My stereo upgrade project

[Matti777]

Ok the car is back together now. The factory 6.5" mounting brackets seem to work and I used some CCF weatherstrip to make sealing rings. Just a warning...the door panels are flimsy and easy to break. I broke one mounting clip peg during reassembly and luckily was able to epoxy it back together.

The next phase of the project is tuning with REW and the miniDSP. I've taken some preliminary RTA readings using my acoustical mic and laptop and played around with REW which has a steepish learning curve for an mech eng tech/bass guitar dabbler who has only dabbled in acoustics.

Initially the sound was not too impressive. I eq'd in a low shelf filter at 60hz at 10db per recommendations from one of the tuning experts just to figure out how this whole tuning routine works. That helped a lot.

I took some more preliminary RTA readings of just the left and then the right side to make sure everything was ok with the speakers. The 6.5's are used JL Audio C3-650s. The curve from side to side are definitely a little different. I figured out how to eq to match the target house curve but it quickly became obvious that with only limited EQ points per channel I would need to take RTA readings of each speaker separately, EQ them separately, update the DSP and then take new readings. I guess at some point I would take readings of the whole system to confirm that I'm matching the target curve. After that I could then decide if the curve is to my liking and experiment a bit. Apparently some people like to eq in more bass and have different slopes on the upper frequencies ie flat or descending. Thats my plan for now at least. I will perhaps post some curves once I get a little further along.

Sponsored

 

[StangTime]

Ok the car is back together now. The factory 6.5" mounting brackets seem to work and I used some CCF weatherstrip to make sealing rings. Just a warning...the door panels are flimsy and easy to break. I broke one mounting clip peg during reassembly and luckily was able to epoxy it back together.

The next phase of the project is tuning with REW and the miniDSP. I've taken some preliminary RTA readings using my acoustical mic and laptop and played around with REW which has a steepish learning curve for an mech eng tech/bass guitar dabbler who has only dabbled in acoustics.

Initially the sound was not too impressive. I eq'd in a low shelf filter at 60hz at 10db per recommendations from one of the tuning experts just to figure out how this whole tuning routine works. That helped a lot.

I took some more preliminary RTA readings of just the left and then the right side to make sure everything was ok with the speakers. The 6.5's are used JL Audio C3-650s. The curve from side to side are definitely a little different. I figured out how to eq to match the target house curve but it quickly became obvious that with only limited EQ points per channel I would need to take RTA readings of each speaker separately, EQ them separately, update the DSP and then take new readings. I guess at some point I would take readings of the whole system to confirm that I'm matching the target curve. After that I could then decide if the curve is to my liking and experiment a bit. Apparently some people like to eq in more bass and have different slopes on the upper frequencies ie flat or descending. Thats my plan for now at least. I will perhaps post some curves once I get a little further along.

From your measurements, can you see a null around 550hz on the passenger side? I am, and I think it has to do with a cancellation between the mid-bass driver and the distance to the center console.

 

[Matti777]

I just came in from the garage after recording an overall baseline RTA and each speaker individually. I updated with the DSP with a sub +12db boost before I started because I know it needed that anyway. It also kept the low shelf on the mid bass speakers. I'm sure some experts would not recommend that but the low end sounds pretty good to me. I haven't had much time to analyze anything yet. I noticed that sitting in the passenger seat to take readings had a significant effect on the readings so I decided to stand outside the car with the laptop on the roof (I also realized that the garage stereo was interfering when I noticed a bass response when measuring the tweets ) It meant that the mic was held in one position but I think that is good enough for me.

The right mid bass seems to have dips at 600 and 800 hz. Honestly I don't know what to do from here. My overall baseline looks to be reasonably close to the target curve (not displayed here) but when I compare the individual curves to the baseline they are not even close.

Overall

Right Mid Bass

 

[StangTime]

I just came in from the garage after recording an overall baseline RTA and each speaker individually. I updated with the DSP with a sub +12db boost before I started because I know it needed that anyway. It also kept the low shelf on the mid bass speakers. I'm sure some experts would not recommend that but the low end sounds pretty good to me. I haven't had much time to analyze anything yet. I noticed that sitting in the passenger seat to take readings had a significant effect on the readings so I decided to stand outside the car with the laptop on the roof (I also realized that the garage stereo was interfering when I noticed a bass response when measuring the tweets ) It meant that the mic was held in one position but I think that is good enough for me.

The right mid bass seems to have dips at 600 and 800 hz. Honestly I don't know what to do from here. My overall baseline looks to be reasonably close to the target curve (not displayed here) but when I compare the individual curves to the baseline they are not even close.

