Best Paid S/W DSP POLL

[rcohen]

I really don't like doing that. It reminds me of all of those phony Hifi magazine reviews. Fortunately, there's no need to try. Anyone can test for free.

Aw.

 

Some flowery words would help motivate me to find the time to try it.

 

Well, did any benefits show up in measurements?

[dallasjustice]

It's really hard to compare because DIRAC and Acourate have very different measurement approaches. Both ultimately use impulse response correction and a form of inversion. The big difference has to do with how the measurement is "prepared" for correction. In DIRAC, multi point measurements are "combined" together in prep for inversion. In Acourate only one synchronous sweep is needed per channel, which is very convenient. However, the signal is prepared using Uli's frequency dependent windowing. I don't fully understand either approach. However, my stupid brain thinks that each method used in " preparation" for inversion is trying to do the same thing. That is, each method attempts to correct what can be corrected and not more. Each method works well so there's no need to battle it out here. Bottom line: it's unlikely measurements are going to reveal why one sounds better than the other.

 

Michael.

 

Aw.

 

Some flowery words would help motivate me to find the time to try it.

 

Well, did any benefits show up in measurements?

[sdolezalek]

All I know is that Acourate sounds MUCH better in my system.

 

You lost me there. Why are you doing a survey if you have already concluded that Acourate sounds the best in your system? Is this a marketing pitch for Acourate or is there a real question here?

[dallasjustice]

You lost me there. Why are you doing a survey if you have already concluded that Acourate sounds the best in your system? Is this a marketing pitch for Acourate or is there a real question here?

 

The poll is closed now. I DID a poll because I wanted to read about other experiences folks have had comparing these softwares. So far, there have been some excellent comments. Everyone is free to express their opinions and experiences. I plead guilty to having strong opinions based on real experience.

[rcohen]

It's really hard to compare because DIRAC and Acourate have very different measurement approaches. Both ultimately use impulse response correction and a form of inversion. The big difference has to do with how the measurement is "prepared" for correction. In DIRAC, multi point measurements are "combined" together in prep for inversion. In Acourate only one synchronous sweep is needed per channel, which is very convenient. However, the signal is prepared using Uli's frequency dependent windowing. I don't fully understand either approach. However, my stupid brain thinks that each method used in " preparation" for inversion is trying to do the same thing. That is, each method attempts to correct what can be corrected and not more. Each method works well so there's no need to battle it out here. Bottom line: it's unlikely measurements are going to reveal why one sounds better than the other.

 

Michael.

Interesting. It would be fun to read about the details. Are details available somewhere?

[dallasjustice]

Interesting. It would be fun to read about the details. Are details available somewhere?

 

Room Correction (Acourate/Audiolense/Dirac/Other)

[rcohen]

Room Correction (Acourate/Audiolense/Dirac/Other)

Thanks!

[dallasjustice]

Thanks!

 

That's thread is great because some really smart folks posted. It gets really interesting after reply #70. :-)

[Kal Rubinson]

BM reroutes signals between speakers. If you do that after the EQ, the rerouting sends some EQ-ed signals to other speakers.

[rcohen]

BM reroutes signals between speakers. If you do that after the EQ, the rerouting sends some EQ-ed signals to other speakers.

Understood, but as long as they are time aligned at the crossover frequency, and they measure right at the listening position, why is this important?

 

My understanding is that Dirac is intended to be used this way, similar to to speaker crossovers (whether active or passive).

 

I could do bass management within JRiver, rather than in my pre-amp. I'm not sure this would guarantee phase alignment at the crossover frequency.

 

Also, it's convenient to draw the summed target curve in Dirac.

 

Of course, Acourate and Audiolense XO are designed with this in mind.

[3ll3d00d]

Some flowery words would help motivate me to find the time to try it.

 

Well, did any benefits show up in measurements?

I've written a few flowery words before on the subject so I'll share those I am definitely an evangelist for acourate because it worked wonders in my system.

 

Measuring the results of convolution - Home Theater Forum and Systems - HomeTheaterShack.com

 

To give some context, my system is not remotely esoteric. It's HTPC -> firewire -> RME FireFace 800 -> 5.1 analogue -> Marantz AV7005 -> cinepro & speakerpower poweramps -> MK 150mk2 series + custom 15" sub and the room is my lounge/cinema, completely untreated except for normal furnishings and typical Victorian UK construction (solid walls, raised wooden floor). I've made a series of changes over the last few years; some incremental (e.g. power amp upgrades, switching to an analogue path through the AV7005), some step changes (all new speakers, new sub). Acourate firmly falls into the latter category, getting this right was as big an upgrade as new speakers were over my previous set.

