How to determine Clipping?

[Kevin_McC]

I've often wondered how to determine when a speaker or amp clips. Is it just added distortion, is it a loud pop or something else? How do you tell? Has anyone here experienced this? if so please describe your experience.

[David Meek]

Kevin, I've never been sure of a definition of clipping, just that when an amp approached it and the warning lights kicked on, the sound wasn't as nice.

I was reading Stereophile last night and saw a definition of clipping by John Atkinson. He said:

"Clipping" is defined, as usual, as the power level where the measured THD figure reaches 1%. . .

So, yes, clipping is distortion.




David - HTGuide flunky
Our "Theater"
Our DVDs on DVD Tracker

[Danbry39]

Here's a definition of clipping (not mine):

Clipping

Amplifier distortion occurring when a high energy wave form (a very loud sound resulting in a large output) is input into an amplifier and the amplifier is unable to fully reproduce it due to power supply limitations or amplifier design limitations resulting in the audio output waves being cut off (the rounded tops sliced off resulting in short waves with flat tops). Clipping creates audible distortion and can be damaging to speakers especially if the clipping is hard and frequent.

Clipping generally occurs when an amplifier is playing at a high level and it is asked to output a large amplitude waveform (“tall” wave with lots of power). The amplifier clips when it does not have the power capability to correctly create the waveform. Instead, as the wave is built it hits a ceiling essentially not allowing the wave to go any higher. Since the amplifier cannot recreate the remaining portion of the wave rising above the “ceiling,” the wave is cut off.

Generally, the more power an amplifier has (especially relating to the quality of the amplifier’s power supply) the more immune it is to clipping. For this reason, larger amplifiers tend to provide better quality sound at loud listening levels since they clip less often (if at all) compared to similar but less powerful amplifiers.

Clipping may be heard in loudspeakers as an abnormal, non-musical sound. It is unpleasant and it may damage speakers (with tweeters being particularly susceptible). This occurs because a speaker cannot produce the flat-topped waveform sent to it by the clipping amplifier.

In order avoid clipping, do not play a sound system at excessively loud levels and make sure the amplifiers being used are large enough to recreate the sound levels in the given space. Generally, use the largest amplifiers reasonable, as they are less likely to clip and damage speakers. It is much more dangerous to clip loudspeakers by using a small amplifier lacking in power than to use a large amplifier even if that amplifier’s power ratings are greater than those recommended for a given speaker (distortion, and clipping in particular, causes damage not clean power from a quality amplifier).




Keith

[Kevin_McC]

David/Danbry39,

Thanks for your replies.

I guess what I was looking for is a description from people who have actually experienced clipping in normal use. I always see people saying these speakers need X amount of watts or I see others buying large wattage amps for relatively sensitive speakers.

In my system I don't think I could clip my speakers without suffering permanent hearing damage due to the sensitivity of my speakers, all are 95db or higher.
I'm using 2-HK Sig 2.1's that are rated at 110 watts for my fronts, center, surrounds and rears.

I bet most of the time I'm using less than 1 watt because I normally listen around 80 - 85 DB
I realize that we need headroom in the amps for dynamic peaks in the audio track, but based on all the coments I see about people needing large amps I would assume that someone has actually experienced clipping at one time or another?

I was curious to learn what was heard or experienced when it happened? Did the speakers make a clapping sound like metal hitting metal or did everything just become distorted?

[dsmith]

I don't know if I have ever "heard" my power amps clipping, but I don't generally listen at high levels. However, I think I have heard my preamp clip, which is not where you normally expect it. My Citation 7.0 has adjustable input levels and a signal level display that flashes red if the input level is causing the preamp to clip. It was while I was auditioning the Innersound ESL amp (ironically a very powerful) that I believe that is what I was hearing. I have two identical CD players hooked up to separate inputs on the Citation, and my first listening with one gave fantastic results. But when I switched to the other CD to listen to some different music things just did not sound as good, and this is music I was very familiar with. The Citation's front display is set to fully dim but will come on when making any adjustments including volume changes. So while adjusting the volume I noticed the signal level display was flashing red, indicating clipping was occuring. I stopped the music and re-set the input level for that source and the difference was very noticeable. How did it sound while "clipping?" As I recall the music sounded congested and less resolved, more like a cassette tape than a CD; definitely not the pure, realistic musical sounds I heard with the other input. I suspect that is what low levels of amp clipping would sound like. Obviously if an amp clips badly enough to damage your speakers it will not sound like bad music - it will probably sound like a small firecracker went off in your speaker, and then it will be silent.

