Dirac: Initial Experience

[twist222]

Flavio,

 

Thanks for the plain english explanation of impulse response. Much appreciated. In my case it looks like the impulse response was improved quite a bit. It does look like there is a lot of ringing though especially about 12 ms out. Is that correct?

 

LBob,

 

It looks very much like the first reflection arriving 2 ms after the direct sound.

 

Regards

Mark

[LBob]

LBob,

 

It looks very much like the first reflection arriving 2 ms after the direct sound.

 

Regards

Mark

 

Ah yes, 2 ms. I plead old eyes. I recently replaced my sonex panels with some from GIK for esthetic reasons. Perhaps I need to check the placement again. I should have that pretty well damped. Or is this not a problem I should worry about?

[LBob]

There's a crossfade region between the gold dot and the orange dot. After the orange dot, nothing gets changed.

 

Still, if everything to the left of the orange dot gets lowered, that's the same as a spike to the right of the dot. The tilt on the default curve for speakers with extended treble past 20khz creates this condition.

 

Thanks for the explanation. I did another target curve by dragging the untreated areas. I could get rid of the right angle in the bass, but there was a tiny area around 20k that I couldn't get the program to address. New filters sounded a bit better. There was a slight "glare" in the previous one that I wasn't aware of.

 

Tomorrow I will do three more projects. So far it's been interesting. Differences and improvements, but nothing I would call "transformative." Perhaps my expectations were too high. I'm in a well treated dedicated listening room so maybe there isn't that much room for improvement. Or perhaps major improvements will occur in new projects. Or perhaps my ears are getting old... More later.

[Fitzcaraldo215]

Did you rerun the measurements when comparing the e28 and the Integra?

 

 

I typically find that the differences between DACs and amps disappear when I rerun measurements and room correction on them.

 

 

Also, sometimes the differences between microphone positions will cause a greater difference than the DACs themselves.

 

 

I assume you aren't using subs?

 

 

Actually, the Integra and the e28 were both run with the same Dirac calibration, which was done via the Integra. Dirac cannot calibrate via an e28 due to latency issues with the driver. That is supposed to be fixed eventually.

 

I think what I did was actually a fairer comparison between Integra and e28 when using Dirac. There were no mike positioning differences involved this way as there would have been with two separate calibrations. Channel level trims and delays were also identical this way. Also, having looked at a lot of measurements over the years, you just do not see frequency response differences in the 20-20k audible band for electronics like DACS, prepros, etc. They are all dead flat. In any case, the great positives for the e28 were also there to the same degree with EQ off, although the sound is much better with EQ on.

 

Yes, I have a single JL Audio sub which was used in all cases with the same xovers done in JRiver. The xover frequency values were the same as those used by Audyssey.

 

I am with you to an extent about differences caused by things like mike positioning. Like most, I am unable to reposition mikes much more precisely than within +/- a few inches. My experience is that in spite of this, the essential sound character remains pretty much the same after recalibration. Tools like Audyssey and Dirac, unlike some others, go to great lengths to spatially average, intelligently filtering out most local, single point acoustic response issues. The real problem occurs with independent measurements using a different tool post calibration where multi point spatial averaging is not done or does not use intelligence in the averaging, as with using a simple arithmetic average.

[rcohen]

Thanks for the explanation. I did another target curve by dragging the untreated areas. I could get rid of the right angle in the bass, but there was a tiny area around 20k that I couldn't get the program to address. New filters sounded a bit better. There was a slight "glare" in the previous one that I wasn't aware of.

 

Tomorrow I will do three more projects. So far it's been interesting. Differences and improvements, but nothing I would call "transformative." Perhaps my expectations were too high. I'm in a well treated dedicated listening room so maybe there isn't that much room for improvement. Or perhaps major improvements will occur in new projects. Or perhaps my ears are getting old... More later.

It makes sense that differences would be smaller if you are already close to where you want to be.

 

It can also be easy to chase your tail with psychological tricks. There have been times that I've found myself imagining differences when Dirac was actually off. My perception of the sound changes with fresh ears in a new session. Also, every recording is different, so it's not like there's a single perfect answer that applies to everything.

 

Still, I found that after multiple sessions of tweaks, I was able to dial it in to a sound that I was happy with...right as I got new speakers and had to start over.

[rcohen]

Actually, the Integra and the e28 were both run with the same Dirac calibration, which was done via the Integra. Dirac cannot calibrate via an e28 due to latency issues with the driver. That is supposed to be fixed eventually.

