Home Theater Forum and Systems banner

21 - 40 of 122 Posts

·
Registered
Joined
·
85 Posts
The "after" looks much better. If you have the ability, it would be good to see the full frequency response.

And the general concensus on dealing with dips is to move the sub or the listening position, or both, rather than throwing a lot of amp power at it with a positive filter.
 

·
Super Moderator
Joined
·
9,240 Posts

Actually, any equalizing requires more amplifier power. Even if you have only a single 10 dB peak, that’s what your overall sub level was based on. Cut the peak and you’ll find your sub is now not as loud as before. So, you increase its level to make up for the lost volume. Voila, now you’re driving your sub harder than you were before – at all frequencies except for the one you cut...

If you were dealing only with a dip in response, assuming it wasn’t a null and could be equalized, you’d only be driving the amp harder at the frequency that’s boosted.

So there’s no free lunch. Any equalizing requires ample headroom going in. In Moonfly’s “After” graph we can see that overall sub level has been increased by several dB. That amounts to increased demand on the amplifier. Equalizing raised the lowest point of the dip from ~62 dB to ~67. So it may respond to additional equalization. Can’t hurt to try.

Regards,
Wayne
 

·
Registered
Joined
·
1,384 Posts
Well, Wayne. That depends a bit on what he used to set the level with in the first place. If he usen an SPL meter, then yes. If he used a tool like REW or Audyssey, then the peak would be above the general level of the sub, and thus would 'save' headroom.

Not saying you are wrong, just saying there are other circumstances that need to be identified.
 

·
Elite Shackster
Joined
·
7,142 Posts
I love a good debate :D (dont we really need a dedicated discussion thread?)

Right, both graphs are obviously produced with REW. Both are measured at the 75db reference level. The sub is in one of only 2 available positions in my room, and is in the better of the two. All the other available room positions would require a smaller sub,a dn have been tested but dont offer a better response. In short the place it is, is as good a spot as any in my room, and there are no better available. The space it is in is large enough to accommodate a decent sized sub, so Ive taken advantage of it with the DIY sub Ive built, and subsequently measured.

As noted, Audyssey seems to have boosted the output of the sub by about 6db-9db, then trimmed the peaks. Audyssey actually ping'd the sub 3 time for each sweep, which I understand is it testing the subs threshold. However, it wont boost much more than 9db anyway, and as the boost seems to be across the entire range I assumed its limited the amount of boost to what the lower end, or majority of the response needs. This to me means that while the dip (caused by the room, every sub Ive tested has this same dip) has been boosted, there is probably a bit more room to add a bit more with the BFD Ive recently acquired, beyond what Audyssey wants to.

All in all I'm happy with what Audyssey has done,the sound is very good, and every time I use it I get consistent and pleasing results. Nothing short of fantastic for a free feature, given the next cheapest auto eq solution is about £250 for me.

Currently my sub is back in bits as I want to get the cabinet finished. When I get it back together, I'll apply the BFD based on the above post Audyssey graph, then re-run Audyssey and see how it comes out. I doubt I'll be disappointed, but I'll update the build thread (see sig) when I finally get there.
 

·
Super Moderator
Joined
·
9,240 Posts

Well, Wayne. That depends a bit on what he used to set the level with in the first place. If he usen an SPL meter, then yes. If he used a tool like REW or Audyssey, then the peak would be above the general level of the sub, and thus would 'save' headroom.
Could you explain further? That doesn’t make much sense to me.

Regards,
Wayne
 

·
Registered
Joined
·
1,384 Posts

Actually, any equalizing requires more amplifier power. Even if you have only a single 10 dB peak, that’s what your overall sub level was based on. Cut the peak and you’ll find your sub is now not as loud as before. So, you increase its level to make up for the lost volume. Voila, now you’re driving your sub harder than you were before – at all frequencies except for the one you cut...


I have bolded the part I am talking about. I agree, IF you use an SPL meter and noise to set the levels. But if you use a calibrated microphone and REW, or Audyssey for that matter, the peak may be above the average level of the sub, and require ONLY a cut to be in line with the rest. In that case there is no additional headroom needen, you have 'saved' a little headroom. Am I making sense?
 

