Dirac, iTunes, Audirvana -- what am I actually listening to?

[OldAudioNut]

Hello. I’ve been an avid reader of this site, but is my first post. I’m been a 2-channel stereo neurotic for 40 years or so. I’m pretty happy with what I have at the moment, but am always trying to get better sound. My listening room is very difficult. First, it doubles as my wife’s living room, so I’m somewhat limited in speaker placement and use of passive room treatments. Also, a cathedral ceiling and stairway to an open loft leaves lots of asymmetry in the layout. But, I suspect I’m not alone in these challenges, with anyone reading this post likely having a similar situation.

 

So, I’ve been thinking about DSP lately. As I still have a lot of physical media in SACDs, I’m thinking that my solution ultimately will be one of the analog boxes such as offered by DEQX or Dirac. I don’t wish to plunk down my money without an extensive audition. Since I use a Mac Mini and Audirvana for computer playback, and since Dirac is offering a 14 day trial of the Dirac Live, I thought I’d mess around with Dirac to get a hint of what DSP can do.

 

I’ve gotten Dirac to work. I suppose the calibration was probably not optimal, as I used the basic Behringer microphone. But it’s close enough for what I need. I’ve created 3-4 different filters, and am listening closely. I’m pretty sure it’s working, as I can toggle the “Filter On/Off” button on DAP and hear quite distinct differences. At the moment, though, I’m not exactly sure what I’m actually listening to.

 

It all seems to work fine. I tell System/Preferences/Sound to use Dirac. I open DAP. Then I open Audirvana, which is set to iTunes integrated mode, using in Direct Mode. Using iTunes, I can get tracks to play. DAP indicates that it is “Streaming”, and toggling the Filter On/Off changes the sound. Audirvana also appears to be working, and the DAC is showing the appropriate resolution and bit rate. If I choose a 96/24 file, DAP still streams, Audirvana still works and the DAC shows the right reading. If I choose a 192/24, or DSD file, DAP goes gray with filter off, and everything else works fine.

 

Here’s my confusion. I don’t actually think it should be working this well. My impression, from this forum and other sources, is that Audirvana and Dirac do not play well together. So I have some questions:

1) What am I listening to?

a. DSP’d digital played through the Audirvana music engine, sent over to the DAC?

b. DSP’d digital played through DAP music engine, sent over to the DAC? In this scenario, maybe Audirvana is just tracking along, and kicks in only when the DAP can’t handle a high rez file.

c. Something else? Listing to iTunes music engine?

 

2) What is happening when I toggle the DAP filter to “Off”:

a. Still listening through Dirac with no settings

b. Bypassing Dirac, listening to Audirvana music engine

c. Bypassing Dirac, listening to iTunes music engine

 

I appreciate any insights this forum can provide. I’m hesitant to form any opinions about DSP in general, or Dirac in particular, until I actually know what I’m evaluating.

 

Howard

[flak]

Hello Howard,

 

I'm not using Audirvana myself (I'm a Windows user) but I'll try to answer your questions anyhow...

 

1) Most likely, the sound is being played back by the Audirvana audio engine and then it is being sent to DAP, then to the DAC. It sounds like it works well and there is no resampling going on in OS X because the DAC shows the correct sample rate.

 

2) With DAP in "Off" mode, the signal chain is exactly the same except the DAP is not applying the filters to the audio stream. So I guess a) is the answer here.

 

Also the Behr ECM8000 is perfectly suitable for the purpose but you should either have its individual calibration file (there are significant differences in high frequency response from unit to unit) or you should at least use its generic calibration file but limit the correction to the low and mid frequencies (you can do that by moving the DLCT right curtain leftwise)

 

Ciao, Flavio

[OldAudioNut]

Hi Flavio

 

thanks for your reply, it is helpful to know that I' listening to the DAP music engine, not Audirvana.

 

I'm hearing pluses and minuses with the Dirac. On the plus side, I'm getting better, richer harmonics especially in the mid-bass area. It does seem to shift the tonal balances downwards, probably because it was elevating low bass in the 20 - 60 hz region. Also, the sound picture became more focused and coherent. For example, on vocals i got more of a sense of a unified center image, rather than an image constructed from 2 speakers. I'm pretty sure I'm hearing more things in the mix.

 

On the minus side, the dynamics were gone. Rim shots, bass drum thwacks, guitar plucks all lost their bite. It seemed almost muffled, as if a dampening mist of electronica had been sprayed over everything. Also, my reaction to the DSP image is analogous to how I feel about 3D movies: the hyper-realism is fascinating at first but I quickly become fatigued. The sound just felt artificial, the illusion of naturalness was gone.

 

I'm thinking that the DAP music engine is not as good as Audirvana which may account for some of dynamics and artificiality issues.I believe someone else in this forum has mentioned this, and it might improve things when Audirvana and Dirac learn to play nicely together. Also, I know that the Eidolon's get their imaging and sound via a pretty sophisticated cross-over design; i have a feeling that they don't like DSP messing around with them. System synergies are important at this level.

 

So, there was enough good stuff that i'll keep my eye on DSP, but it's not yet for me. Thanks for taking the time to reply to me, and thanks also to Dirac for making the trial available.

 

Howard

[EuroDriver]

I have just been playing with Dirac on the miniDSP platform. I was unhappy with the first filter I created, but quite satisfied with the second filter where I reduced the high frequency window down to 500 Hz for one speaker, and 1,000 Hz for the other speaker.

