Sound United U 30 Update (Denon X3800H / X4800H / A1H, Marantz Cinema 50 / 40 / AV 10)

[multisport4me]

This for me is where "subjective" wins over "objective". If too me it sounds better, I'm going with it. Even if the facts don't agree. I've been in several home theaters where the bass is just overwhelming and covers over everything else. But that's how they like it.

Exactly. Don't get me wrong, I love empirical measurements but room correction gets you in the ball park - your ears still matter to deliver what you like in your room. I love how guys will spend days and weeks to get some measurement from pretty much ANY room correction out there then crank the subs up 6+dB because the bass sounds weak to them.

Same goes with the measurements that Amir or Gene does. I love it. I think its important that they do it. But in the end, nobody on this forum can hear a difference of 4dB SINAD. I just don't buy it. To me, the measurements tell me that the engineers got the product in the ball park - some better than others - but in the end, its what my ears like - not an analyzer.

 

[chych7]

Would like to see some time-domain plots for Audysey vs. Dirac, including RT60, clarity, and GD. Also the target curves really need to be matched for a fair comparison; setting upDirac with an individually calibrated mic vs. Audyssey with the batch calibrated stock mic is bound to create target/tonality differences (there's the ACM-1X/MultEQ-X as well).

 

[Newman]

I love how guys will spend days and weeks to get some measurement from pretty much ANY room correction out there then crank the subs up 6+dB because the bass sounds weak to them.

Hang on. There is a right way to do that. The Harman Target curve specifies a range of bass boost per a curve. Flat is not within the range. +6 dB is (just).

So, that’s not about people ignoring measurement guidelines.

 

[stupidvillager]

Hang on. There is a right way to do that. The Harman Target curve specifies a range of bass boost per a curve. Flat is not within the range. +6 dB is (just).

So, that’s not about people ignoring measurement guidelines.

Isn't the Harman curve just based on subjective analysis of what sounds the best to try to find a "one size fits all" curve.

 

[Newman]

Isn't the Harman curve just based on subjective analysis of what sounds the best to try to find a "one size fits all" curve.

Isn't the Harman curve just based on subjective analysis of what sounds the best to try to find a "one size fits all" curve.

Fixed it for you.

 

[Vacceo]

Exactly. Don't get me wrong, I love empirical measurements but room correction gets you in the ball park - your ears still matter to deliver what you like in your room. I love how guys will spend days and weeks to get some measurement from pretty much ANY room correction out there then crank the subs up 6+dB because the bass sounds weak to them.

Same goes with the measurements that Amir or Gene does. I love it. I think its important that they do it. But in the end, nobody on this forum can hear a difference of 4dB SINAD. I just don't buy it. To me, the measurements tell me that the engineers got the product in the ball park - some better than others - but in the end, its what my ears like - not an analyzer.

Even in that case, measuring allows you to get an idea of what you like, so of you want to dial, you know what you're looking for.

 

[Anttikt]

Long time reader but first post


Yes that's definitely a visible difference and may increase the "fun" when listening while being non linear. But there's something more totally unmentioned: Impulse correction. Dirac does it, Audyssey not, or maybe only as a sideeffect. So this may be the reason why Dirac sound's better to him, too.

My X4800H arrived at saturday but didn't have any time to install it, yet. Hope to do it this weekend and then I will compare Audyssey and Dirac, too - with my ears and with my UMIK-2 (calibrated by CSL).

One question: Do you use the 0°, 45° or 90° calibration when calibrating setups with height or ceiling speakers? 0° doesn't make sense since you could only point at the fronts. But 90° doesn't make sense, either, because this will point directly to the ceiling speakers. My thought is to use the 45° file and let the microphone point 45° upwards. This still has an angle error when looking at the front heights (the angle will be something between 0-45° in reality, depending on the room and the distance of the listening position in relation to the speaker position) and also when looking at the ceiling speakers at the back (maybe >45°). But I think this is still the best. What do you think? Best approach or fallacy?

Audyssey has been doing impulse correction always. Where did you get information that it doesn't?

https://audyssey.zendesk.com/hc/en-us/articles/212347483-How-does-MultEQ-apply-room-correction-

Audyssey's week point remains their mic. Now they do offer the calibrated one but it's no Umik. It still has the really poor pattern especially at negative elevation. Which leads to poor correction on center channel (which is the most important one) if it's placed lower than ear height and user doesn't understand to tilt the mic forward for better measurement result.