 

I think what I did was actually a fairer comparison between Integra and e28 when using Dirac. There were no mike positioning differences involved this way as there would have been with two separate calibrations. Channel level trims and delays were also identical this way. Also, having looked at a lot of measurements over the years, you just do not see frequency response differences in the 20-20k audible band for electronics like DACS, prepros, etc. They are all dead flat. In any case, the great positives for the e28 were also there to the same degree with EQ off, although the sound is much better with EQ on.

 

Yes, I have a single JL Audio sub which was used in all cases with the same xovers done in JRiver. The xover frequency values were the same as those used by Audyssey.

 

I am with you to an extent about differences caused by things like mike positioning. Like most, I am unable to reposition mikes much more precisely than within +/- a few inches. My experience is that in spite of this, the essential sound character remains pretty much the same after recalibration. Tools like Audyssey and Dirac, unlike some others, go to great lengths to spatially average, intelligently filtering out most local, single point acoustic response issues. The real problem occurs with independent measurements using a different tool post calibration where multi point spatial averaging is not done or does not use intelligence in the averaging, as with using a simple arithmetic average.

10-4.

 

When I measure from the speakers, I actually have often found some subtle FR differences between amps and DACs, despite the common evidence that they are all the same. These are generally small - in the <1.0 db range. Audibility is subtle and dependent on the content. The small differences are easily compensated for by room EQ software without side effects, though, assuming you remeasure through the equipment. After remeasured EQ, I can't hear the difference.

 

I'm actually experimenting with JRiver vs pre-amp bass management now. Pretty confusing actually, with some clipping problems coming up with doing it in my pre-amp, and it's tricky to get Dirac set up right with doing it in JRiver.

[twist222]

Ah yes, 2 ms. I plead old eyes. I recently replaced my sonex panels with some from GIK for esthetic reasons. Perhaps I need to check the placement again. I should have that pretty well damped. Or is this not a problem I should worry about?

 

LBob,

 

I am not an expert but if my maths is correct 2 ms equates to a difference in signal travel path length of just under 0.7m. I do not know the details and dimensions of your listening room but 0.7m would be consistent with the first reflection from the floor or depending on the height of the ceiling, possibly the reflection off the ceiling if you don't have a higher ceiling.

 

In my room I only have a rug on the floor between me and the speakers and nothing on the ceiling (well not until I figure out how to get something stuck to the ceiling past the wife's scrutiny). The rug I use is not really thick enough to absorb a wide range of frequencies. What I am trying to say is that for practical and aesthetic reasons many of us do not have effective acoustic treatments on the floor and ceiling so many of us will get significant reflections from the floor and ceiling. This is a different problem from standing waves between floor and ceiling.

 

If I understand it correctly one of the of the problems with early reflections is the effect they have on measurements of lower frequencies where the microphone receiving both direct and reflected sound at the same time.o You may have seen photos of companies undertaking speaker measurements in anechoic chambers to obtain measurements that do not include reflections. But in our case we do not have access to anechoic measurements and we are trying to measure in room response so we need to measure with some early reflections included.

 

As for how Dirac deals with early reflections I have no idea but perhaps flac may be able to help or point us in the direction of some further reading.

 

Sorry for the long post and for not being able to be more help on this one. As you may gather it is something I am really interested in but have very limited knowledge on.

 

Regards

Mark

[LBob]

Mark,

 

I have low ceilings (7 feet) and a concrete floor with very thin carpeting. Maybe a rug would help. I have the same problem with sticking something on the ceiling. WAF is hard to fight.

[flak]

As Mark pointed out 2ms is 68.8 cm. only.

The reflection is present also in the original measurements and it's important to notice that the ratio between the main impulse and this 2ms reflection is improved after correction.

It is the ratio between these, not the absolute amplitude that you see in the normalized graph that is important for the end result.

 

Ciao, Flavio

[TommyCat]

apologies if in the wrong discussion, new to CA. Has anybody has static issues creeping into the music with Dirac enabled post the Mac Yosemite upgrade? Seems to happen on a sporadic basis but what seems to trigger the issue is switching between files of differing bit depths and sample rates. Once it starts to happen, any audio source I play (A+, itunes radio, Pandora etc) the static remains. The only fix is to shut down Dirac and reload. Good for awhile but then reappears. Sent a communication to Dirac and they said they are opening a ticket (which may not go anywhere unless they can reproduce the issue). Any body else with this strange behavior?

[TommyCat]

As of this morning disabling Dirac. Borderline unusable until this issue with Yosemite is fixed. Of course might be Yosemite causing the issues. Hopefully between Dirac and Apple this can be ironed out.