 

This second filter seems to have all of the benefits, with little or no perceived loss of the sonic goodies which are important to my old ears.

 

have you tried reducing the frequency window ?

[OldAudioNut]

I did play with filters, actually tried a few different things. On reducing frequency window, I tried one at 700 and also 250Hz. I found the 250 was ok in that it was more listenable in the bass, but the imaging magic and deeper window into the mix was gone. I also created a flat filter from 25 - 3000Hz and allowed it to roll-off after that, actually i kinda liked this one the best but again I didn't think it sounds natural. Probably ok for heavily processed studio mix music, not so good for acoustic or live performances.

 

I haven't created different filters for each channel, that might be interesting.

[EuroDriver]

Can you post screen shots or a link to your frequency plots ?

[OldAudioNut]

Actually ... no, I can't post a screen shot. I tried but failed to get a screen shot. I'm running the Mac Mini headless via Tight VNC. Turns out that 1) keyboard manipulation didn't work; 2) Grab facility didn't work; 3) Preview facility didn't work. Must be that the Mac requires a video monitor to be active to get screen shot.

 

Do you suppose there is a file in dirac that I could access and post the link? All i see is the *.dprj file under the Dirac\Projects folder.

[EuroDriver]

What hardware are you running Tight VNC on ?

 

Let's see if Flavio will take a look at your *.dprj file

[OldAudioNut]

Well, scratch my last post, i figured out a way to upload a screenshot of my frequency plot. Kinda small image, though.

 

[KnockKnock]

Hi OldAudioNut

 

If I were you, I´d tweak the frequency a bit. Try to follow the measurements, that in your case would be making a dip at 700 Hz and 3 kHz, roll off from 5 kHz and upwards, make a boost at 500 Hz. Just follow the measurement, and level out the small ups and downs. Works life a charm at my place, with Audirvana+, iTunes and Dirac ;=)

[flak]

Hi OldAudioNut

 

If I were you, I´d tweak the frequency a bit. Try to follow the measurements, that in your case would be making a dip at 700 Hz and 3 kHz, roll off from 5 kHz and upwards, make a boost at 500 Hz. Just follow the measurement, and level out the small ups and downs. Works life a charm at my place, with Audirvana+, iTunes and Dirac ;=)

 

These are interesting remarks... i.e. not everybody knows about the BBC dip but if you google for "BBC dip" you'll find several discussions about it.

 

In general several manufacturers apply some kind of "voicing" so that their speaker's frequency response is intentionally modified... as a result when somebody has selected a specific speaker model among many others it is possible that he will prefer a general behaviour at the mid/high frequencies that is similar to the "before correction" one.

 

There is nothing wrong with that but if that is the case the KnockKnock suggestion is the one to follow... you will anyhow benefit from a full bandwidth smoother response and impulse response correction.

 

An alternative is to limit the region of intervention as Eurodriver is doing, OldAudioNut's frequency response at the higher frequencies before correction is reasonably close to the target curve so he may prefer this option.

 

Flavio

 

P.S. As far as "dynamics" are concerned it is most important that the DSP gain has been properly adjusted, this is an often overlooked aspect

[KnockKnock]

OldAudioNut (nice nick!;=)

 

As your left and right speaker has nearly the same measurement, this is a VERY good starting point. So don't tweak them separately ;=)

 

Another EQ hint: when i look at your sreenshot, I´d try to cut every thing below 45 Hz (there is not much going on there anyway), raise the bass energy between 50 and 65 Hz quite a bit (8 dB is a good starting point). You before mention 250 Hz, that frequency is the most silly ever invented! You can successfully make a narrow dip there, and get a more precise all-round sound image. 250 Hz is similar to putting your hand before your mouth while you speak, it just sounds boomy.

 

Enjoy

 

Best Regards

KnockKnock

[OldAudioNut]

KnockKnock

 

I do have some bass traps and absorption panels up, i think that helps with the left/right speaker similarity. I tried your suggestion about fitting the curve to the speaker response, that certainly made the voicing much more similar to unfiltered. I also tried moving the low filter cutoff up to 45Hz, that didn't seem to make much difference. I'll keep playing with these ideas, as I have a couple days more on the trial. This stuff is fascinating.

 

Flavio: I do think I notice that the volume is slightly lower when the filter is applied compared with when DAP is shut down. I used the gain slider to increase gain a bit on the filter. I also tried adjusting the preamp volume control up a notch with DAP engaged. I didn't use a SPL meter so I'm sure I was imprecise at matching levels. These narrowed the dynamics differences a bit, but I can still hear the difference.

 

To me the major improvement of the DSP is a tighter focus, mainly on the center image, especially female voice. The major drawback seems to be that i can hear the extra layer of control, the music doesn't seem to breathe as well. The big benefit I'm gaining from this exercise is knowing that my starting point is pretty good, specifically there are no real bad problems with elevated low freq room nodes which kinda corroborates my sense of no boom in the room. The little bits of deviation from flat in the higher freqs may be some interaction of speaker voicing and my difficult room.

 

My conclusion may turn out to be: leave well enough alone.

[KnockKnock]

Old Nut

That's very good, I would stay with the 45 Hz cut, as you hardly notice it, why bother to put to much strength to the speakers - for no reason enjoy