TOPIC:

After some trials and tribulations getting DL running on my Mac (using VMware Fusion and Windows 10), I was able to get some successful DL runs on my car audio system. While the target curve correction is great, I'm running into issues with imaging. I have run several DL tunes and my image is always pulled to the right of the driver's seat. Here is how I'm set up:

In the plugin:
FL > Input1/Dirac1/Output1
FR > Input2/Dirac2/Output2
RR > Input 3/Dirac3/Output3
RL > Input 4/Dirac4/Output4
Sub > Bass Mgt/Dirac5/Output5

In DL:
Custom 5 channels mirroring the above config

When looking at the Dirac Live settings in the plugin, I notice a couple of things: 1) the right side speakers are a little more than 3 dB higher than the left, but that doesn't sound right to my ears - it should be closer to ~1 dB 2) even when compensating for the levels, the delays are still off a little as fixing the levels doesn't quite pull the image back to the left.

One thing I noticed in the custom 5 channel setup was that DL seemed to expect config L/R/C/RL/RR and I was changing this to L/R/RL/RR/Sub before my runs. I thought maybe that DL might be thinking that my RL speaker was supposed to be a center channel, thus jacking up all the settings. To test this, I ran a couple of custom 3 channel tunes where DL is looking for L/R/Sub. Unfortunately, I still ran into the same problems of a too-hot right channel and the image pulled to the right. A poster in another thread suggested that perhaps I mirrored the mic positions accidentally, but double-checked that and even ran a tune with the mic positioned mirrored intentionally but that didn't yield any better results.

So far, I'm pretty stumped. I'm hoping someone here might have some suggestions. TIA

In DLCT - Are the L&R side target curves linked on the filter design screen before optimizing? This ensures they are at the same level.

I believe delays are determined primarily by the first "sweet spot" measurement. And levels are directly proportioned to the target curves and how they compare level wise to each other.

I did have them linked for the stereo tunes I did, but not for the 4-channel measurements. The stereo tunes, not surprisingly, were a little better. I'll incorporate that into my next attempt, thanks!

I was going to write a long post asking questions, etc., but I found this video and it clears up a ton of how to position the mic for DL measurements:

I ran one more tune with all four channels and the sub making sure to link the channels on the target tab. I still had an image I didn't like, so I just "fixed it in post" in the CDSP plugin by adjusting the front right delay and gain a little.

This got me to thinking about what I should expect in a good image. My expectation is that the center of the image is directly in front of me (i.e. centered on the sweet spot). In looking at a couple of more tutorials, it seems that others like the image to be centered on the dash from the perspective of the driver's seat. Now, I understand all this is subjective, but is DL looking to place the image centered on the dash or on the sweet spot? If it's the former, then I get why the image seems to be "shifted" from my preference

itlnstln wrote:

I ran one more tune with all four channels and the sub making sure to link the channels on the target tab. I still had an image I didn't like, so I just "fixed it in post" in the CDSP plugin by adjusting the front right delay and gain a little.

This got me to thinking about what I should expect in a good image. My expectation is that the center of the image is directly in front of me (i.e. centered on the sweet spot). In looking at a couple of more tutorials, it seems that others like the image to be centered on the dash from the perspective of the driver's seat. Now, I understand all this is subjective, but is DL looking to place the image centered on the dash or on the sweet spot? If it's the former, then I get why the image seems to be "shifted" from my preference

Bingo. Yes, the center stage image should be centered between the far left and far right boundaries... So basically the center of the dash. If center image were in front of you then the left side would be compressed compared to the right.

I also found the image pulled to the right after dirac cal. I adjusted with +2 on left channels in the mixer. That gives me an image in front of the driver.

Season to taste, of course. But my understanding is DIRAC wants your "center" to be the center of the dash. I also found this a bit odd as I've always tuned myself to get it to the center in front of me (driver). But after living with the center "in the center of the dash" for a few weeks I really do feel this is better. It makes for a much better soundstage that is deeper and wider. It just seems like you're a bit to the left of the performer, as if you got some cheaper seats the tradeoff was much less of one after getting used to it. I like it more now. It's not like sitting in the sweet spot in your home theater but of course this is a car and it's going to be far from perfect no matter how much we work on it.

Phatfos1 wrote:

Season to taste, of course. But my understanding is DIRAC wants your "center" to be the center of the dash. I also found this a bit odd as I've always tuned myself to get it to the center in front of me (driver). But after living with the center "in the center of the dash" for a few weeks I really do feel this is better. It makes for a much better soundstage that is deeper and wider. It just seems like you're a bit to the left of the performer, as if you got some cheaper seats the tradeoff was much less of one after getting used to it. I like it more now. It's not like sitting in the sweet spot in your home theater but of course this is a car and it's going to be far from perfect no matter how much we work on it.


its not so much what Dirac wants , it’s what’s proper...

a properly imaging system will have the vocal between the speakers , regardless of where you sit.... 

If you use tape measurement distance to ms and plug into raw delay so the arrival times are equal, guess what....  need I say more