I have read that most amps will start to lose the perfect waveform well before they reach their maximum rated power levels, which means low level clipping can start to occur before 1% THD is reached, and I suspect it sounds something like what I was hearing. Keeping the output waveform smooth is what is called "clean" power, and lets your system sound its best. Just another reason to load up on the power to get the best sound.


Don

[Kevin P]

If you want to "hear" what clipping sounds like without damaging your speakers, take a boom box or transistor radio and crank it up until it distorts.

Minor clipping will just sound distorted, the music won't be clean sounding. More severe clipping results in the sound "breaking up", much like what you hear when you crank a boom box up too loud. At this point dangerous DC signals are going to your speakers and they'll fry. They probably won't pop but they will just stop working. You might get a nice scent of burnt voice coils if you're lucky too.

If you hear metal hitting metal, that's not clipping, that's a driver bottoming out. Clipping is caused by not enough power; bottoming is caused by pushing a driver too hard (the maximum power ratings of drivers, especially woofers, are determined by the amount of power/excursion it takes to bottom them).

KJP




Official Computer Geek and Techno-Wiz Guru of HTGuide - Visit Tower of Power
My HT Site

[JonMarsh]

Amplifier power versus speaker sensitivity all comes down to the question of how loud do you listen to music, and what is the crest factor of program material you like? (peak to average).

For example, if you take a typical pop CD, even very good ones, like Keb Mo or Alison Krause, the average signal level is pretty high, and as a result the peak to average ration is low. What this means in practical terms is that it sounds "loud" even when the peak power isn't all that high.

Now, take on the otherhand something like a good classical piece, or Jazz/big band music like Jacque Loussier Trio Plays Bach, or Harry James. If you listen to "Sweet Georgi Brown" from one of the Harry James Sheffield disks, (even the CD's), it will seem strangely quiet or low in volume at first- because the overall recording/mastering level is set to handle the highest peaks. The subjective volume, most of the time, is not so high, but there's no artificial limiting of the peak SPL, so when they come along, they hit a pretty high level- especially if you've turned up the volume so that the "average" level sounds normal compared with your pop recordings. THIS is when you really need a more powerful amplifier, often 10 dB or more, as compared with pop program material. This is ten times the power, BTW! So the "average" requirements that could be met with an good 15 to 25 watt amp now require a 150 to 250 watt amp to avoid clipping. If you don't ever listen to the "demanding" material at moderate average SPL's, you may never need the extra power.

For typical high performance home speakers in the 86-87 dB efficiency range, I recommend at least 100-150 watts for a moderate size room.

BTW, the easiest way to assuredly detect clipping is with a scope- and rather than just look for flat topping on the waveforms (hard to spot, and sometimes an artifact of limiting in the program material), connect the line level signal (say, from tape out on your preamp) to one of the scope vertcial channels, and connect the the output of your amp to control sweep on the horizontal channel, or if you scope supports it, another vertical channel, set for an X/Y mode with the first channel. You'll have to adjust the channel sensitivty much lower on the output from teh power amp; 5V/div is a good place to start, maybe 10V for a high power amp.

Under "normal" signal conditions, you'll just see a varying length diagonal line on the disply, but when the amp clips, the end of the diagonal line will tip up vertically(on the right) and down on the left, because the input will still be increasing, and the output won't be going anywhere.