·
Super Moderator
Joined
·
9,240 Posts
the peak may be above the average level of the sub, and require ONLY a cut to be in line with the rest.
I don’t think anyone sets their sub level (before equalization) based on a response average.

My premise has nothing to do with SPL meters, Audyssey, etc. It’s based on the assumption (right or wrong :D) that if there is a peak in the room, then before adding equalization, people will have adjusted their sub level based on that “hot” frequency. If they had set the level based on the response average, they would ultimately perceive the sub to be too loud and turn it down. I don’t think anyone’s going to steadfastly endure an overpowering subwoofer just because “that’s what the measurement says is right.” :huh:

Consider this: Assuming there is a peak in the room – The peak is acoustically induced, so it is “free” gain. That translates to “free” headroom, since the peak means that the sub’s level has been set lower than it would be otherwise. Eliminating the peak means that the “free” gain now has to be made up elsewhere. That means the “free” headroom will be lost. Make sense?

Basically, it’s all about the sub’s “before vs. after” gain setting. It doesn’t matter if you cut the peak electronically (via equalization) or acoustically (by relocating the sub). If "after" ultimately means the sub’s level is increased, then additional amplifier power is required.

Regards,
Wayne
 

·
Registered
Joined
·
2,431 Posts
I don’t think anyone sets their sub level (before equalization) based on a response average.

My premise has nothing to do with SPL meters, Audyssey, etc. It’s based on the assumption (right or wrong :D) that if there is a peak in the room, then before adding equalization, people will have adjusted their sub level based on that “hot” frequency. If they had set the level based on the response average, they would ultimately perceive the sub to be too loud and turn it down.

Regards,
Wayne
I agree. That's the way I perceive the issue.

I always end up calibrating the sub to the mains about 3 or 4 times as the process of dialing things in takes place. The first one is rough and like you mentioned usually has the largest peak or 2 dominating the calibration level. After you get your placement/phase/ crossover to the mains dialed in more you may have significantly smoothed things already and then your sub is a bit low, so re calibrate. From there perhaps you add a second sub or even a couple and dial them in if you are using a multi sub approach (you should be) at which point you may need to recalibrate overall levels again. At that point you might be EQing any remaining peaks down and you guessed it...time to calibrate levels again. After all of that it's time for Audyssey to bat clean-up (which I have yet to use.)
 

·
Registered
Joined
·
678 Posts
The following graphs were measured on a Denon AVR-989 running MultEQ XT with a calibrated mic.

To make the level setting clearer, I am posting separate graphs of the sub & left and sub & right curves. In each graph, light green is Audyssey Off, blue is Audyssey Reference, and the red is Audyssey Flat. 1/3 octave smoothing was used in each full range graph.

Left:
100221.leftaudysseyoffonflat.jpg

Right:
100221.rightaudysseyoffonflat.jpg

One can see where the Audyssey Reference curve introduces its "BBC dip" near 2kHz, and tails off above 10kHz.

I should explain the unusual sub curve with Audyssey Off. I deliberately configured my Hsu VTF sub in its overdamped mode, only one port open but the high-pass-filter set to two ports open. I found that, after Audyssey equalization, this gave the flattest response in my room.

The room is not treated, most of what dispersion there is comes from numerous bookcases.

Bill
 

·
Registered
Joined
·
85 Posts
Bill -

Thanks for keeping the thread alive. Those Audyssey curves are very respectable overall, and exceptional in the bass. All I would probably do it bump up the sub output by a few dB.

The difference in the high treble between your left and right speaker measurements is a bit odd though. Are your speakers set-up assemetrically? Or did the mic move bewteen readings?
 

·
Registered
Joined
·
678 Posts
The sub output appears level when compared to a single front speaker. When I measure the two mains together with the sub, the curve looks like more typical Audyssey curves. The sub level then appears +3dB higher than the mains; I spent several weekends trying to understand the level difference. Since late 2008, Audyssey includes a DynEQ function that I leave on, except for REW measurements, and that gives a low end boost similar to a house curve, which may be what you were suggesting.

You have correctly identified the next major area of improvement. The speakers are reasonably symmetric to the axis from the TV, but the whole arrangment is offset in the room. The left front is near the corner, whereas the right front is mid wall. So the reflections are very different, with a fireplace on the left, but the wall on the right 8 feet away. Also, there is a couch on the right parallel to the right hand wall, and its far arm blocks some of the sound from the right speaker. I've already broached the idea of re-arranging the room with my wife, but it will take some convincing.