Overall

Right Mid Bass

What do your plots look like with a 5db scale no smoothing? 50db is not enough resolution.

 

[Matti777]

The posts are just for interests sake. I will of course be working with 5 db scale. My understanding is that 1/6 or 1/12 smoothing is normally used. I can post more later. Also, sub is crossed over at 60 and mid is 80/2000 (hence the 12 db boost on the sub for the 1/2 octave difference)

 

[mikes2017gt]

I haven't had the time to do any tuning on my new install yet. But I think that works out because I'm learning a lot from this thread! I'm pretty clueless when it comes to REW so you guys that know what you're doing, please keep doing it. I'm taking notes.

 

[Matti777]

I haven't had the time to do any tuning on my new install yet. But I think that works out because I'm learning a lot from this thread! I'm pretty clueless when it comes to REW so you guys that know what you're doing, please keep doing it. I'm taking notes.

I'm one step up from clueless so I wouldn't pay too much attention to what I am doing quite yet. I should be a little smarter once I have a couple iterations of making eq changes, updating the DSP and taking new readings to see what kind of mess I have made.

 

[mikes2017gt]

I'm one step up from clueless so I wouldn't pay too much attention to what I am doing quite yet. I should be a little smarter once I have a couple iterations of making eq changes, updating the DSP and taking new readings to see what kind of mess I have made.

LOL! I hear you. Please do continue to share what you're doing and how you're doing it. Good luck with tuning!

 

[Cathul]

It also kept the low shelf on the mid bass speakers.

Did you also use a low-shelf on the bass? I ask, because if you do a low-shelf on midbass you need it also on the subwoofer to keep the phase in sync. If you boost the sub and keep a gap in your crossover you usually don't use a low-shelf on the midbass.

To make things easier you could also use Jazzis spreadsheet to generate target curves for your drivers.

I'm sure some experts would not recommend that but the low end sounds pretty good to me. I haven't had much time to analyze anything yet. I noticed that sitting in the passenger seat to take readings had a significant effect on the readings so I decided to stand outside the car with the laptop on the roof (I also realized that the garage stereo was interfering when I noticed a bass response when measuring the tweets ) It meant that the mic was held in one position but I think that is good enough for me.

No it's not. Sit in the drivers seat and wave the microphone around like Kyle Ragsdale is doing in his videos or take several sweeps on different positions and generate the average response in REW when measuring with you outside of your car.
Measurements can be drastically different when taken only 2 inches apart due to all the reflections in the car.

The right mid bass seems to have dips at 600 and 800 hz. Honestly I don't know what to do from here. My overall baseline looks to be reasonably close to the target curve (not displayed here) but when I compare the individual curves to the baseline they are not even close.

Well, if the overall baseline (left/right together) is significantly different than the idividual measurements (left only and right only) then you have significant phase problems that lead to cancellations at different points in your response.
If everything is in phase at all frequencies then the combined response should be up to 6db higher at every point of your measurement and look like your individual measurements.
On the other hand you don't have done any serious eq yet (which you cannot really see due to the low resolution of your screenshots).
If your left and right side are within a 1-2 db margin everything should look a lot better.

Right Mid Bass

Not enough resolution tbh, but almost 50db difference between subwoofer and tweeters are not that good. 20db difference between 20 and 20k Hz would stretch it imho and is subject to personal taste, but 50db is insane.

f.e. this is what my right midbass looks like with a 5 db resolution and 1/48th octave smoothing before equalizing and after equalizing.

 

[Matti777]

Did you also use a low-shelf on the bass? I ask, because if you do a low-shelf on midbass you need it also on the subwoofer to keep the phase in sync. If you boost the sub and keep a gap in your crossover you usually don't use a low-shelf on the midbass.

To make things easier you could also use Jazzis spreadsheet to generate target curves for your drivers.

Thanks for your thoughts on the situation. Its a steep learning curve. The local shop had a $150 special to tune stereos and maybe I should have taken them up on it but what is the fun in that ?

I originally had a low shelf on the sub but removed it. I'll probably remove the low shelf on the mid bass too as I only realized recently the gap and 12 db sub boost keep things in phase. The low end does sound good though....maybe it can be better.

I've got the Hanatsu curve loaded right now and generally I am following the Audiofrog guideline although at this point I am just dabbling. I'm a functional/built for purpose guy so all I want is something that makes the most of my equipment without going to the nth degree. Last night after studying the individual curves I began to understand why the Audiofrog guide makes sense. After all we hear the speakers together anyway. They go on about correcting imaging but maybe that part is for later.