Only way to use Umik with Audyssey is do measurement and PEQ filter calculations with REW+Umik. Then import those filters to MultEQ-X and it will convert those PEQ filters to Audyssey FIR filters. Just I think the timedomain correction is lost here.

 

[peng]

Only way to use Umik with Audyssey is do measurement and PEQ filter calculations with REW+Umik. Then import those filters to MultEQ-X and it will convert those PEQ filters to Audyssey FIR filters. Just I think the timedomain correction is lost here.

I think I might have found another way without spending $200 on the X, that is, to use the Editor App ($20) with Rat., to create a non flat (irregular shape) target curve based on REW measurements for corrections. It would be like using a manual feedback system for error correction. I discussed that with Audyssey at length on this concept and they never agreed nor disagreed with it. At the final conversation the gentlemen asked me to do a couple more experiments and send the results, but I ran out of time and did not follow up. I only mention this because you seem to know the nitty gritty of Audyssey. There are so many misconception about Audyssey in general, many from self proclaim experts on forums and Youtube, it is unreal...

 

[Flak]

Audyssey has been doing impulse correction always. Where did you get information that it doesn't?

https://audyssey.zendesk.com/hc/en-us/articles/212347483-How-does-MultEQ-apply-room-correction-

The information that Audyssey doesn't correct the impulse response comes from fully independent measurements by forumers, for example:

https://www.avsforum.com/threads/room-correction-bake-off-2018-audyssey-xt32-dirac-2-0-accueq-2018.3037658/#post-57324882

or

Dirac Live RC Suite Stereo PC version

I tried it with one Atmos movie, when I set the player to bitstream via HDMI, it worked perfectly except it downmixed to Stereo and PCM. Apparently, and unfortunately, it cannot do what you (and me too...) want it to do. So it is PCM only, not really a problem except then you won't get...

forums.audioholics.com

 

[Anttikt]

The information that Audyssey doesn't correct the impulse response comes from fully independent measurements by forumers, for example:

https://www.avsforum.com/threads/room-correction-bake-off-2018-audyssey-xt32-dirac-2-0-accueq-2018.3037658/#post-57324882

or

Dirac Live RC Suite Stereo PC version

I tried it with one Atmos movie, when I set the player to bitstream via HDMI, it worked perfectly except it downmixed to Stereo and PCM. Apparently, and unfortunately, it cannot do what you (and me too...) want it to do. So it is PCM only, not really a problem except then you won't get...

forums.audioholics.com

Thanks these were interesting. However I couldn't flnd proof that Audyssey would not br doing corrections for impulse response.

The first one where there are impulse response plots with ARC vs XT32 vs DL from AustinJerry. DL clearly does the best job. But Audyssey is better than ARC. Since there is no impulse response plot without room correction there no reference point. If Audyssey would be doing nothing the ARC would be really messing things up. If ARC is closer to the response without any correction then Audyssey is clearly tidying things up. Not as well as Dirac though. There is also no mention what are the avrs or processors used in this. I would just like to know that Audyssey is not done with Marantz that has HDAM to f** things up.

On the second one. PENG shows impulse response comparison between Audyssey and DL PC version. Audyssey are with AV8801 so Marantz processor and they are from long time ago and he can't confirm if the mic is in the same position. Again there is no without plot for reference. Dirac seems to be doing good job again. But how can you conclude that Audyssey is doing nothing?

It would be really interesting to see if someone could provide impulse response plots with one and same avr or processor without room correction, with XT32 and DL, now that it's posdible with selected D+M models. And preferably Audyssey with MultEQ-X that can do as many filters as your computer can calculate according to Audyssey.

 

[broncogr]

@Flak
I have been messing with Dirac the last few days on my Denon AVC-A1H.
I believe I have discovered a quite serious bug that also existed for the Onkyo RZ50 before it was fixed by a firmware update.
I checked the speaker distances after a Dirac calibration

Notice how they nearly all are at 20 ms ?
Then I read the troubleshooting FAQ for Dirac Live and stumbled upon
Troubleshooting: By Firmware raise the 20ms delay limit on Onkyo - Dirac Live Support - Confluence

I checked the graph for impulse response on the Dirac Live app, selecting all speakers

So, Denon, as Onkyo previously, has set the max for speaker Delay (Distance) to 20ms....
When Dirac time aligns the speakers, it asks for a larger than 20ms delay to be used.
The AVR cant set a larger delay than 20ms so sets it at 20ms which makes the distance calibration WRONG.