 

To give some less flowery thoughts;

 

- acourate locks the soundstage into place, no acourate = a stereo image firmly shifted to one side and voices/instruments that tend to shift within the soundstage (by frequency), listening to a LEDR demonstrates this quite clearly

- it provides a completely immersive surround soundstage, all channels feel perfectly balanced

 

This is backed up by the data. You can see my room in a post in that thread on the jriver board & the results of my correction

 

Room Correction (Acourate/Audiolense/Dirac/Other)

 

Meanwhile this post shows the prime seat alone

 

acourate: an audio toolbox | Page 2 | AVForums

 

The surround and centre channels fit on top of this v closely so that the all channels are +/- 2dB in the prime seat across the full frequency range. The average across the seating area is not that far outside this either, at least above 100Hz anyway which is to be expected as I only have 1 sub.

 

re the "mixed phase" point. Acourate is v flexible and lets you do all sorts of things, the basic approach though allows you to;

 

- apply arbitrary IIR filters that get combined with the final correction

- do near field correction/alignment of a single channel (including XOs where necessary) as distinct from the room correction

- determine how to apply frequency dependent windowing to the recorded response (this allows the window applied to the signal to be measured in terms of cycles not time and hence the window gets shorter as the frequency increases thus serving to gate out later reflections .. the practical impact is that it influences the shape of the response that you aim to correct, i.e. you either do broader brush correction or end up with a much more jagged, fine resolution correction that will be more position dependent ... one thought is that you can use your room's RT60 to decide on the window size though I suspect this work betters in a treated room, it certainly doesn't work in my untreated one)

- decide on the target response

- correct amplitude to that target (trimming the LF and HF extremities if required)

- correct phase to optimise the step response (this involves iterating over the strength of the phase correction & the compensating pre ringing correction until you achieve a response that is stable & maximises the similarity of the stereo pair of channels)

 

The one thing I will say is that you need to be prepared to invest time and effort into this, it's by no means an easy piece of software to learn how to use but it is extremely powerful. This is especially true if you run a multichannel system. For example, it took me months to come up with this workflow

 

acourate: an audio toolbox | Page 2 | AVForums

 

Personally I found it worth the effort for both the SQ improvement and the knowledge gained in the process.

[rcohen]

Thanks for the details!

I haven't tried Acourate or Audiolense XO, yet, but I have tried Dirac Live, Audyssey XT32, Audyssey Pro XT32, and some older flavors of Audyssey.

 

Compared to older flavors of Audyssey, Audyssey XT32 was the first time I preferred room correction over none.

 

The Pro kit gave me better sounding results. The ability to disable midrange compensation really helped in my system, and it also helped to be able to select my favorite among their preset curves. On the other hand, the process of taking measurements was really, really slow. The curve editor was so bad it was almost useless. Also, bugs cause problems when trying to reuse measurements, so fine tuning the target curve was impractical.

 

Switching to Dirac, I immediately liked the sound better. It removed a lot of muddiness. In the past, every time I've tried to fine tune sound with EQ, I felt it would introduce something unnatural and muddy in exchange for getting the right frequency balance. That tradeoff doesn't seem to happen with Dirac. Also, the target curve iteration process is wonderful. It's fast, easy, and fun to make small or large tweaks to the target curve, and A/B test until you zero in on a favorite. Small changes can make a big difference. The only downside about this process is that one's perception of sound changes as you listen, so the curve that sounds best at the end of one session often sounds off at the beginning of the next session. It's easy to end up chasing your tail.

 

I've used Dirac in two rooms, now (theater and bedroom), and in both cases, it makes a very large improvement. The imaging locks in. The room disappears (kind of like putting on headphones). Voice intelligibility is much better. The tonal balance matches my taste, which is to say natural sounding, impactful bass, clear but not fatiguing treble.

 

Comparing my two setups, my theater has Triad Platinums (got them used) + 4 Rythmik F25s. My bedroom has a couple Adam Audio ARTist 6s (smallish active speakers with AMT tweeters). Dirac certainly can't make those speakers sound the same. The theater setup is a much bigger more dynamic sound. The Adams are more transparent, with sweeter extended treble, but nowhere close to the bass and dynamics of the theater. Both rooms could benefit from more acoustic treatment, but Dirac really reduces those problems. One big difference is that the theater is symmetric, but with my bedroom, the speakers are off center, and my listening position is off center in the other direction. Dirac made a huge difference in compensating for that (time alignment, levels, timbre matching, helping with early reflections).

 

Here's my current target curve for my theater. I did run into some trouble with clipping when I added the bass boost, since JRiver was sending it a 100% signal. I was able to avoid that by boosting the gain on my subs above flat prior to taking Dirac measurements. Turing down the output level on JRiver probably would have also worked.

 

Incidentally, according to the REW room sim, heavily treating my front wall would be amazing for reducing the jaggedness in the un-EQ'ed bass response.

 

[3ll3d00d]

I think we can all agree that good EQ is well worth the entrance fee I think it would be v interesting to hear your thoughts on a listening session based on your existing target curve & the acourate demo.

