Help With Tune Please

I'm a novice when it comes to car audio and this is my first system ever. I've attempted a tune using information that's available on the internet. I was hoping the members here could help me with additional tuning. I'm not even sure what to post but I see this graph from REW posted often so I'll start with that. What additional information is needed to critique my tune? Would individual graphs for each speaker be a better start?

Any and all help is greatly appreciated!
 
Yes, individual measurements for each driver would be a great start including the raw measurements before setting EQ and crossovers.
What electrical crossover settings do you use?
Maybe even attach the MDAT file from REW or make it downloadable from any location.

At first look it looks like you got some serious phase issues between midbass and subwoofer.
So please also post the settings for time alignment with the corresponding tape measurements.
 
-24db linkwitz

All my measurements were done after setting crossovers and TA. I didn't know I was supposed to take measurements before. I read that there was a possibility of damaging speakers if crossover weren't set properly.

I used a tape measure and measured from each speaker to my head for TA. I measured the back of the sub box to my head when I did TA for the sub. I measured through the access panel in the back seat.

Here's all my measurements before and after eq. https://drive.google.com/drive/folders/1cko7WOscOmFu_yZYlLfBm5FgrZ57Cttz?usp=sharing

Here's the TA numbers from my dsp.

Driver side woofer
92.08 cm
36.25 inch
1.56 ms

Passenger side woofer
137.42 cm
54.10 inch
.23 ms

Sub
145.21 cm
57.17 inch
0 ms

Midbase

Before eq
View attachment 15518

After eq
View attachment 15519

Sub

Green is before and purple is after.
View attachment 15520
 
Sending a full range signal to tweeters will fry them unless you put a capacitor in.
What speakers are you using?
GB60 and what sub? No tweeters?
 
Thanks for the measurements, but as you've put them into individual files you lost the relative information between drivers, i.e. i cannot see at what level the drivers are at, and therefor not the acoustical crossover points.
You also didn't mention the crossovers, except for the slope, but not frequencies for low- and highpass.
When i just open the measurements without the information about the gain between drivers I see you have an acoustical crossover of about 125Hz, is that correct (probably not, but shows the problem)?
So please make measurements with the same volume on the headunit and within one mdat file, so we can directly see the acoustical crossover points in REW.

What is causing that spike in the midrange EQed responses around 35Hz? Have you loaded a calibration file?

When I simulate a LR24db crossover at 80Hz and 300Hz with my housecurve it looks like this compared to your drivers midbass.

View attachment 15521

So you're missing some volume between 60Hz and 120Hz. What is your midbass highpass set to? If it's set at 80Hz, then the slope is too steep for an acoustical LR24db alignment. Either try a lower frequency or a shallower slope like BW 12db and remeasure. Slope on the sub seems to be correct, but as said before, I cannot really see where you've set your acoustical crossover at.
 
So you're missing some volume between 60Hz and 120Hz. What is your midbass highpass set to? If it's set at 80Hz, then the slope is too steep for an acoustical LR24db alignment. Either try a lower frequency or a shallower slope like BW 12db and remeasure. Slope on the sub seems to be correct, but as said before, I cannot really see where you've set your acoustical crossover at.

To further explain this, as it took me awhile to understand it, the crossover you set in your DSP does NOT have to be LW 24dB. What you're trying to achieve is what looks like a LR24dB slope on your RTA measurement. The 'acoustical' crossover, as Cathul said. Your 'electrical' crossover can be whatever it needs to be in order to achieve that. Justin Zazzi has made a spreadsheet that makes it very easy to determine what your LR24dB acoustical crossover should like like between your sub and mid.

https://www.caraudiojunkies.com/sho...ion-for-room-eq-wizard&highlight=justin+jazzi
 
Thanks for the measurements, but as you've put them into individual files you lost the relative information between drivers, i.e. i cannot see at what level the drivers are at, and therefor not the acoustical crossover points.
You also didn't mention the crossovers, except for the slope, but not frequencies for low- and highpass.
When i just open the measurements without the information about the gain between drivers I see you have an acoustical crossover of about 125Hz, is that correct (probably not, but shows the problem)?
So please make measurements with the same volume on the headunit and within one mdat file, so we can directly see the acoustical crossover points in REW.

What is causing that spike in the midrange EQed responses around 35Hz? Have you loaded a calibration file?

When I simulate a LR24db crossover at 80Hz and 300Hz with my housecurve it looks like this compared to your drivers midbass.

View attachment 15521

So you're missing some volume between 60Hz and 120Hz. What is your midbass highpass set to? If it's set at 80Hz, then the slope is too steep for an acoustical LR24db alignment. Either try a lower frequency or a shallower slope like BW 12db and remeasure. Slope on the sub seems to be correct, but as said before, I cannot really see where you've set your acoustical crossover at.


