficult to find a "merge point" without making the graph wonky. Ex: my midbass curve is off by about .02dB at 122hz because that's the closest I could get. No big deal really.

The thread is titled "REW for active crossover measurement" from late 2011 if anyone want's to search for it on HTS.

I uploaded a couple curves in case anyone wants them. The midrange is 300-5k with 4th order L-R slopes on either end, and the midbass is 50-300 same slopes.

 

Open up notepad and write. Use 1/6oct spacing. The software will extrapolate the numbers in between

Tapaaatalk!!

 

So if I shared my FR at 1/3oct along with xover points, slopes, gain on each channel etc, you could plug that into your processor and get a sound that is somewhat similar to what I have. Of course cars, equipment and install locations are different, but still As a house curve / starting point it may work.

I stopped measuring a couple of years back and all my tuning is by ear. Sure I fell on my face about a million times but eventually tuning became instinctive. If nothing else its a different sound to experience and take it forward in your car.

I'm ready to pull out my spl meter and measure, except I don't have the white noise tracks for 20-200hz. Han's tuning disk only starts at 200hz. I know this is most unscientific and there are a ton of variables that will be different in each car but yeah it would be cool to share the overall system response and see how that pans out for different cars.........as a basic starting point

 

It would sound similar, yeah.

Tapaaatalk!!

 

do you have tones from 20-200 that I can download?

 

Download Audacity (free) and create your own tones.

 

https://www.dropbox.com/s/qhgic8b8ee517gf/Disc%203%20%28Tones%29.rar

Here ya go

 

Hi SPLEclipse!

The latest version of Room EQ Wizard (V5.00 30th January 2011) allows you to set the high-pass target curve inside the EQ module to anywhere between 5hz and 5khz. This is a "new" feature I was not aware of because I was using an older version, which is why I built my house curves using excel and the equation included in the photo below. I highly recommend using either this built-in feature of REW, or trying the method Hanatsu suggested earlier. If you'd like to learn more about my method, I'll try to explain a little bit, but it requires some super-sleuthing in excel and isn't easy for a beginner.

I found the above equation from the Texas Instruments - Basic Introduction to Filters on page 13 and modified it a little bit. The entire equation is squared because a 4th order linkwitz-riley is two 2nd order bessel filters on top of eachother. For the variable n, use the number 2. For the variable fo, use your crossover frequency (300 for 300hz for example). The variable X is your independent variable (like in a X-Y graph). The values of Y will be spit out in terms of decibels, and you can put the values into your "house curve" text file.

I know this is a little complex, but keep asking questions if you would like to learn more. I can explain more and add some more screenshots if needed.

 

^^^Thanks Jazzi (sorry I misspelled that before)! I forgot that REW can present a high-pass curve, it's something I've used before but forgotten somewhere along the way with more complicated house curves.

I think the stumbling point for me is using Excel to spit out the numbers, and that's something I'll just have to work on. I have to admit you've got a talent when it comes to those equations; it makes sense to me but isn't something I could do on my own.

What I'm working on is taking a little step backward. It's easy to set up every speaker to be "close enough" to ideal, and then just bulldoze over the response with an overall house curve and push the biquads out to my minidsps. While I image this can work most of the time, my measured results and subjective impressions have been mostly good but sometimes lackluster. Soooo...with your inspiration (and some recent wacky behavior with my minidsp, see my thread "I win for most distortion") I'm going back and starting at a more basic level by paying attention to what each individual driver is really doing.

 

You can 'simulate' shelf filters, low and highpass, notch filters as well as PEQ in REW. The AutoEQ function is great

Tapaaatalk!!

 

You must find a measuring technique that correlates with what you hear. At low frequencies (below 200-300Hz) it doesn't matter much how you place the mic, if you're in the car or not, spatial averages etc. In the midrange and tweeter range it's critical how you measure and average as it will have an impact on the target response. I do 8 averaging measurements while sitting in the car. If I'm measuring left midrange I do 5 sweeps with the mic around left ear and 3 sweeps around right ear. If I'm measuring right side I do 6 sweeps with the mic around right ear and 2 sweeps around left ear. This generally works best for me, it takes crosstalk into account. I actually back engineered this technique from centering stage by listening to pink noise.

Tapaaatalk!!