 

Why do you roll off so sharply above 12kHz btw?

[rcohen]

I think we can all agree that good EQ is well worth the entrance fee I think it would be v interesting to hear your thoughts on a listening session based on your existing target curve & the acourate demo.

 

Why do you roll off so sharply above 12kHz btw?

 

Since I was taking a 90 degree measurement (necessary for a surround setup), it's hard to be sure of how accurate the measurement is at the upper limits, so it might not be as steep as it appears.

 

When I added more treble, though, it starts sounding harsh. It doesn't sound like any treble is missing, though. The Adams do have a sweeter, more extended treble. It might be cloth dome vs. AMT, or it might be something with the room. Probably a mix of both.

[rcohen]

When thinking about how much rolloff there really is, it's helpful to to look at the distance between the two curves, not just the shape of the target curve.

[rcohen]

Here's my bedroom curve, with the Adams. They are actually more similar than I realized. Maybe this is just my preference. The Adams are naturally flatter for the last octave. Also, you can clearly see the LR differences from the asymmetric layout.

 

I forgot to mention, in the prior graph, the Triads are crossed with the subs at 100hz.

[rcohen]

I think we can all agree that good EQ is well worth the entrance fee

Totally.

 

One thing I noticed is that it seems to bulldoze subtle system differences, like DACs, pre-amps, and amplifiers (assuming nothing is wrong). I could sometimes hear differences before, but after good EQ/impulse correction, they seem to be below the line of audibility.

 

I haven't tried these, but it would be interesting to do experiments to see if the differences between ported vs. sealed, servo vs. non-servo subs, and active vs. passive crossovers are still audible after this type of EQ.

 

I can definitely still hear the differences between speakers and rooms. Definitely less than before, but those are still above the line.

[Kal Rubinson]

Understood, but as long as they are time aligned at the crossover frequency, and they measure right at the listening position, why is this important?

But that is impossible. How do you get a proper correction filter for the rerouted main channel bass when it is applied only to that channel but, instead, rerouted to the sub?

[rcohen]

But that is impossible. How do you get a proper correction filter for the rerouted main channel bass when it is applied only to that channel but, instead, rerouted to the sub?

The sweep measurements and playback goes through the same bass management crossover.

The corrected bass is what is rerouted to the sub, and summed with the independently corrected LFE channel.

If they sum correctly at the listening position, isn't that what matters?

 

Perhaps that introduces the possibility of phase differences for sounds panned between two speakers?

 

When Dirac sets different delays on each speakers, I guess that could also cause problems with BM.

[wgscott]

A different way of putting it is there is no need to do bass management after you do room correction, and doing so could actually screw up the room correction.

[3ll3d00d]

Bass management and room correction are not related except that they concern the same frequency range (and can be optimised if done together).

 

This repeats the AVS thread on this subject but I think it is definitely suboptimal but likely to be perfectly workable in practice (especially if you know what you are doing). It feels like a case of perfect is the enemy of good tbh (though that might be the very definition of our interest!)

[3ll3d00d]

Since I was taking a 90 degree measurement (necessary for a surround setup), it's hard to be sure of how accurate the measurement is at the upper limits, so it might not be as steep as it appears.

 

When I added more treble, though, it starts sounding harsh. It doesn't sound like any treble is missing, though. The Adams do have a sweeter, more extended treble. It might be cloth dome vs. AMT, or it might be something with the room. Probably a mix of both.

I think what I was thinking is that you are chasing the HF rolloff by treating the 12kHz peak as the anomaly rather than the dip just before it. Your ears rule though, I am just looking at a graph

[rcohen]

I think what I was thinking is that you are chasing the HF rolloff by treating the 12kHz peak as the anomaly rather than the dip just before it. Your ears rule though, I am just looking at a graph

If you imagine the whole line up 5 db higher, think about it as running through the middle of the natural curve of the speakers/mic/whatever, not trying to connect the valleys. I just preferred a sloped frequency response. Maybe that will change as I add more sound treatment.

[rcohen]

I guess I'm going to have to try doing bass management in JRiver.

[dallasjustice]

I've noticed the same thing. But it's hard to explain these phenomena. I wonder why I used to care so much about power cables. :-)

 

Great DSP makes one re-evaluate everything else.

 

Totally.

 

One thing I noticed is that it seems to bulldoze subtle system differences, like DACs, pre-amps, and amplifiers (assuming nothing is wrong). I could sometimes hear differences before, but after good EQ/impulse correction, they seem to be below the line of audibility.

 

I haven't tried these, but it would be interesting to do experiments to see if the differences between ported vs. sealed, servo vs. non-servo subs, and active vs. passive crossovers are still audible after this type of EQ.

 

I can definitely still hear the differences between speakers and rooms. Definitely less than before, but those are still above the line.