I have opened a ticket at the Dirac helpdesk, so hopefully they can ask Denon to raise the limit of the delay and fix the issue.
It seems logical that since the issue existed with Onkyo, Dirac upon the implementation of Dirac Live for D&M receivers, should have checked that the same issue didn't exist.
Lets hope this gets solved in a future firmware update

 

[peng]

Thanks these were interesting. However I couldn't flnd proof that Audyssey would not br doing corrections for impulse response.

The first one where there are impulse response plots with ARC vs XT32 vs DL from AustinJerry. DL clearly does the best job. But Audyssey is better than ARC. Since there is no impulse response plot without room correction there no reference point. If Audyssey would be doing nothing the ARC would be really messing things up. If ARC is closer to the response without any correction then Audyssey is clearly tidying things up. Not as well as Dirac though. There is also no mention what are the avrs or processors used in this. I would just like to know that Audyssey is not done with Marantz that has HDAM to f** things up.

On the second one. PENG shows impulse response comparison between Audyssey and DL PC version. Audyssey are with AV8801 so Marantz processor and they are from long time ago and he can't confirm if the mic is in the same position. Again there is no without plot for reference. Dirac seems to be doing good job again. But how can you conclude that Audyssey is doing nothing?

It would be really interesting to see if someone could provide impulse response plots with one and same avr or processor without room correction, with XT32 and DL, now that it's posdible with selected D+M models. And preferably Audyssey with MultEQ-X that can do as many filters as your computer can calculate according to Audyssey.

Just a couple points to clarify things a little:

- I couldn't guarantee the mic positions were exactly the same but I did use measuring tape to make sure it was dead center, every time as much as I could, and elevation would typically be about 35 o 36 inches. So, when I said several inches I was being very conservative, in reality I would say more likely within +/- 1 inch. In any case, I don't think even +/- 2 inches would affect the impulse response performance, based on results obtained from measurements done on different days.

- Dirac Live's impulse response looks consistently cleaner regardless of which pairs of speakers I used, have tried it on the LS50, R900, and Energy Veritas 2.3i, when compared with Audyssey, the KEFs are in the room where the AV8801 was and the Veritas are in the room where the AVR-X4400H used to be. Audyssey was run on the AV8801 itself because it was not compatible with the App.

- I do think Audyssey could improve impulse response, just not nearly as good as Dirac. I guess Dirac Live RC was named Dirac for a good, even obvious reason. It reminds me of the Delta function, aka Dirac Delta function.

I have also compared the Anthem's with DLBC, but again, in that case when DLBC was used, the two subs were in .1 configuration whereas when ARCG was used, they would be in .2, with ARCG's "auto phase" used.

 

[Anttikt]

Just a couple points to clarify things a little:

- I couldn't guarantee the mic positions were exactly the same but I did use measuring tape to make sure it was dead center, every time as much as I could, and elevation would typically be about 35 o 36 inches. So, when I said several inches I was being very conservative, in reality I would say more likely within +/- 1 inch. In any case, I don't think even +/- 2 inches would affect the impulse response performance, based on results obtained from measurements done on different days.

- Dirac Live's impulse response looks consistently cleaner regardless of which pairs of speakers I used, have tried it on the LS50, R900, and Energy Veritas 2.3i, when compared with Audyssey, the KEFs are in the room where the AV8801 was and the Veritas are in the room where the AVR-X4400H used to be. Audyssey was run on the AV8801 itself because it was not compatible with the App.

- I do think Audyssey could improve impulse response, just not nearly as good as Dirac. I guess Dirac Live RC was named Dirac for a good, even obvious reason. It reminds me of the Delta function, aka Dirac Delta function.

I have also compared the Anthem's with DLBC, but again, in that case when DLBC was used, the two subs were in .1 configuration whereas when ARCG was used, they would be in .2, with ARCG's "auto phase" used.

View attachment 273152

View attachment 273153

View attachment 273154

Thanks. Really good information.

I'm surprised that Anthem wasn't any better.
Dirac really seems to be the best in time domain issues. Hopefully we get to see comparison with MEQX soon.

 

[Dj7675]

Are there any audibility studies regarding better impulse response leading to better, more preferred sound?

 

[Anttikt]

Are there any audibility studies regarding better impulse response leading to better, more preferred sound?

That's the question. If we think of placing absorbers in your room. Should you have them so much as to kill all reflections. No. Well maybe you shouldn't do it with room correction either. Just because you can do it doesn't automatically mean you should. Besides absorption you usually need some amount of diffusion. In smaller room more absobtion and less diffusion and in larger room a bit more diffusion. Because you want sound to decay at even rate for the room to sound pleasant.

