"Official" Audyssey thread (FAQ in post #1) - Page 1976

What matters is the total power being delivered to the speaker driver. When an amplifier clips, it's usually because its input signal voltage is so large that the amp is delivering its maximum possible power long before and after the input signal gets to its peak. I wouldn't expect square waves at very small signal levels to cause problems in most cases.

One can describe a clipped signal (or square wave) as the sum of a large number of sine waves of varying amplitudes and frequencies. (That's called Fourier analysis.) When compared to a pure sine wave at the same frequency as the square wave, the square wave contains many additional high-amplitude, high-frequency sine waves. Saying that those high frequencies damage speaker drivers is just another way of describing the same situation. People often forget to mention that it's their amplitudes which matter, though.

Is it true that most blown speakers are due to clipping of inadequate amplifiers than from too much undistorted power being delivered to them? That has always been my belief.

When I listen to hi-res multichannel music at high volumes - and it seems easier to do that because it is so clean - some material causes clipping. My ears tell me it is my amp beginning to clip, and this causes me to consider bigger amps.

Jeff

For anyone who wants to learn about active and passive loudspeakers, and the differences between them, here are a couple of straightforward articles:

 

http://en.wikipedia.org/wiki/Powered_speakers#Differences_between_passive.2C_powered_and_active_speakers

 

http://www.enjoythemusic.com/Magazine/manufacture/0403/index.html

 

As one of the articles points out - "active speakers may have greater fidelity, less intermodulation distortion (IMD), higher dynamic range and greater output sound pressure level (SPL) with fewer blown drivers."  

 

I often wonder if I should have gone for the active versions of my own M&K S150s LCR set instead of the passive versions I have. If I was starting from scratch I would, but as I have good amps I decided to save the money and chose the passive designs. There is no denying that the benefits of an optimised coupling between matched amplifier and loudspeaker that is found in an active system is the best solution - and when the active speaker has separate amps for each driver set (as most do), even better still.

True, a low power square wave will not harm a speaker but when clipping a reasonably powered receiver/amplifier it would by definition be a powerful signal, some tweeters will burn out while others will tolerate it. Thermal overload is also the cause of many speaker failures. I know my mid-woofers are rated at 80 watts for 8 hours at 80Hz. I of course am unlikely to drive them at such levels especially since they are crossed over from the sub at 80Hz. Since this is the Audyssey thread, I should mention that when it is run it sets the crossover to 40Hz which is probably the -3 db point for my speakers but they cannot tolerate that excursion well and are subject to thermal overload at that frequency.

 

Or, Jeff, the active versions of your M&K S150s (see my post above).  I am seriously tempted... best of all, Mrs Keith wouldn't even notice as they look the same :) She might spot the gaps in the rack where the amps used to be, but I could easily explain that away. LOL.... 

 

EDIT: Jeff - the second article I link to above explains very well the increased output (typically 6dB) that you could get by 'going active'.... I wish I hadn’t started this train of thought My experience of me is that when I get these ideas...

Right, I think it's clear that the theoretical problems being discussed would only be an issue when you are playing LOUD.

- you can damage your speakers with an unclipped signal if it's too loud
- you can damage your speakers with a clipped signal if it's too loud

The specific question, which seems to be getting lost in the shuffle here, is whether a large voltage output at the SW pre-out, created by the combination of loud volumes and a high positive trim, can potentially pose a threat to damage the subwoofer. There seem to be two hypotheses as to what could cause the harm:

1. The output signal itself clips, before it even gets to subwoofer amp (this is what was measured with an oscilloscope by the poster I quoted)
2. The voltage output at the SW pre out is too large for the input sensitivity of the sub's amp

It seems to me in both cases that the end result would be a clipped signal is amplified and sent to the woofer itself. Which of course can destroy the driver, just like overdriving it with a non-clipped signal. So it seems to me that the operating question is whether it's true that -- given the same actual volume -- you are more likely to damage the speaker with the clipped signal.

Another area to consider with Active speakers is the XO. They do not have passive XO's, instead they are electronic with DSP. The benefit - you can do very steep slopes (100db/Octave or more) which let's individual drivers operate more comfortably within their specs. You can also XO lower with such a steep slope (for the tweeter) and higher for the for the midrange - getting a much better blend between drivers. This is one if the reasons most recording studios use active speakers.

Most electronic crossovers are limited to 48db/octave slopes which is still very steep. I use 48db/octave between my mid-woofers and tweeters. It certainly does reduce the load on the tweeter and they are crossed over at 2kHz which is lower than recommended.

Theresa,

Thanks for the correction - typo (meant approaching 50db/octave). I've worked in some studios that used active Genelecs - they sounded pretty amazing (also heard quite a few Meridians). All out of my personal budget though...

The point I was trying to make was the optimization of XO and Amps used for each driver and the optimal blend that can be had with a good active speaker.

So to sum it up with a specific example:

Situation 1 = sub amp gain is low so you get a high sub trim (+10 or whatever) in the receiver
Situation 2 = sub amp gain is higher, so you get a lower sub trim (-5 or whatever)

Now you crank the volume up loud and blast War of the Worlds or whatever insane LFE movie you desire. You hit some heavy LFE action where the sub is being asked to output 115db in order to hit the proper SPL at the listening position.

In situation 2, the signal doesn't clip. In situation 1, the voltage of the sub pre out exceeds the input sensitivity of the sub's amp and thus the signal clips.

So same SPL required, but are you more likely to damage the sub in situation 1?

I've often wished to hear Meridian equipment, their totally out of my budget though. When I came across miniDSPs several years ago I knew I had to try them. I've been very pleased with them and in active crossing over in general.

Very good point.

So what speakers are you using and do you by-pass the passive XO's to get true active crossover/amplification? Sorry, I know we are going way OT from this thread now...