 

Ok here goes. First up, all you guys who measure, hats off. It took me an hour to do 4 sets of response readings which I then averaged out. This is not the reading from my main car since my bit10 finally called it a day last weekend. I never got around to doing a FR for the Ford, never thought I'd need it. It would have helped while setting my new processor as and when that happens.

So for now all I have is the wifes car which has a 2 way+sub active off the p-80rsII. Mids are Polk Audio MM6501 and the tweeter is the vifa XT25 clone that came with the SR6500. Amps are Polk Audio and the sub is a 10" ID in a 1cuft sealed box. The car is a small hatchback. There is a night and day difference in span of control and resolution between the p80 and a full processor.

But surprisingly I was able to dial in a fairly decent baseline. Some frequencies stick out and need correction, where I have no control eg, 160, 600, 1 khz, 1.6 khz and 4khz. I have adjusted these in the measured response below, based on what I'm hearing. Sub and mid bass crossed at 50hz with sub on 30db slope and the mids at 24db/oct. Mid and tweeter crossed at 3khz, 24db slope all round. Tweeters are cut 6db.

1. Set the baseline volume for measuring by using uncorrelated pn track. HU set at 54/62 and the measured level was 80db. Ambient / floor noise was 31 db.

Measured Response is as follows:

20hz - 86db
25Hz - 88db
30hz - 89db
40hz - 90db
50hz - 90db
60hz - 88db
80hz - 82db
100hz - 83db'
125hz - 81db
160hz - 79db
200hz - 78db
250hz - 82db
315hz - 84db
400hz - 84db
500hz - 86db
600hz - 85db
800hz - 84db
1khz - 84db
1.25khz - 80db
1.6khz - 76db
2khz - 74db
2.5khz - 71 khz
3khz - 67db
4khz - 61db
5khz - 62db
6khz - 58db
8khz - 60db
10khz - 60db
12khz - 57db
16khz - 52db

In my car this is a fairly decent baseline (for a couple of days of tuning). I'm curious how this would shape up in your vehicle. Look forward to comments. Once you have dialed this in, if you can describe what sounds right and wrong we can try tweaking it a bit.

Hope ya'll have fun

 

Ok here goes. First up, all you guys who measure, hats off. It took me an hour to do 4 sets of response readings which I then averaged out. This is not the reading from my main car since my bit10 finally called it a day last weekend. I never got around to doing a FR for the Ford, never thought I'd need it. It would have helped while setting my new processor as and when that happens.

So for now all I have is the wifes car which has a 2+sub active off the p-80rsII. Mids are Polk Audio MM6501 and the tweeter is the vifa XT25 clone that came with the SR6500. Amps are Polk Audio and the sub is a 10" ID in a 1cuft sealed box. The car is a small hatchback. There is a night and day difference in span of control and resolution between the p80 and a full processor.

But surprisingly I was able to dial in a fairly decent baseline. Some frequencies stick out and need correction, wheher I have no controleg, 160, 600, 1 khz, 1.6 khz and 4khz. I have adjusted these in the measured response below. Sub and mid bass crossed at 50hz with sub on 30db slope and the mids at24db/oct. Mid and tweeter crossed at 3khz, 24db slope all round.

1. Set the baseline volume for measuring by using uncorrelated pn track. HU set at 54/62 and the measured level was 80db.

Measured Response is as follows:

20hz - 86db
25Hz - 88db
30hz - 89db
40hz - 90db
50hz - 90db
60hz - 88db
80hz - 82db
100hz - 83db'
125hz - 81db
160hz - 79db
200hz - 78db
250hz - 82db
315hz - 84db
400hz - 84db
500hz - 86db
600hz - 85db
800hz - 84db
1khz - 84db
1.25khz - 80db
1.6khz - 76db
2khz - 74db
2.5khz - 71 khz
3khz - 67db
4khz - 62db
5khz - 63db
6khz - 60db
8khz - 61db
10khz - 60db
12khz - 57db
16khz - 52db

In my car this is a fairly decent baseline (for a couple of days of tuning). I'm curious how this would shape up in your vehicle. Look forward to comments. Once you have dialed this in, if you can describe what sounds right and wrong we can try tweaking it a bit.