That's why I would like to see comparison of waterfall and decay plots and for all the channels. I have a feeling that for these modern object based surround formats it could be more important if room correction could make all of your speakers decay similarly.That plus similar frequency response and you could have real accuracy in imaging for every direction.

At least I would like to be able to tune my system like that and see how it sounds. With these room corrections you can have your own target curve for frequency response. What I would like to have is target curve for decay. Now you're stuck with the decay target that's hard coded.

Just my thouhts based on what I've read about sound reproduction from different sources. Would like to hear if someone has other thoughts.

 

[chych7]

Did my own study a while back on how Audyssey and Dirac correct impulse:

Audyssey's Next Generation of Room Correction (MultEQ-X)

We all understand that incorrect position. There is no miscommunication. Have you tried taking confirmatory measurements properly - same area, random sample points? Incorrect position, aping measure point? He's only talking about an apples to apples comparison in a specific sense. You are...

audiosciencereview.com

Dirac did look like it cleaned it up better, but Audyssey did improve it too vs. no correction.
Nevertheless, subjectively I found that I preferred Audyssey's sound in stereo-only listening, in my system (Philharmonic BMRs, well treated room, subwoofers disabled). This was after matching the target curves.

 

[jhaider]

I did not edit the target curves at all. Standard curves on both, Dirac and Audyssey. The purpose of this comparison was to find which room EQ does a better job without doing any tweaking or changes to their target curves. At least for me, Dirac did a better job in my room and sounds more natural to my ears.

Well then your results are a no brainer, given Audyssey’s default target curve is basically defective and Dirac’s default target is either based on your speakers similarly to RoomPerfect or more or less the target that did best in the room correction study by @Sean Olive, depending whether you use paddles or points.

However, your measurements of Audyssey are puzzling. They show neither the expected midrange error nor neutered bass. Are your labels switched? The speakers seem to have a bit of a sound power dip around 2kHz, so maybe that explains why the difference that is so obvious with other speakers does not show here. Otherwise green looks to me like Dirac and red like Audyssey + DEQ at low volume. Are your labels switched?

FWIW my own experience is both basic Dirac and Audyssey+iOS app can be made to give very good results. You have to fight against Audyssey much more because its defaults are pretty much all bad, but if you have put in the effort to do so then switching is a sideways move. I would prefer error-corrected Audyssey with loudness compensation over basic Dirac without loudness compensation, though someone who always listens loud may not. DLBC is another story; I can only create one circumstance where DLBC is not clearly better than basic Dirac or error-corrected Audyssey: an x.0.y channel system. As soon as one introduces subs DLBC wins IMO.

Are there any known plans to make it possible to run DIRAC + Dynamic EQ together?
If I remember right, I read somewhere that D/M will be one of the first AVR's with a function like this running together with DIRAC...
When this would come, then it would be a no-brainer for me to buy one.

If HTP-1 comes back, it has a better loudness compensation (IMO) than DynamicEQ…does Masimo offer Dolby’s new version of Dolby Volume?

 

[broncogr]

If HTP-1 comes back, it has a better loudness compensation (IMO) than DynamicEQ…does Masimo offer Dolby’s new version of Dolby Volume?

It doesn't at this point in time

 

[chych7]

However, your measurements of Audyssey are puzzling. They show neither the expected midrange error nor neutered bass. Are your labels switched? The speakers seem to have a bit of a sound power dip around 2kHz, so maybe that explains why the difference that is so obvious with other speakers does not show here. Otherwise green looks to me like Dirac and red like Audyssey + DEQ at low volume. Are your labels switched?

I was thinking the same. The FR responses make no sense vs. the default settings/targets. Dirac should have a downward slope (around -0.8 dB/dec I think), but it is flat and peaks at high freqs, which is really odd. Audyssey should have a stronger high freq roll-off and a -3 dB 2 kHz mid-range dip by default. The -3 dB shelving from 150 - 600 Hz is also kind of odd, but both show it.

 

[enricoclaudio]

I was thinking the same. The FR responses make no sense vs. the default settings/targets. Dirac should have a downward slope (around -0.8 dB/dec I think), but it is flat and peaks at high freqs, which is really odd. Audyssey should have a stronger high freq roll-off and a -3 dB 2 kHz mid-range dip by default. The -3 dB shelving from 150 - 600 Hz is also kind of odd, but both show it.

No, they are NOT switched. That's what I get in my room. For Audyssey, I have mid range compensation turned OFF in all my speakers.