^^^ Forgot to add, do you find the miniDSP still beneficial even when calibration with XT32?

bp,

Me thinks no for situation 1, because even though the trim set by Audyssey is high and still delivers an un-clipped signal, in the next stage the gain control of the sub input attenuates that signal because now its set to low. Let's remember the gain knob is usually a potentiometer that can only decrease the input signal, just like in this example. Whaddaya think?

On another note, it would also be interesting to discover why the avr makers allow Audyssey a +12 dB upper limit? Even +15 dB for Onkyos. Anyone?
Edited by mogorf - 1/20/13 at 9:47am

 

Stop it right now!  You are tempting me... :)  I am not good at resisting temptation.... I have the ex-wives to prove it ;)

 

It might be worth remembering too that some subs have inbuilt protection against being overdriven. (Eg SVS). Whether that works to protect against clipping, IDK.

Sorry to bring this back, but all this Congo talk reminds me of a line from a famous Dire Straits song

Also a certain cab ride that Lebowski took after a rough night.

 

 

I assume it is because they have to make sure that their AVR works with an almost infinite range of subs, from different manufacturers, all with different specs. As they cannot possibly, obviously, test their AVR with every sub, they have to cater for a wide range of eventualities.

 

Remember, nobody has said this will definitely happen to your sub if you set a trim level of +11dB. It is a possibility. Whether the sub is damaged will depend on the input sensitivity of the sub in relation to the output voltage of the AVR. Because this is an unknown combination, this is why I say that it is always saftest to set a trim range of approx +/- 3.5dB. The user then knows for sure he will be in a 'safe' range.  That is not to say that +11dB will not be safe too, but it may not. By erring on the side of caution, safety is assured. 

 

Other than academic interest, I cannot see why this has become such a big issue - a crusade even - when I am sure that you, like everyone else who has commented, would agree that a range of +/- 3.5dB cannot do any harm, so there is no good reason not to use it.

 

OTOH:

 

• Setting a trim of -11dB may cause problems with the auto-on feature not switching the sub on if it has gone to sleep,
• A trim at either extreme, say -11dB or +11dB also allows for no user preference to set the sub a little hotter or cooler (and this would need an avoidable recalibration if this was desired after some days or weeks of listening, as is often the case)
• And finally (phew!) a trim setting of +11dB MIGHT cause a problem with overdriving the sub amp input, which COULD be dangerous and is therefore best avoided.

 

Meanwhile, there are NO obvious benefits at all to be had from setting trim levels near the extremes of their ranges in normal circumstances and setups, so it can't really be the best advice to do so. 

Keith, all very true.

 

Thanks, Theresa. I appreciate your comment. Sometimes I think I have gone mad and have started speaking Chinese without realising it :)

I would love to get some thoughts from someone like Ed Mullen or another sub mfgr. It does seem like most tend to recommd a combo of lower sub trim at the AVR leading to some more gain at the sub amp, but it would be interesting to see the hyopthetical "danger" validated (or not) by someone who does it for a living.

IMO the "best" setting is as low as possible while still consistently triggering the auto on circuit of your sub, and still allowing enough room for adjusting to taste.

I think most go the external amp route because they start on that path with an integrated amp/receiver. Through all of the upgrading, I have never upgraded amp and speakers simultaneously. If that is true for others, then there's an impediment. But the active speakers companies don't seem to be expending much on getting out the story.

Jeff

Exactly, it is the average power that needs to be converted into mechanical motion with a lot being dissipated as heat. Fail to dissipate the heat and voicecoils overheat and windings lift and/or burn out. At any given peak power output, the average is higher with clipping present.

 

+1 to all that. I aim for (and usually achieve) -5.0dB. This let me go as much as 3dB hot (the most I want) and still be 'comfortably negative'.

 

I can’t recall upgrading speaker and amp at the same time either. I think the active speaker companies see their market as pro sound studios which is why they ignore us aficionados. 

Sit #1 - The heavy LFE content causes the receiver/processor to run out of DIGITAL HEADROOM and compresses the loudest parts of the signal; is that the same as clipping? I don't think it would cause audible distortion, but does it, too, raise the average power and cause problems downstream?

I don't know, just asking ....

Jeff

There is no headroom issue in the digital path. All these gain issues happen after the DAC.

I recently did an 8pt XT32 calibration using my Integra DHC-80.3.  Here's the room configuration with LCR behind an AT screen, sub in right corner behind AT GOM FR701 mesh and 4 surrounds in ceiling (I know...inherited from previous owner!):

 

This is view with the left wall and back corner wall removed.  The two grey locations on the right wall are door openings (back right is the stairwell and the front right leads to the unfinished section of the basement):

 

 

Dimensions are for furniture layout.  Current configuration is "similar" but not exactly the same.  Having moved a bunch of furniture around during the crawl, I can say I didn't see major differences in the sub 100Hz region with different configurations:

 

 

I did some measuring with my OM kit (track 2, monophonic short sine sweep) and the before (red) and after (blue) Audyssey results are shown below with mic at ear level and centered in room approx. 13' from the screen:

 

 

Noticing the rather large suckout at 85dB, I attempted my first sub crawl today.  The sub is currently front corner right in the room.  The results of each corner as well as mid points along each wall and even the middle of room even though I know that's not feasible are shown below:

 

 

 

 

 

 

 

 

I'd like some help interpreting these results as based on my limited experience with measuring and interpreting, it doesn't appear as though there is a slam dunk in any of these graphs?!  I hope I'm wrong!  Thanks.

 

 

Thanks, Roger. I thought I had read that it ... "headroom" i.e. running out of bits in this digital context ... could be an issue under certain circumstances. Guess not.

Jeff
Edited by pepar - 1/20/13 at 1:53pm