This could be the cause of the 60-120hz issue, although that is pretty wide for this issue. The crazy amount of rise below 60 out of the midbasses is also pretty interesting.

https://www.youtube.com/watch?v=aIAV-jDafq0&t=1s
 
Cathul

gb10 - 3k to 20k
gb25 - 300Hz to 3k
gb60 - 80Hz to 300Hz
sub - 20Hz to 68Hz

Just to be sure I would need to take measurements of my speakers in their current eqed state at one volume level? How do I turn it into 1 mdat file?

I have no idea why the response spikes at 35hz. I honestly didn't even give it a second though because it was out of the speakers range. Could it be noise from outside the car? I'm measuring with a UMIK-1 using the 90deg calibration file.

I think I understand what you and jrwalte are try to say. I should try using -12db and then tune within REW to a -24db slope using the selectable settings? If that doesn't work I should move the cutoff down to 70 instead of 80?

jrwalte

I think I understand what you're saying. I should try using -12db in the dsp and if that fixes it then tune to -24db in REW? I'll look into Justin's program too.

jdunk54nl

I initially brushed the dip off thinking it was some kind of null or cancelation from me sitting in the drivers seat. I now know that isn't the case.
 
We're saying it could be anything. It doesn't have to be 80 Hz crossover or even matched between the sub and mid (they don't have to both be the same #). You can mix and match crossover types too and use LW on the mid and BW on the sub, for example. You just have to make sure you aren't causing it to be out of phase and you know this by seeing a dip around the 'acoustical' crossover you're shooting for, typically 80Hz 24dB LW.

A big part of your null looks like your 68Hz crossover of the sub. Try raising it.

In your last graph of the mid test, you should raise the volume. The results should be 75+dB I'd say. I think you see the jump in the low end because you're near the noise floor.
 
Cathul

gb10 - 3k to 20k
gb25 - 300Hz to 3k
gb60 - 80Hz to 300Hz
sub - 20Hz to 68Hz

Just to be sure I would need to take measurements of my speakers in their current eqed state at one volume level? How do I turn it into 1 mdat file?

Just make consecutive measurements of all drivers, midbass and sub without any EQ and crossovers (and after that with everything turned on) at the same volume level and then save the measurements.
You will have lots of measurements in the left part of REWs window and they will all have the same relative volume and in the "All SPL" tab you can then see the acoustical crossover.

View attachment 15522

I have no idea why the response spikes at 35hz. I honestly didn't even give it a second though because it was out of the speakers range. Could it be noise from outside the car? I'm measuring with a UMIK-1 using the 90deg calibration file.
[/quote]

Raise the volume on the headunit.

[quote]
I think I understand what you and jrwalte are try to say. I should try using -12db and then tune within REW to a -24db slope using the selectable settings? If that doesn't work I should move the cutoff down to 70 instead of 80?

Basically, yes.
And that's what you need the measurement without crossovers and EQ for.
In REWs EQ window you can then try different crossover simulations to bring the slope near the desired ideal slope of the target curve.
The acoustical curve should be a LR24db as this gives the least problems with phase in the crossover region. Two acoustical LR24db slopes will be in phase in the crossover region.
But, you can surely use a BW12db electrical filter, or an 18db filter, or a 12db LR or whatever make the acoustical measured slope match the desired acoustical slope. By doing this you may achieve lot less EQ in the crossover region.
Ideally you want to use the same crossovers on both sides, but it may be that you have to use different crossovers for your left and right midbass.
If the acoustical crossover matches this shouldn't be too bad, although it is not ideal in my opinion.
 
Thank you very much for the advice. I just want to clear up one thing before moving forward. If I take measurements using mono pink noise with all crossovers and eq settings disabled I won't damage my speakers?
 
Thank you very much for the advice. I just want to clear up one thing before moving forward. If I take measurements using mono pink noise with all crossovers and eq settings disabled I won't damage my speakers?
Set a steep crossover (24db electrical) on the tweeters at 1.5x fs to give them protection and you're good to go.
And you won't damage your midbass speakers if you're not going crazy with the volume knob while measuring.
 
Thank you very much for the advice. I just want to clear up one thing before moving forward. If I take measurements using mono pink noise with all crossovers and eq settings disabled I won't damage my speakers?