I did not move the mic between the measurements, but it is very difficult to measure from exactly the same point as was used for the Audyssey setup. Also, in my asymmetric environment, it is not surprising that Audyssey, evaluating 8 points around the bubble, calculated different filters for the two fronts.

Thanks,
Bill
 

·
Registered
Joined
·
85 Posts
Hey Bill -

I've got the same speaker placement constraints, with the front stage scrunched over to one side. I've been using manual PEQ for the past year based on an average of 4 mic readings. You may find that if you tighten up the radius of your mic positions, making sure to keep all of them between the speakers, you could get more consistency in the HF between the L/R.
 

·
Registered
Joined
·
2,431 Posts
Audyssey actually ping'd the sub 3 time for each sweep, which I understand is it testing the subs threshold. .
Is this what it is doing? I just ran Audyssey for the first time a few days ago and it swept each speaker 10 times for every measurement. They all seemed to be the same volume to me and not all that loud.
 

·
Registered
Joined
·
3,069 Posts
I have just re done my graphs again tonight and have used less filters and less cuts which in turn is giving me more between 30 and 80hz. The first is my graph after Audyssey and no EQ (no BFD) which I think Audyssey has done a good job. I don't have the graph prior to Audyssey as I didn't save my measurements but its pretty good all the same. The second is my Sub only with Audyssey and with the BFD filters applied. And the last is my sub and speakers with Audyssey and BFD.

To re cap:

1. Sub only with Audyssey and no BFD

2. Sub only with Audyssey and BFD

3. Sub and speakers with Audyssey and BFD.

Sub crossover: 80hz Speakers: 80hz

I have also increased the sub trim level to +5 and reduced the EP4000 gain down to 24db (it was on max). I am running the sub to about 82db so slightly hot.
 

Attachments

·
Elite Shackster
Joined
·
7,142 Posts
Is this what it is doing? I just ran Audyssey for the first time a few days ago and it swept each speaker 10 times for every measurement. They all seemed to be the same volume to me and not all that loud.
I'm sure somewhere I read that Audyssey does this for the sub channel (not sure on the speakers), and each time it pinged the sub it increased the gain. Looking at the before and after graphs of my sub, the low end had been boosted just so to give me a nice flat response, so I know it certainly pushed my sub more than any other I had owned previously.

An e-mail to Chris at Audyssey might be in order though as the easiest way to get it cleared up.
 

·
Registered
Joined
·
2,431 Posts
Here is what I'm looking at currently with Audyssey. I will be tweaking this more and re-doing them soon but you can see the cutting and then overall level boost that Audyssey is making to the sub channel. My averaged response after Audyssey is looking good through the bass range, but I need to make some improvements to the main listening position still.



Electrical response comparing Audyssey on and off. 6.5db boost.




Same as above. Electrical with Audyssey levels readjusted to show cuts.




Main position before Audyssey



After Audyssey





6 position average response before Audyssey




6 position average response after Audyssey.
 

·
Premium Member
Joined
·
27 Posts
New owner of a Denon AVR1610. I used REW a few years ago to set up a home recording studio, and now I'm trying to set up my family room television and sound systems. I'm trying to run REW on a Windows 7 computer, and there are some issues, which I'm sure have been explored on the REW forum. I was able to get it to calibrate everything, and run normally (as far as I could tell). The results are attached.

The first graph shows triplicate runs of REW after Audyssey setup was run and Multi-Eq was on. I used 1/12 Octave smoothing to get rid of some of the jitter but keep the smaller peaks and valleys. I was not able to set the axes to exactly 15-200 Hz and 45-105 dB, but got close. All were run with a 75 db target level.

The second graph shows 2 runs with Multi Eq turned off. I was surprised by the small amount of difference other than the generally lower sound level. Also, unlike the graphs with Mulit Eq on, these don't replicate very well. Both graphs contain what seems to me to be a similar degree of peaks and valleys.

The third and last graph shows all of the data on one plot.

I would appreciate comments on the results, and on whether I am doing things improperly (as far as you can determine from my post).

Thanks in advance for the guidance.
 

Attachments

21 - 40 of 122 Posts
Top