Hope ya'll have fun

Anyone who wants to try this can copy the below into a text file. I formatted it so that REW can read it and dropped the dB scale a bit so it will be easier to get around the same level as the measurements. I'll give this a shot later this afternoon...I'm interested to see how that midrange peak sounds.

20 26
25 28
30 29
40 30
50 30
60 28
80 22
100 23
125 21
160 19
200 18
250 22
315 24
400 24
500 26
600 25
800 24
1000 24
1250 20
1600 16
2000 14
2500 11
3000 7
4000 2
5000 3
6000 0
8000 01
10000 0
12000 -7
16000 -2

 

I'll take a look at it later

Tapaaatalk!!

 

Plug it into your helix and see what you come up with

 

Some things that stand out from the FR:

1. 40-200 : There is a 12 db roll off from 40-200. I'm not a bass head and with the drivers in phase its enough bass for me. If you feel like you want more, raise 40-50hz a bit and then 100. Or you can try rolling off the very top end for a slightly darker sound, more on that later. 50hz is mass in your base and 100hz is energy. You want a good balance between these two. 80hz and 200hz both bloat your bass but in different ways. 160hz is critical, a touch more and its too fat a touch too little and its paper thin. Mac drilled that into my head . You want just enough of the lower end. Excess will start masking clarity and muddying the sound.

2. 200-1khz : From 200 and up there is a 5 to 6db rise in response till about 1 khz. This rise sets up the entry to your mid range frequencies. This range is vital for adding clarity to the sound. 300hz is bottom end for female vocals and it gives power to male vocals. 400hz adds chaff to the sound as does 4khz at the top, grainy vocals. 500hz is where you really start to hear the clarity, on my eq this is cut the least. If the attack on drums sound woody and dull try raising 600 a bit.

The 500-800 range if overcooked will make the sound thin, hokny, telephone like. 1 khz is great for dynamics. I will try and keep 1 khz a touch higher than 800 and 1.2. It is great for the harmonics of your mid bass, the snap that you look for starts in the 600-1khz range. I will start rolling off the response 1.2khz and beyond.

3. 1.25-5khz : Watch 1.6khz. It is a typical peak in most setup and you want to cut here. 1.6 makes the sound screechy, sharp however you want to put it. I would cut 1.6 agressively and then flatten out the dipping response at 2 before rolling it off again. 3&4 just need to be cut to the point where they are not offensive and you don't lose clarity. Boosting here never works, period. Try opening a bit at 5khz viz a viz 4khz. 3khz adds chaff to the sound. If you want to make your sound less grainy cut at 400 and 4khz.

4. 5-15khz : There is a big roll off here. I'm sure partly due to the roll off of the equipment. The PN tracks sound strong by ear till about 15khz even though measured numbers are tailing off. so its got to be the equipment. keep this in mind when correlating my measurements.Use 6-10khzto manage sibilance. I use 12khz and up like a brightness control on a TV. A steeper roll off means a slightly darker sound with a more prominent lower end. A more gradual roll off means brighter, clearer more open sound, but if 1.25-4khz or 300-800 is set too aggressively, the sound will be a bit too sharp. You want it clear and sharp only when the sharpness is in the recording. Having it clear but sharp[ all the time is not what you want.

[Edit]With the limited dsp on the p-80 my response is about the best I could achieve in this car. Its about decent in my car. But based on your install and environment, some things may get exaggerated or just not sound right. Try to identify whats wrong and then look at the above to see if you can identify where to cut. Hope this helps.[/Edit]

 

^^good eye. I didn't even notice that. Just got done setting up my individual driver EQ/TA last night. I'll try yours sometime tonight.

 

SQNut curve



Hanatsu curve



Well, I got a 4-way system so crossover points are different. But wow... the two curves sound A LOT different as you can imagine. Sqnut's and mine are similar from 4kHz and up but I got 20dB attenuation in the midrange vs the lows and highs. You could say that I got an enormous amount of bass compared to his since the frequencies are relative to each other after all.

I tried to simulate sqnut's HC but it didn't work out in my system. No lows, too much midbass/upper midbass, way too much midrange and non existent highs.

I think the way we derived the curves differ too much.

 

Ya'll listen to way too much bass. But seriously when you compare two curves that are so radically different, the brain always picks the one you are more accustomed to. The way to do this is to run my curve for a few days, get used to the different sound and then tweak for what you feel is more or less and see what you end up with. A 20db difference in midrange is just way too much.

