A Test. How much Voltage (power) do your speakers need?

[Pano]

Did what? Set a -18dB RMS pink noise signal to 85dB SPL?

[AndrewT]

Originally Posted by Pano

..............I will freely admit that I should have been more clear about using dynamic recordings in the first posts. My mistake. I will go back and edit that so as to be more clear...........

Not a mistake.
Simply that the test as proposed is flawed.
It does not predict how little power we need to avoid clipping when we play our systems at the loudest we want to hear at.

[JRKO]

Unless Pano was planning on publishing a White Paper, can he simply add a rider and be done with it

[Pano]

Originally Posted by AndrewT

Simply that the test as proposed is flawed.
It does not predict how little power we need to avoid clipping when we play our systems at the loudest we want to hear at.

Andrew, this is so completely wrong that I'm surprised at you.
What don't you understand about the test? All your concerns have been met and explained by me and by others.

Until you can demonstrate, point by point, how the test is flawed, I must consider your responses irreverent.

[Pano]

For those of you who own an SPL meter - or want to buy one - here is some pink noise for you. It is recorded at -18dBFS RMS. There are two files. One is mono, the other stereo. In the stereo file, left and right are uncorrelated, I.E. they are random compared to each other. Each is 11 seconds long.

I suggest playing them in loop, one after the other. Move your SPL meter around the listening area and look for an average 85dB reading. Your meter should be set to Slow and C weighting.

This will set your system to the SMPTE PR-200 standard level. The EBU uses the same level. Movie soundtracks are supposed to be mastered at this level. This should give you 18dB of headroom over the average level. or peaks of 103dB, if your amp and speakers can do it.

I think you will find that this is much too loud for normal CD listening, as many CDs are mastered much hotter than this. The peaks won't be higher, they are still at 0dB on the recording, or 103dB SPL at your chair. It's the average level of most CDs that will be higher. But that's OK, it may be close to your "As Loud as I Ever Want to Go" setting found by ear. I doubt you'll ever run louder than this, but please tell us if you do.

[Pano]

FWIW, I just tried the above test of setting my system level at 85dB. It came up right at 12:00 on the volume knob. Surprise, surprise. 12:00 was my by ear "plenty loud enough" setting.
I do find it much too loud for most listening. Checking at an enjoyable listening level I find the input signal is 8-12dB lower on average.

[GM]

Originally Posted by 5th element

Thanks for your response GM this is actually a very good example of what this test isn't about and can hopefully serve as a good example for those who perhaps don't quite get it.........

You're welcome! Thanks for the thread summary, I don’t have time these days to do more than a quick scan of long threads.

The test does seem a reasonable alternative to the time honored SPL measurement way to determine a 70V distributed system’s power requirements in today’s digital ‘world’, but it's what it claims to do that concerns me, i.e. provide the information required for someone to determine how much power they need ‘worst case’ if they don't want to clip the amp with whatever digital sources they currently have on hand [or may buy in the future, not to mention some analog recordings] plus I see no way to use it for setting up a proper HT and/or multi-channel music system, so don’t want folks to be mislead into believing this procedure will necessarily do what is being claimed.

I mean how does one know how much their existing amp and/or speakers is probably clipping doing this test unless it’s already sufficiently oversized, especially if it’s a tube amp and/or ‘FR’ driver? Or how dynamic the signal?

Folks tend to set max levels based on audible distortion, so any conclusion seems predicated on a test that may be ill suited for the app since the more dynamic it is, the more it’s likely to be clipped, so it may be just less clipped with the calculated power available unless the normal average is somewhat, to well, below the test setting.

Worse, the most dynamic CDs I have don’t have the most power hungry spikes in its loudest passages, so how is some arbitrarily chosen 120 or 220 Hz test tone going to account for these spikes?

Indeed, the only way I can [or at least once, could] hear them was by noticing more top end ‘air’ around the instruments after seeing the spikes using an SLM, i.e. one step closer to what I’ve experienced in concert halls, piano bars, etc., so how likely is one to notice such subtlety using this test unless their amp, speaker system was already grossly oversized [at least by the tone of this thread] and know what to listen for?

105 dB peak/channel at the Lp is a goal for those folks that want to reproduce movie soundtracks at DD/THX/DTS reference. This is based on an 85 dB average at the LP. Since the gain is fixed at 20 dB for the mains and 30 dB for the LFE channel, it’s scaleable to whatever average or peak one desires.

If not sitting up close, it also happens to be in the ballpark for the piano bars, jazz clubs, small concert venues or similar I once frequented before mind numbing ‘loud’ at the expense of SQ became a prerequisite, so has been my personal reference level for long before these standards were proposed, though of course the dynamics of live venues are somewhat higher than all but a precious few recordings I’ve owned/auditioned.

Not that I haven’t enjoyed some of the loudest concerts on record to the point of ~mimicking them at home with the required highly efficient, high powered sound system required that might not have been adequate based on my understanding of this test if done today. As TD noted:

“How much Voltage does it take to reproduce the variety of things I like to listen to in their dynamic entirety?

The answer to me seems that it depends on the dynamic peak to average ratio in the music and the dynamic nonlinearity of the speakers and not just 4X the .5VRms margin.
That margin IS plenty for a lot of recordings, especially modern pop but well short of others, way short of reproducing real life events in a normal noise floor.”

All that said, if it’s a high output impedance tube amp, I’m ambivalent about clipping the really fast transients as the amp’s harmonic distortion is normally perceived as adding top end ‘air’ regardless, so this test should be adequate for finding its reasonably low distortion power minimum or conversely, the speaker’s efficiency to meet a specific average SPL.

GM

[Pano]

Sorry GM, you simply don't understand the test. All the issues you mention above have been covered. I've tried to explain it, but it seems to be out of my reach to do so.
It's super simple and easy to do.

The ONLY variable is where you set your volume control. I've asked for folks to do that using dynamic recordings and have now even supplied pink noise for an 85dB reference.
I just don't know how much simpler I can make it.

I'll go over your points 1 by 1 in a few minutes.

[kevinkr]

Several off topic and personal posts have been deleted. Please let's focus on the discussion and not each other.. Thanks..

[Mooly]

Test is done, apologies for the rotten pics... it was so bleedin' loud ya know

And that's with a fresh set of earplugs well fitted lol

So I did the test, you can see the SPL meter here, C weighted. Very little difference between the mono and stereo tracks.

I then went back and ran Panos 220 Hz track at this same volume setting and measured 3.7 volts RMS.

Keeping that volume setting I played the Eroica track that I posted "Cooledit" shots of earlier. That peaked at around 93 ish db. And it was to loud for comfortable listening for me at that.

The Mooly Amp has power meters accurately calibrated in volts onto 8 ohms and so I guestimate the Pano noise test at 0.5 w RMS. Clear shot of the meters so you can see the scale.