more on digital clipping..

[wes480]

when that new RHCP album "by the way" came out, i bitched about how bad it sounded and how i could hear clipping all over it.

well, that new single has been on the radio some, and I kinda like it...so, I pulled the cd out again..put it on...and again, i was just astounded by how a major label could put it out.

so, I ripped a WAV file of a couple of the songs off of BTW, and watched the meter in cool edit pro while playing it.

Sure enough, it slams RIGHT on 0db basically the entire song...yet..never crosses. According to Cool Edit, it never clips. And, why should it...after all, it was done at a "great" mastering house, yada yada.

But, when I put it on my stereo, and crank it even a LITTLE, it is just *dddaaaamnnn obvious*, and others have said the same thing, so I know it's not just me.

Bottom line is I can hear digital clipping. So, my question is why? If CoolEdit says that the file never clips - does that mean that even being at the borderline of clipping could produce the audible effect?

I can hear it at low volumes on my stereo as well...but, it's mainly noticeable at high volumes of course.

I'm gonna through it on my HR824s later...audio box is just down temporarily.

[Slackmaster2K]

Firstly, nothing digital can ever, ever, go above 0db. That is why digital clipping sounds so bad. All values that sum to a value greater than 0db are simply chopped off. While it is system dependent, I would assume all systems to deal with overflow values by simply truncating them. This absolutely destroys waveforms. If you look at the waveform for the song in question, you should see areas that look like they have flat tops (& bottoms)...it's as if somebody took a lawnmower at 0db and chopped everything off, which is essentially what happens. In the analog realm, there are no hard limits because it would be impossible to design a circuit that could handle a line signal measuring "0db" and absolutely nothing higher...there is always some wiggle room...headroom!...at any rate, while you still may damage your signal by clipping in the analog realm, the resulting waveform will still have an organic shape - think more "rounded off" instead of "chopped off".

Secondly, just about everything released on CD will bounce up to 0db regularly. The last step of any mastering stage will be some sort of hard limiting and/or normalization. There's no reason NOT to have a digital recording have a peak value of 0db.

Slackmaster 2000

[wes480]

slack - thanks for the info...i understand that it can't go over 0db..what i really meant was that cooledit doesn't read the waveform as a clip....however, my point is in reference to that particular cd, 'by the way', from rhcp.

the thing about that mastering job is that there are absolutely no dynamics left in the recording, it's ALL at 0db...

it is very fatiguing to my ears...and i can definitely hear the clips happening.

so again my question is - I can hear the clips - so can many others, on californication and by the way, it's been discussed before.

so why doesn't the software say it is clipping?

i'm not just imagining these clips!

[strmkr]

This might help. Record some audio and make it clip. normalize(never do this to audio you want to keep) to 99.5% or there abouts and run your stats. It should show no clipped samples,but will probably still sound ugly. Just because it isn't showing clips in the final project doesn't necessarily mean it never clipped somewhere during the process.

[wes480]

yeah, that is true.

i dunno, i just don't get it. its a million dollar record <shrug>.

[bleyrad]

i can't say for sure without hearing it myself, but what you're describing sounds more like the typical modern victim of extreme hard limiting. go over to http://www.prorec.com/ and read Rip's article "Over the Limit" if you want to know more.


that said, if there's actual digital clipping in the album, as you say, then i would be extremely surprised that a major label would put that out.

[Blue Bear Sound]

IT is also quite possible that the CD is NOT clipped, but the levels are hotter than your player's circuitry headroom will alow for and it's the player itself that is clipping....

Years ago I had bought a JVC cd player that did that -- promptly returned it and bought a Sony -- problem solved.

[TexRoadkill]

You could be hearing the acutal distortion off of the Limiter that was used in mastering. There is a big thread on RAP about how high slew rates can cause the DAC to clip. As usual the thread turned into a big pissing match about digital vs. analog and very little useful information was added to counter the argument.

[Slackmaster2K]

As I said, you will typically not see clipping when you playback an audio track on a digital system. Clipping cannot be recorded to a CD or any other medium. The resulting sound of digital clipping can of course be recorded, but not an actual "clipped" value.

It is possible for a digital meter to trigger a clip light when it receives a particular frequency of 0db samples...like if you have 10 samples in a row all at 0db, then the waveform was probably clipped, and a light could go off.

No two digital meters will be equivalent though, and it's up to the software writer to determine what colors flash when. In most cases, there will be no clip indicator in a software player, because it's just not necessary. In your recording software, there will definately be a clip indicator because it's designed to manipulate sound, and in this instance, real clipping can and does occur, and you need to know about it! Whether this clip indicator fires only on values greater than 0db (in relation to your converters) or if it also fires based on a certain number of 0db samples is hard to say. You could test it by simply turning the volume up on a track until it's constantly clipped, and then write that to a file....open that file in the player you're interested in and see if you get a clip indication. If you do, then it clips based on 0db sample frequency. I would assume this NOT to be the case in a multitracker, because all you're concerned with when mixing are values that go ABOVE 0db, which is only possible when manipulating and summing tracks.

At any rate, if you want to see if your ears are correct, just rip a track or two from the RHCP cd and take a close look at the waveforms....if you see ANY flat tops, then they put it through limiter hell.

Slackmaster 2000

[Toki987]

Everything I did initially had flat tops. I didnt know that what I was seeing in SF was bad thing. I was too hot tracking and having to back it off in SF, resulting in the flat tops and a really squashed and muddy sound. I appreciate you laying it out like that, I understand much better now. I also dont track to 0 anymore unless it may be a transient cymbal, occasional snare pop or something.

[wes480]

Blue Bear - interesting about the player overloading. It happened whether I was playing it off the CD player (marantz), or playing the WAV file from the computer (hooked up to marantz reciever, from an SBLive!). So I don't know.

Here is a look at the wav file. I took a snip of this and the clipping sound is really obvious to me on any playback system I used (including my HR824s).

http://www.hometheaterdesigns.com/we...screenshot.jpg

this is a one minute snip of the last part of the song featured in the jpeg above - where you see the cursor marked.

http://www.hometheaterdesigns.com/wes/BTW-section.wav

you be the judge. I think this is closer to Rip Rowans thing,

[bleyrad]

i'm only listening on headphones right now, so my opinion may be completely invalid, but it sounds to me like it's the lead vocal clipping in some way, and nothing else.