Based on your post I went back to the car and listened to the Tracy Chapman CD I used while tuning. It sounded decent and then I took the CD and played it on my home setup with the Scan floor standers and yeah similar sound. So conclusion is, you're listening to too much bass and not enough of the 300-1khz range. Make the 1.25-4khz slope steeper if you want to cut the midrange, cut the 300-1khz range by a couple of db's if you want, but most importantly give the new curve a chance to settle on your ears. After a few days go back to your old curve and see what you think. If you don't have the mid range right nothing else matters.

PS : Just so that we are on a level field, I am going to try to implement Hans curve in the car and see how that sounds and run with it for a couple of days. Let's see what happens. Han, do you have a ref sound that you are tuning for? Not the FM/ELC but a real sound that you are using as a reference?

 

My midrange real FR look like this last time I tuned the system, can't remember if I did any changes since then...

 

Just for the hell of it here is mine:

 

Looks pretty similar to mine. I generally prefer I little less energy above 2kHz though.

(Might have something to do with the fact that I continuously listen at ~100dB in the car xD)

 

^^Han, I've been experimenting with yours. Definitely a lot of bass, and taking a look at my screen name, that should tell you something.

How did you simulate those graphs? That's along the lines of what I'm looking for with this thread.

 

How did you simulate those graphs? That's along the lines of what I'm looking for with this thread.

Did you get what you were looking for? I can't really tell if we answered your question or not

 

Nothing fancy. RoomEQ can do it.

The HC's need to be put in relation to which volume you're listening to. I've taken precautions regarding distortion in the lows. My sub is designed to induce very little non-linear distortion. Low distortion in subwoofer frequencies and you subjectively feel there's "less" bass, so I increase it to compensate. You generally need 1-3dB less bass with sealed boxes cause they induce lots of distortion at louder volumes. Vented alignments can be almost "distortion free" if used and designed correctly. Imo, vented is superior to sealed in all ways but one, size.

 

Bass isn't an issue in my car.

 

^^^If you were to overlay the two curves so that 50hz was at the same amplitude, they don't look as dramatically different except for above 1khz and below 50. The SQnut curve is only about 5dB hotter through the midrange.

Still tuning, but mine is looking like a combo between the two, lol. I like a little more at 1khz and a midbass drop that extends out to about 250 vs. the often advised 160.

 

Keep going, you're on the right track. Listen to how it sounds and use the frequency guide that I wrote out. The 5db hotter mid range can be 3-6 db depending on where you're hearing that range from. If you've got a 3 way with the mid range mounted up high with little or no obstructions even a 3db rise would do. Remember that is a range we are sensitive to so a 1-2db rise is very audible.

If on the other hand 300-1khz is from a 2 way (like mine) where the mid bass is is low in the door and blocked by my legs, passengers legs etc you may need a little more. Nothing is cast in stone, get a solid reference eg with a 2ch setup and then tune towards that.

 

I used the ELC/FM curves a lot as well. As shown we have a loss of sensitivity starting around 500 and peaking around 1.5 khz. One way to do it keep it flat from ~200-800 and then lift 1-1.6 a few db. I've done that and it sounds bad. The sound just turns thin and screechy.

But that dip in sensitivity is still there. You're still perceiving 1khz lower than 300hz even if the two measure the same. So I prefer to gradually raise the response from about 300 to ~1khz. Ideally I would like it more gradual resulting in 1khz only being 3-4db louder than 300, rather than the current 5-6. That is down to the fact that the resolution on the p-80 sucks when compared to a processor. On my bit 10 I could cut/boost in 0.5db steps on the p-80 it's +/- 2db so, what I got was the best I could do.

 

If you follow the FM/ELC curves, the volume you listen to also plays a role in how your ear is perceiving the sound. So at Han's avg level of 100db 300-500 is much flatter and the dip at ~1.5 is not as pronounced. So he can run 300-500 flat and even a 1-2 db gain from say 300-1khz would be enough. My normal listening level is 80-90db so I'm more inclined to follow the 80db/phon curve.

 

I agree 1kHz is v. Imp

 

Anymore house curves anyone would like to share?
I'm currently working on mine and testing.