Set the gb 10 at like 2000hz with a 24db crossover

Set the gb25 at like 150hz with a 24db crossover

No crossovers on the gb60

No crossovers on the subs

Then I personally like to sit in the back seat, recline the drivers seat all the way down, put my microphone on an extension (like a 1x1 board that is 4ft long) and play mono pink noise and measure each speaker moving the microphone around my listening area. I recline the front seat based on reading Floyd Toole's work on measuring theatres and the issues of using the microphone close to the headrests of the theatre seats. The seat causes reflections our brains can't hear, but our microphones pick up. So remove it and you remove the microphone picking up those reflections, then when the seat is back, we do not hear them anyway.

I then get out of the truck, go inside where it is much cooler (I live in phoenix) and use REW to EQ everything. I do all manual EQ now but their auto eq works ok.

- Once in the EQ window, turn off all but the predicted curve and target curve (See light green circles in the bottom middle of the picture).

- Make the Frequency response take up as much room as possible (use the little arrows on the bottom left to close the bottom screen as circled in red)

- Set your crossovers on REW first by hitting the EQ filter button (even before if you use Auto EQ) in the top middle of the screen. Scroll all the way to the bottom.

- You can see on the right side, my house curve wants a 130hz 24db/oct ACOUSTICAL crossover (as in measured by a microphone in my listening position).

- In the EQ filters window at the bottom, also circled in blue, you can see BU2 or Butterworth 2nd order (2nd order means 12db/oct) at 100hz ELECTRICAL crossover is what I would need to enter into my dsp to get that to happen. This is because the electrical setting and the natural roll off of the speaker combine to form that 24db/oct I want.

Do note, this was a home bookshelf speaker that I built and was measuring. This wasn't my final work but what I could find easily and illustrated the point. I wouldn't use the target settings window on car speakers, I would import the house curve using Justin Zazzi's REW helper spreadsheet. A single bookshelf speaker is just easier to do this way for me. Car speakers 100x easier using Justin's spreadsheet.

View attachment 15526





Just for reference, if I put in an electrical setting of 130hz 24db/oct, you can see the acoustical response falls off way too quick, as in it is like a 30db/oct slope (at 100hz it is 64db, at 50hz it is 34db or 30db change). This would impact the phase and make it harder to get the speakers in phase with the subwoofers. I want to spend as little time as possible messing with open sound meter or smaart to get my phase right. The above picture would allow that. The below picture would require more phase alignment work and some "trickery" to get it right, and it would probably still be off compared to the above.
View attachment 15527
 
Last edited:
Don't use "Bass limited speaker" in the EQ window of REW. Use "Speaker Driver" instead as this gives you the option to set LR24db acoustical crossover slopes instead of the default Butterworth for the target curve (the blue line in Jdunks screenshots). ;)
 
Don't use any of those on a car. Import the house curve from Justin's spreadsheet and set REW to none.

For my home, if you set to speaker driver, you can't add a room curve, and I wanted to match the gentle slope from low frequencies to high that you would expect in a room.
 
You can. I do it all the time. I even imported the single driver curves from the spreadsheet and compared it to the speaker driver curve in REW after i loaded the overall target curve in the preferences. Ideal curve was identical to the imported single driver target curve.
 
You can. I do it all the time. I even imported the single driver curves from the spreadsheet and compared it to the speaker driver curve in REW after i loaded the overall target curve in the preferences. Ideal curve was identical to the imported single driver target curve.


The problem with the speaker driver REW curve compared to Justin's spreadsheet, is his spreadsheet takes into account all parts of the total curve. You can't do, for example, a dip between 2khz and 4khz, and have REW account for this in the individual speaker curve. His curve being imported does account for this and takes and makes the individual speaker curve appropriately for that change. Similar issues with high frequency roll off and bass increases.


If I am wrong about this, please let me know how you can. Same as if you can add a "room slant" to the curve with a speaker driver like you can with a bass limited speaker. You can see in my pictures above, I have the slant at 0.6db/oct slope from low frequencies to highs. This is matching the natural roll off that should be predicted in a room. A completely flat response would be unnatural in a room and would probably mean the speakers would have too much energy at the high frequencies aka a "bright" speaker.
 
If you set Jazzi overall curve in preferences as the house curve this exact curve is used to calculate the ideal response and slopes in the EQ window when you select speaker driver and set the acoustical crossovers for this ideal curve.

View attachment 15528

Try it yourself and be amazed. ;)
 
Well I'll be....REW must have changed how that worked because it never worked like that before (I haven't tried in a few years after the first time I tried and it very well could have been me making a mistake somewhere too)! Thanks for the tip!

Just did it with a midrange, normally when this would happen it would just be like the 2nd picture and not the first.

Still can't do room tilts unfortunately unless you have a house curve loaded that already has one.

View attachment 15529View attachment 15530
 
Last edited:
Back
Top