I was looking at the individual curves and how I might eq I had more questions than answers. For example, how do you eq when the curve crosses the target curve and doesn't come back? This isn't near a crossover point. You can't just drop the target curve down. Perhaps this a level issue? I'll redo the RTA readings recording only left and right side per the Audiofrog guidelines and see what I get.

No it's not. Sit in the drivers seat and wave the microphone around like Kyle Ragsdale is doing in his videos or take several sweeps on different positions and generate the average response in REW when measuring with you outside of your car.
Measurements can be drastically different when taken only 2 inches apart due to all the reflections in the car.

I agree that sitting in the drivers seat would be more realistic. Perhaps I will try that this time. I see lots of people taking readings with a mic holder mounted in the drivers seat. Some will move it to 3 different positions and average them. Some will just move the mic in a figure 8 pattern (not sure where they are positioned). Having legs in front of speakers seems to make a big difference. I can see the average response change as I reposition my legs. I didn't see that moving the mic a few inches back and forth. Perhaps I will test that again and I will be proven wrong. The question is will it change the readings enough that the eq would be different? That I don't know. I hardly see this as an exacting science anyway seeing as the acoustics in a car are a long way away from ideal but I am shooting for a decent result.

Well, if the overall baseline (left/right together) is significantly different than the idividual measurements (left only and right only) then you have significant phase problems that lead to cancellations at different points in your response.
If everything is in phase at all frequencies then the combined response should be up to 6db higher at every point of your measurement and look like your individual measurements.
On the other hand you don't have done any serious eq yet (which you cannot really see due to the low resolution of your screenshots).
If your left and right side are within a 1-2 db margin everything should look a lot better.

So far I think the individuals are looking ok. Obviously there should be some differences with the mic on the left side and there are a few nulls on the right mid bass.

Not enough resolution tbh, but almost 50db difference between subwoofer and tweeters are not that good. 20db difference between 20 and 20k Hz would stretch it imho and is subject to personal taste, but 50db is insane.

After looking at curves and the eq last night I don't think those high level plots are accurate on db. I can show you another high level plot from the EQ app that shows my overall baseline (all speakers) following the target curve quite closely. This is with no EQ except on the low end. The stereo sounds pretty good so I don't think there are any radical issues. Of course no one works with the curves at that level anyway and the issues show themselves as you change the scale down to 5db as mentioned. Onwards and upwards.

 

[Cathul]

Well, Hanatsus curve is for a bigger car. In such cars you usually have a steeper roll-off to the highs as the bass needs to be louder in larger cars. Check the Whitledge curve f.e. it's even steeper from bass to highs than Hanatsus curve as far as i remember, but John Whitledge has a big van, not a small car like the Mustang. ;)

But (!) always keep this in mind. The only judge is you. If it sounds good to you then great, but i usually recommend to try different curves with different slopes and take a listen and adapt your own curve to the best of all curves combined. Maybe you like the bass in the current house curve, great, but maybe the mids are not exactly what you want, then raise/lower the mids as needed. Same with highs... it's an iterative process and can take a while, but this makes it a hobby, right?

 

[Matti777]

Yes iterative for sure. I often hear of people getting a good tune and then still fiddling with it trying to get a bit more. I think the lows are close but mids and highs need work. I would like to find an overlay of some house curves to compare. I remember seeing one with a big cut in the 4K area. That was interesting.

 

[StangTime]

I started off using the JBL curve, and found it to be way too "hot" once the car was rolling down the road. Hanatsu's curve was better for my taste. The Mustang needs a lot more low end and mid-bass to overcome road and exhaust noise than what the JBL curve or Audio Frog curve will give you. Sitting still they all sound good. But Hanatsu's is my preference.

 

[Matti777]

I created a Crutchfield curve text file too but I think I will work with the Hanatsu based on its popularity.

The other issue I have to resolve is how to import REW EQ settings into miniDSP 6x8. miniDSP wants a different file type than REW exports. One is a txt file and one is an eq file....I can't remember which is which off hand. I have the basic miniDSP equalizer selected in REW.

 

[Cathul]

I can hear you. For my curve i also added some serious bass and midbass. ;)
Around 3-4 db more in midbass than the JBL curve. But i also deadened the shit out of the car. Trunk, doors, rear quarter panels inside, beneath the rear seat. Everythings treated with sound dampening. ;)

Sponsored