There's also a factor in the audibility of clipping between different amplifiers- many amplifiers in the past (and probably some still in the present) have problems with "sticking" in high frequency clipping. That is, due to feedback, once the circuit clips, the internal circuitry is overdriven on an interstage basis, and the most of the current from the input gain stage is available to "overdrive" the final voltage gain stage. With this excess charge in the base collector region, the output voltage gain transistor takes a while to "clear" after the over drive is removed, so, for example, when clipping a sine wave, it doesn't come back to the right signal level immediately, but may "hang" for a significant portion of the sine wave. Very nasty, and very indicative of interstage design problems.

Even in the old SP300 I designed for Intersound in the late 70's, I borrowed a technology from advanced switch mode power supplies called the "Baker Clamp", which depending on the voltage saturation of the output gain stage, shunts or diverts the excess base drive to the collector side of the circuit, and allowing nearly instantaneous recovery. For this reason, this "Pro" amp was very forgiving of light clipping situations, and in many applications sounded as clean and power ful as amps rated twice the RMS power- a desirable feature in a live sound environment. But also desirable in the home.

BTW, non feedback designs like the Ayre, Theta, or BAT products inherently don't have this problem, because there's no feedback loop correction trying to "fix" the output with overdrive.... another reason they sound more relaxed and powerful than their rated power would lead you to believe, at times.

Best regards,

Jon




Earth First!
_______________________________
We'll screw up the other planets later....

[JohnT]

Thats a great explanation...but what does it SOUND like when it clips

[Burke Strickland]

Thats a great explanation...but what does it SOUND like when it clips

click-click-click-click-click-click-click-POP !

Burke

[Ricky]

I've been hitting 110dB peaks with my Aragon Soundstage/Parasound/NHT system. Certainly doesn't sound like clipping :P

[Kevin_McC]

Thanks for all the replies lots of good info.

Ok let's say we have speakers with a sensitivity of 86db and we have 120 watt amp driving them and we're sitting 10 feet away and measuring 95 db avg on a radio shack meter.

At what point would I hear clipping in reference to the meter reading and what would the clipping sound like? How much head room should the amp have to account for dynamic peaks in a dynamic music passage if one listened at a avg db of 95?

The other question this raises is at what point will your ears give out versus the gear. There has to be a threshold for you hearing?

I know this is an extreme example and one shouldn't listen at these levels, but using these will give a reference point.

[GregLee]

Thats a great explanation...but what does it SOUND like when it clips

click-click-click-click-click-click-click-POP !

Burke

Seriously? I've heard clipping in other circumstances, and it doesn't sound like that. In fact, it sounds pretty nice: bright or edgy. I was working on a music synthesizer someone else had written, and noticed a clipping routine in a late conversion of floating point numbers to signed 16 bit numbers to be sent to a DAC. Out of curiosity, I stuck in a diagnostic to count and report the number of times a floating point number had to be changed to prevent a fixed-point overflow.

It turned out there was *lots* of clipping -- thousands of clips per second. So naturally I changed the earlier processing routines to fix this. When the program was repaired, the music generated got warmer and smoother, but it didn't sound "better" in an obvious way.




Greg

[Burke Strickland]

Seriously? I've heard clipping in other circumstances, and it doesn't sound like that. In fact, it sounds pretty nice: bright or edgy. I was working on a music synthesizer someone else had written, and noticed a clipping routine in a late conversion of floating point numbers to signed 16 bit numbers to be sent to a DAC....

We are probably talking about two different kinds of "clipping". The one you describe as part of a purposeful musical creation process probably wouldn't sound the same as a seriously underdriven speaker/amp combo attempting to overachieve when it simply "runs out gas".

While the output from that unintended situation may not always sound the way I stated, it pretty well describes what I heard when I tried to drive some very inefficient "acoustic suspension" speakers with a fairly low wattage power amp at loud volume. Blew a tweeter, blew the amp, and learned an expensive lesson in what NOT to do.

