Issued a challenge and need some guidance.

[jcx]

do you want to correct Bob Cordell too - if I'm wrong on this then he needs school'n too

Hi Richard,

Unfortunately, clipping happens. Unless, of course, you have VERY big amplifiers with decently efficient speakers. At RMAF 1996 Peter Smith and I demonstrated the issue using a peak/average digital power meter I designed. We played Rickie Lee Jones "Getto of My Mind" from her Flying Cowboys album in a typical hotel room at realistic (but not overly loud) levels using quality loudspeakers with 84 dB efficiency. The Denon 250 wpc amp we used clipped, while average power out of the amp per channel was less than 2W. You have to see it to believe it.

Clipping happens more often than you think. Its too bad virtually no audiophile amplifiers have honest clipping indicators.

Bottom line: don't dismiss clipping issues.

Of course, the plethora of crappy, compressed program material lulls us into complacency about clipping.

Cheers,
Bob

so crediting that peak factor and adjusting the 240 W peak power with 6 dB for the different speaker's sensitivities under consideration we're still talking 60 W for "a realistic but not annoying level in a typical-sized hotel room"

 

[Arty]

what is that 250 wpc ?

 

[GambitAudio]

Change of plans.

Challenge is off. He had to cancel his order to pay for a new transmission in his car and wants to postpone for the foreseeable future until he can save back up for the purchase.

Looks like this project's due date just got wiped clean. I guess the silver lining is this just gives me time to experiment a bit. Still open to suggestion.

 

[jcx]

could ask Bob - but does it need much interpretation if average was 2 W and he claimed it clipped on snare hits

even a heavily gamed 250 Wpc "music power" spec should hold for single millisecond or less

I believe Cordell points out elsewhere that his peak meter hardware catches much shorter peaks than any SPL meter's "fast" setting

 

[radiosmuck]

do you want to correct Bob Cordell too - if I'm wrong on this then he needs school'n too
Originally Posted by Bob Cordell View Post
Hi Richard,

Unfortunately, clipping happens. Unless, of course, you have VERY big amplifiers with decently efficient speakers. At RMAF 1996 Peter Smith and I demonstrated the issue using a peak/average digital power meter I designed. We played Rickie Lee Jones "Getto of My Mind" from her Flying Cowboys album in a typical hotel room at realistic (but not overly loud) levels using quality loudspeakers with 84 dB efficiency. The Denon 250 wpc amp we used clipped, while average power out of the amp per channel was less than 2W. You have to see it to believe it.

Clipping happens more often than you think. Its too bad virtually no audiophile amplifiers have honest clipping indicators.

Bottom line: don't dismiss clipping issues.

Of course, the plethora of crappy, compressed program material lulls us into complacency about clipping.

Cheers,
Bob





so crediting that peak factor and adjusting the 240 W peak power with 6 dB for the different speaker's sensitivities under consideration we're still talking 60 W for "a realistic but not annoying level in a typical-sized hotel room"

So, 2watts RMS and +250 watts peak?
Perhaps Bob's self designed power meter needed calibrating?

 

[jcx]

http://dr.loudness-war.info/album/list?artist=Rickie Lee Jones

17, 18 dB on some albums by their SW - that's already 60 x the average power

and the flying cowboys does get a +24 dB which supports Bob's post


maybe some here should recalibrate their expectations and include reads of Doug Self and Bob Cordell's books?

even the 14 dB that is the bottom of the DR site's "good" range is 25x power

 

[Arty]

according to that, on my usual weekend music hours i would need a past 2,5 Kw / channel continous power amplifier.
and none of the nelson pass amps are suitable for anything more than flea farts.

 

[jcx]

I did allude to rare clipping not being all that objectionable if the amp doesn't make it worse with sticking/overhang or oscillation bursts

for one thing its hard to know what a transient should have sounded like - some are literally pulled down with a mouse on the data points in digital editing if compression isn't heavy enough

temporal masking is strong too so it would be hard to judge from other things being lost during a short clip


continuous clipping is a audible problem though

but even there you can find some stories of "those kids" given access to the volume knob of a real audiophile high power and clean system and lunching the tweeters because the sound didn't "get fat" from amp clipping when turning it up

 

[Arty]

hehe, yess..
well this is a battlefield for many people.
let me explain.
one point of view, is that what we want from an amplifier is to be as close to the ideal as possible. so the signal on the input , and the output of the amplifier may only differ in amplitude, or if the gain is set to 1, then should differ in nothing.

the other point of view is, that amplifiers have a certain character.
an ideal amplifier may not sound "right" at all.

tube fans could describe the thing far better i think.
certain forms of clipping and distortsion actualy make sound "alive and vivid" .

finding a balance of what amount of lcipping and distortsion is "just right" is verry hard. i had a chanse to listen to surgically precise amplifiers and some that where far from perfect, yet i found both of them have a certain charm.
"best sounding" -> i think it absolutely depends on the person in question.
i dislike tube amps for anything expect guitars. naturally

i'm not sure this field is really understood too well.
thing is, more dynamic range and no clipping at all is supposed to deliver the best of the best when repoducing audio. Yet, all my favourite amps do have certain clipping patterns.
some of my amp builds had opamp input stages, on purpose set with some gain, and not enough supply voltage to keep them from clipping. and they had a bit grain in the sound. i can best describe them as giving "balls" to the amp. even at low volume settings (just as you described the kids popping tweeters) they sounded far more "full".
yet , i know i removed detail, and reduced dynamics.
i killed the sound, yet it made it sound "right".


now, i think a competition like this would rather revolve around producing the "right" sound rather than producing an "ideal" amplifier.

