TOPIC:

Hi folks,

Apologies if this has been covered before, I had a cursory look through and could not find answers so hopefully someone can help.

I am running a DDRC24 for the front left and right channels in my car. I have performed one successful Dirac measurement set and generated a filter in slot 1 which I like, but on modifying the target, generating a new filter and storing it in a spare config slot, and on subsequent measurements, there is some unexpected behaviour.

1 - The subsequent filters (modified target curve, same measurement set) are attenuated (overall) compared to the first one - by maybe 6 dB. I have not changed the gain on any output channels or overall gain setting.

2 - When starting a new measurement set, the delay and attenuation for inputs set by the previous filter are still audibly and visibly (in the response histogram) on. How do I clear them, as I don't want to correct on top of a previous correction.

Cheers! -Rick

Hi Rick,

1 > Are your subsequent targets about -6db overall? This would explain.

2 > You mean they are audible with the sweeps?
I don't think Dirac keeps any filter activated during measurements.
Anyway you can just close/restart the program for new measurements to be sure.

Hi Grosso,

Thanks for the reply, I had another session measuring today and the issues are straightened out. Coming from the desktop application, there is a different workflow to generating filters and transferring them to the DSP.

Definitely a crucial step to unload any filter already in the slot and saving the configuration before attempting a new measurement set and transferring it over, in hindsight this makes sense but I was expecting the measurement sweeps to always bypass any active delays or filtering. Not so apparently.

I have now taken 3 measurement sets and made three filters with the same response target to move the sweet spot from the driver's seat, the front passenger's seat and a general setting for the centre of the cabin all accessed via the remote. Very cool! This was exactly how I had hoped it would work!

Indeed that's strange! I don't recall having performed multiple measurements in a row within the same Dirac session. So you had to delete/remove each filter of its slot, and delete each already measured points to restart?
I may test but I'd be curious to get the devteam's feedback.

Btw, what kind of shape did you use for chair/sofa etc? Because I understood that’s in in car too.

Hi Grosso,

Yes, if I recall correctly I closed and relaunched the Dirac Live utility, and I intended to take a new series of measurements because I re-evaluated where I wanted the sweet spot. However, I did not connect to the unit via the MiniDSP utility and switch to a config file with no Dirac filters active, as I assumed I could continue to measure and Dirac would bypass the filters when measuring. I noticed as soon as I started though that the channel attenuation was still active and presumably the EQ/delays.

Just a learning curve I guess!

I have used the 'chair' mode, I did not find any measurement mode for specific use in car, but I did read about this somewhere. Any idea if it is there?

Cheers, Rick

Hello Rick,

Ok so it’s still unclear, on my side yesterday I’ve made 3 sets of measurements again (head/chair/sofa), while closing/restarting Dirac between each set.
I tried at the beginning to just delete all points one by one after the first filter creation for the next set, so without closing the program, but it was just too long and it’s not something I’ll do later anyway.
So the results: filters were bypassed for sure.
For my 3 sets, all with already some active filters in slots, the resulting FR showed my typical pre-Dirac response.

So I don't know what happened on your side.
As you said, just connecting the minidsp utility could change the attenuation level by itself depending of the setting, but it's not that.
Maybe it’s just the dirac output level that was not exactly the same?
Or did you touch your HU/following dsp volume control by inadvertence?

About the measurement method, there isn't anything specific for us in caraudio.
Even non official, it’s only trial, errors, and shared experience.
You can find some good informations in Andy’s threads ("denver8me4dinner" here and « oabeieo » on Diyma), or in my build thread on Diyma too).
I'm now using most of the time some variations of the chair method (one seat tune only).
But even after a year I'm still trying many different mic placements, from super small around the head to the entire volume of the cabin.
The good thing is that most of them can sound good to very good, I'm just chasing the best.