Burke

[lonbove]

my sp300 finally needs a repair

Jon,
its Lon,
Finally looks like I´ve tracked you down. How are you? I turned up the volume of my search out of necessity. Your bomb proof design for the sp300 has proved true for, gosh, over 20 years of use?, and now in the salt air of Brasil, it finally needs a small repair and I can´t find the part. its the Heinemann on-off switch, jc1-a8-cc-a. Can you tell me where I can order one. My folks are coming down in a couple of weeks, March 11 and if I get on it, I can get one to them before they get here.
Thanks
Lon

[JonMarsh]

Hi Lon,

Brazil? Now, how on earth did you wind up down there?

You can't order that part off the shelf anywhere- it's a custom circuit breaker, that I specc'd out. We have an amp in Denver that had an "issue" a few years back, and ThomasW can get the breaker out and send it to your parents if you can provide an address. We have a couple more in working shape, but they're not used much...

These days I'm mostly using Aragon's or Ayres, or modded Aragons.

Hope you're doing well, you can see some of what I do here on this site, but my day job keeps me so busy I don't have nearly the time for this I'd like.

[ThomasW]

Hi Lon,

Needing something that looks like this?



Send me your parents addy and I'll mail it out ASAP. If you need/want more parts let me know.

Regard
Thomas

[lonbove]

That´s it

Hi Tom,
Is your last name by any chance Wally? And yes, that the part-- This is wonderful news!! I´´m waiting for my folks to send me their new address and then I´ll email it to you! Any other functioning boards, output transisters, smaller stuff, stuff you think I can´t get anymore, maybe even the volume pots would be great
thanks
Lon

[bigburner]

Thanks for all the replies lots of good info.

Ok let's say we have speakers with a sensitivity of 86db and we have 120 watt amp driving them and we're sitting 10 feet away and measuring 95 db avg on a radio shack meter.

At what point would I hear clipping in reference to the meter reading and what would the clipping sound like? How much head room should the amp have to account for dynamic peaks in a dynamic music passage if one listened at a avg db of 95?

Hi Kevin,

I think that this document from Musical Fidelity is quite a good guide to the required power of an amp in relation to the sensitivity of the speakers. What do others think?

http://www.musicalfidelity.com/

Nigel.

[whoaru99]

I guess what I was looking for is a description from people who have actually experienced clipping in normal use.

...

I was curious to learn what was heard or experienced when it happened? Did the speakers make a clapping sound like metal hitting metal or did everything just become distorted?

I'd speculate that many people have driven their systems to "soft" clipping and may not have even known it. I guess a slight softening of the dynamics one might say is a possible symptom.

"Hard" clipping is generally pretty noticeable inasmuch as the sound tends to become really flat/non-dynamic and harsh sounding.

[DL86]

Looking at some recent recordings on CD, I had a look of the actual waveform of some CD's on my PC and have found there to be a heap of clipping especially on louder transients. This is clearly audible on my sound system. Playing other CD's that arent recorded with clipped signals sounds a lot better and I can play it louder with less fatigue. Example red hot chilli pepper's stadium arcadium is just rediculous, what were they thinking when making this CD?

[Karma]

HI,
John Atkinson's defintion is a technical one that is generally accepted by the industry. It is used to establish the power of an amplifier. He uses the 1% THD number when measuring Stereophile's reviewed equipment. It is used because at 1% THD clipping becomes generally audible. If more power is demaded from the amp the distortion figure rises very quickly. To a degree, it is arbitrary but useful.

Also, the industry needed a standard to keep everyone honest when they quoted their power ratings. Not all companies use the 1% standard in order to inflate their amps power spec.

Sparky

[whoaru99]

Looking at some recent recordings on CD, I had a look of the actual waveform of some CD's on my PC and have found there to be a heap of clipping especially on louder transients. This is clearly audible on my sound system. Playing other CD's that arent recorded with clipped signals sounds a lot better and I can play it louder with less fatigue.

Example red hot chilli pepper's stadium arcadium is just rediculous, what were they thinking when making this CD?

Probably thinking this won't matter on a radio - likely where many people will hear it first and decide to buy the CD. Then, it's too late for the purchaser...