 

[radiosmuck]

Cordell says his device can indicate a 20 microsecond clip, can you hear it, does it really matter?
His eye's cannot see the speed of the clips, he needs a "peak hold" applied to the indicator bar.

 

[Arty]

uu yess, that is far off the scale.
and i think its just wrong as it is.
there is a point when improvement makes no difference.
however, clipping is -lucky for us- is just one aspect of an amplifier.
20us clipping is nonsense.. really.
i did not know the test included instrument that can indicate that kind of clipping
actually i can't decide if its funny, or sad.

 

[dwferg]

This may be a little heretical on a chip-amp thread but what the heck. Have you considered a class D amp? Super easily done within budget. Check out 3116 thread and whats being done to get top notch sound for very little $$$.

 

[AndrewT]

...............
20us clipping is nonsense.. really.............

A 20kHz transient riding on the lower frequency waveform lasts just 25us for the half wave.
Remove 20us of that 25us half wave signal and you have removed the transient.
Then there are all the other slightly lower frequency transients.
Over tha range of 10kHz to 20kHz the half wave lengths are from 50us down to 25us.
Clip them for 20us and the treble is going to sound very different. Yes chopping 20us from a 10kHz signal will be audible.

I'd suggest that we should be looking at transients lasting as little as 5us to determine the peak to average signal levels.

 

[radiosmuck]

A 20kHz transient riding on the lower frequency waveform lasts just 25us for the half wave.
Remove 20us of that 25us half wave signal and you have removed the transient.
Then there are all the other slightly lower frequency transients.
Over tha range of 10kHz to 20kHz the half wave lengths are from 50us down to 25us.
Clip them for 20us and the treble is going to sound very different. Yes chopping 20us from a 10kHz signal will be audible.

I'd suggest that we should be looking at transients lasting as little as 5us to determine the peak to average signal levels.

There are music lovers and then there are those tortured souls that listen to equipment.

 

[jcx]

and there are those that use tortuous rationalization to avoid learning something new about the real world

I think it is sad that some of you feel the need to minimize this info

how did any of what you think you know now ever enter your worldview if you are so defensive/dismissive


even odder the is the "listen to the equipment" comment about what is a fact of musical signal source being treated more naturally, preserving the dynamic peaks in real life performances

rather than relying on compression either from yesteryear's analog magnetic tape saturation or deliberate compression hardware or digital plugins

not having the peaks is a consequence of poor equipment, artificial practices in recording, choices to "punch up" the sound with compression in mastering, production

some compression is now ingrained cultural expectation of the sound of certain instruments, musical genre - some is stupid "Loudness War" exploitation of human foibles


if you are a flea power fan - your limited power is a "equipment 1st" choice - if you like "the sound" that's fine - just don't argue backwards from that personal affectation that dynamic peaks in real performances don't exist, aren't audible

 

[radiosmuck]

and there are those that use tortuous rationalization to avoid learning something new about the real world

I think it is sad that some of you feel the need to minimize this info

how did any of what you think you know now ever enter your worldview if you are so defensive/dismissive


even odder the is the "listen to the equipment" comment about what is a fact of musical signal source being treated more naturally, preserving the dynamic peaks in real life performances

rather than relying on compression either from yesteryear's analog magnetic tape saturation or deliberate compression hardware or digital plugins

not having the peaks is a consequence of poor equipment, practices in recording, mastering, production


and reproduction if you are a flea power fan - which is a "equipment 1st" choice

I have accepted the fact that I dare not reproduce the dynamic peaks of a real life performance. It's being considerate, one man's music is a neighbours noise pollution.

 

[Arty]

you can hook up a headphone to an 50 watt amp to avoid all that nasty clipping, and comapre to a no more than 1 watt headphone amp if you like.
if it makes a differnce for you then you should go that route.
in that case the bare minimum amplifier worth for you for listening cds would have to be able and get away with least 116 dBL.
and you will need speakers that can actually handle that stuff.

don't see nelson pass amp owners complaining over clipping for sure, actualy most of the respected amplifier constructors don't really aim for überpower amps.
i'm confident they know what they are doing, and it is in harmony with my own experience, so why should i doubt it ?

 

[DPH]

If your listening level is in the mid-70's, handling 24 db becomes a lot more manageable. Plus, if you design around a certain average loudness, you don't need a much, much bigger PSU/amplifier design to handle a larger DR (as your average power consumption doesn't dramatically change).