Audessey, SPL Meter and time on my hands.

Discussion in 'AV Receivers & Amplifiers' started by Member 96948, Mar 19, 2006.

  1. Member 96948

    Member 96948
    Distinguished Member

    Joined:
    Nov 18, 2005
    Messages:
    10,238
    Products Owned:
    1
    Products Wanted:
    1
    Trophy Points:
    166
    Location:
    The Fen Edge
    Ratings:
    +2,253
    I posted this on the Monolith DF thread, but for me at least, it raises more questions on the Audessey front which has more relevance here.

    Had a bit of spare time this afternoon so I got the SPL meter out and recorded some 1/6 octave test tones off the BFD Snapbug site. Sorry the graph is a bit poo, but I've only got Appleworks and I couldn't convince it to do a log scale on the X axis, so I labled it up in Photoshop as best I could.

    It should be noted that I have all of my speakers set to small and 80Hz crossover. The 3806 was set to -15dB. The SPL meter was set to C weighting and slow. It has raised a few questions.

    People have complained that Audessey Room EQ makes things subjectively quieter which this hasty graph seems to back up. I assumed it was because it ironed out room induced peaks in response, but over the subwoofer generated part of the scale, apart from the 20-25Hz peak, it just seems to make everything 6 to 10dB quieter without significantly effecting the curve profile.

    There is a notable dip in response centered on 80Hz which is the amplifier defined crossover frequency. The subs crossover is set to maximum/bypass so all bass management is performed by the 3806. All of my speakers are large enough to reach well below 80Hz so there is no premature roll-off on their part. Therefore, I believe, I cannot tune this out.

    The sub only has about 10 hours use so is still about 90 hours short of being run in. Never the less, the Audessey EQ should compensate for this. I will re run the sweeps a couple of weeks from now, having re-EQ'd it. If Audessey is consistent, I would expect similar results.

    Have a look, pick the bones out of it and please feel free to comment.

    Russell

    [​IMG]
     
  2. Gary_W

    Gary_W
    Active Member

    Joined:
    Jun 6, 2005
    Messages:
    450
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    18
    Ratings:
    +21
    Hi Russ,

    Hope you're enjoying your Monolith :thumbsup:

    Using the SPL meter with the Snapbug tones was a real disappointment to me, as it also appeared that Audyssey did nothing whatsoever apart from turn everything down by 3 or 4 dB.

    The only reason I knew it was doing a lot more that that was that it sounded a lot better with it on, and normally when you turn things down it sounds worse :D

    Denon tech support over didn't have the answers I needed, so I wrote to the CTO of Audyssey to ask what gives. Long story short is that (according to him) an RS SPL meter plus Snapbug tones are a long way from being a real world listening test as they create standing waves that can make problems look a lot worse than they are. According to the company, correcting the room with EQ based on those makes little sense. The other point is that there is a limit to what Audyssey can do as far as correction goes (+9dB, -12dB). This range is inferior to the BFD, but has a little more boost capability than the SMS-1 (now watch the Velodyne fans slap me :hiya: )

    The rest of this (very long) email is the correspondance between us. Hope it helps you. I am in blue, he is in red:-

    Hi

    I just bought the Denon 3806 product purely on the strength of your technology being embedded within it; I have read many articles on the technology and thought I had a reasonable understanding of matters, but I’m finding myself very confused.

    I have Acoustic Energy speakers and an SVS PB-12 Ultra sub. My problem with my setup is that the room had a huge spike at 50Hz, which I was using a parametric EQ to tame. I was able to measure this with a cheap and cheerful RS sound meter and a range of test tones from 160Hz down to 16Hz. Calming down my room was the reason that your technology appealed to me.

    Upon getting the Denon, I ran the auto setup with the maximum 6 positions. I am obviously no longer using the external EQ device.

    The Audyssey MultEQ really does seem to work according to my ears which is what matters at the end of the day. However, according to the sound level meter it is doing nothing for the curve whatsoever; I still have a huge peak at 50Hz, a huge trough at 30Hz and a pretty uneven response elsewhere. In short, the response after applying the Audyssey curve is about a 3 or 4 dB cut at all frequencies, but the curves seem to follow each other almost exactly. I can send you the curves if you’re interested.

    Is there any way you can explain to me why this is? I figured that I should get a fairly flat curve with the Audyssey on but it isn’t the case. It is obviously doing a lot more than the SPL meter is telling me (my ears tell me that), and I wonder if you would be good enough to help me understand my strange results?

    Many thanks


    Hi Gary,

    I hope I can clear up some of the confusion. The first thing to know is that MultEQ has built-in limits to the amount of correction it performs at any given frequency. So, the maximum boost that it will perform is 9 dB and the maximum cut is 12 dB. There are many reasons for this that have to do with maintaining the proper headroom, not clipping, etc.

    In your room, it is entirely possible that the peak at 50 Hz is larger than 12 dB. But here is the tricky part: The peak at 50 Hz (or at any frequency for that matter) is highly dependent on where it is measured. MultEQ tries to create a filter for each loudspeaker that smoothes out the response that is measured at multiple positions that encapsulate the listening area. The 50 Hz peak at each of those positions is very different.

    One could argue (and many have) that it would be best to use a single position for the mic and just remove the peak at that position. It turns out that this is not useful unless the listener's head is the size of the mic capsule (0.25"). Also, doing that would cause such problems at other parts of the room that the overall sound quality would degrade. In fact, many of the single position EQ systems suffer from that problem.

    The other problem comes from the way one measures the frequency response. I understand that from a consumer standpoint using an RS SPL meter is fun and easy, but the results you are getting are not at all correlated to the actual frequency response in your room. One big reason for that is that you are likely using a sine wave tone to measure each frequency. By definition a sine wave will suffer greatly from standing wave problems and give you a measurement that is not representative of what a real signal would be doing in the room. That's the reason we use a special sinusoidal sweep that covers all frequencies in a prescribed way. Then a transformation is applied to the collected data to identify what is happening at each frequency that is representative of what a real-world signal would suffer from in your room. The best way to verify the performance of room correction is to use a real-time analyzer that performs both time and spatial averaging. The problem is that these are quite expensive and difficult to use.

    I hope this helps with some of your questions. Please contact me directly if you have others and I will be happy to chat with you some more.

    Best regards,


    I then replied to Chris, and here is his post back in response to my further questions:-

    Hi Gary,


    From what you have said, I think I need to re-think my methods. Using the RS meter, a 50Hz sinewave in my room needed a 22 dB cut to tame it, achieved by a Behringer Feedback destroyer connected to my old receiver. The BFD really did help a great deal, but your explanation makes me doubt that the problem is quite as bad as my testing made out. There is certainly a problem at this frequency at my listening position as playing any piece of pop music will tell my ears. Audyssey does a wonderful job of sorting this out, but I obviously want it to have the best possible chance of improving my room.

    Wow, 22 dB is indeed quite a peak. I suspect that at least two dimensions of your room must be very similar to be causing such a strong standing wave.

    If I could prove a consistent problem across my 6 listening positions that exceeds 12dB at 50Hz, then I could actually use the PEQ on the SVS; it can do a 12 dB cut. As such, I could help out at the problem frequency and then run the Audyssey EQ as I’ve got a total of 24dB worth of cut if needed. If I could prove such a problem at this frequency, would the Audyssey EQ find it helpful for me to apply the cut on the SVS or am I best to leave it to its own devices?

    Yes, additional cut at 50 Hz by the PEQ would help if you indeed have such a big peak. It would be tricky to find how narrow to make the bandwidth of the parametric filter on the BFD. That part is impossible to do with an SPL meter and can only show with a plot of the frequency response in your room.

    If you feel it would help, I have the dilemma of how to evaluate the severity of the 50Hz problem. Would connecting the RS meter to free software like Room Wizard (which does give a sweep) be helpful, or would I be better to trust my ears than the low end test gear I have?

    I am not familiar with Room Wizard. If you can feed the output of the RS meter to the software then it would be useful to measure the response that way. The RS mic is not calibrated for frequency response measurements and is too big (0.5") to be accurate at frequencies above 1 kHz. But for looking at 50 Hz it will be fine. Let me know what you find.


    And that is about as far as I got. I've had a bit of a play with RoomWizard, but nothing conclusive (dire quality cables that aren't long enough to reach the listening position and lack of time when being that noisey is possible).

    What I have done is to bung the BFD in as well. With the old amp, I had a huge trough at 30Hz and a huge peak at 50Hz. What I did was cut absolutely everything so that the bottom of the trough became the new baseline then turned the sub up a bit. I put this BFD curve onto the 3806 then ran the Audyssey EQ. It is better than the 3806 with Audyssey alone, which surprises me. But I think in my room the response is that bad that it needs all the help it can get...

    Not sure how to address your 80Hz problem, but consider Room Wizard to get a closer idea as to what the Audyssey does; your trough may not be quite as bad as Snapbug makes out...

    Gary
     
  3. Member 96948

    Member 96948
    Distinguished Member

    Joined:
    Nov 18, 2005
    Messages:
    10,238
    Products Owned:
    1
    Products Wanted:
    1
    Trophy Points:
    166
    Location:
    The Fen Edge
    Ratings:
    +2,253
    Hi Gary,

    I think I remember seeing you post this transcript before and it is indeed interesting reading. I note that you also noted the apparent cut across the bass frequencies.

    Having had the chance to run a movie or two this afternoon I applied the trusted 'ignore the measurements and trust your ears method' and actually ended up turning the sub DOWN by 3dB. I understand that Audessey is far more than a cut and boost device and that a sine wave is a bit of a blunt tool, but the booming during, for instance, War Of The Worlds when the first machine rises from the street was unbearable. It took me right back to the days of my Kef PSW2000. The problem I was suffering is precisely standing waves set up by prolonged bass frequencies.

    I think, therefore, I may, as you have, invest in a BFD. I had ignored them until now, but given how cheap they are, it seems a small price to pay. Question is, there are two. Which one is the one to go for? There are a few DSP1124s for sale on Ebay at the moment but none of the apparently better (newer?) FBQ2496s.

    Finally, after reading your post I wondered that, if Audessey was limited to the amount of cut and boost it could apply, that the same may also be true of phase. It throws up a warning if you wire a speaker out of phase so it's ability to compensate to any large degree may be limited. Having thought a bit harder, the 80Hz dip I am suffering may be cured simply by me turning the phase knob on the sub. After all it is centered on exactly the crossover frequency. As it appears to be such a problem over such a small bandwidth, it should be a matter of 10 minutes work with 3 test tones and the SPL meter to sort it out. I'll let you know.

    Many thanks for your help again.

    Russell
     

Share This Page

Loading...