difference between red and orange clipping indicators

Does anybody know the difference between the red and orange clipping indicators on individual channels?

In the past I thought it was meant to indicate that red on an output channel is really the max, no headroom. And that orange was the 'clipping color' on individual tracks at mixdown as 0.0 dB is not actually the maximum as Logic has a lot of headroom above it, so not really 'RED'.

But now I see both red and orange clipping indicators on the individual channels, both on audio and software instrument channels. It seems completely random if the clipping indicator is red or orange. Is this a bug or a feature?

Mark

I only see orange indicators on Software Instrument channel strips. They can theoretically go to +1500 dB without clipping. They are orange because they could arrive at the output that hot, and then it will clip your output.

I have seen orange clip indicators on both software instrument and audio channels. It also seems not consistent. One time I open a project and I see only orange clippings on different audio channels, next time I open the same project, some audio channels clip red and other clip orange.

Sounds like a bug to me.

The premise is as Erik states, different indications.

Aux and Outputs are always going to be red.

Audio and software instrument channels are Orange with one exception. A selected audio channel or an audio channel set to record will show the red indicator.

The premise is as Erik states, different indications.

Aux and Outputs are always going to be red.

Audio and software instrument channels are Orange with one exception. A selected audio channel or an audio channel set to record will show the red indicator.

Thank you both, but what is the *Logic* behind the clip indicator turning Red when you select an audio channel at playback?

I really don't know for sure.

Probably just a side effect of that a selected audio track is technically in Record ready mode.

I also think it's something like that.

Well, good to know the ins & outs of orange and red

My two cents:

Orange means "you clip, man, but its in the 32-bit float domain, so no big deal" (*)

Red means "you clip,man, and its either at the AD (record) or DA (output) level, so that's rue clipping that's gonna ruin your work if you don't take care of it".

(*) All internal processes within Logic itself are computed in 32-bit float data format, which means (float) that contrary to 24bit or 16 bit (or any other type of integer format), there's no practical upper limit to any signal value and "full scale" (like in dBFS) is meaningless. in other words, you cannot loose information by increasing the signal level => you won't actually clip it so long as you saty in the 32-bit float domain. As a matter of fact, digital clipping is only "loosing" information because you try to convert from digital to analog or from analog to digital a signal produces an amplitude value wich is above the maximum value that the integer format (16 bit or 24 bit can encode). With 32-bit Float, as said above, there's virtaully no maximum value (or it's so high that it's meaningless) to what you can encode.

Anyways, put it more simply :

Orange - the level of that channel strip is too hot but that's not an issue at the level of that channel strip itself (however, you should ckeck your outputs in case gain staging is not appropriate when unltimately delivering that channel strip to either of them).

Red - the level is too hot and you are clipping (actually destroying signal information) either at your input (record, trim down the recorded input) or at your output (lower your output fader).

As regards auxes, the "red" behaviour is probably a "just in case" behaviour of loging, as it can't necessarily tell where the signal is coming from (retrun from an outborad gear?) or to further down in the signal path...

As far as Auxes, I'm not fully sure I'm right, but for channels in recod more and outputs, I'm quite confident (and I'm 100% sure about integer vs float discussion above).

regards,

Arnaud

Yes Arnaud, that's how I understood it as well, but as audio tracks were sometimes clipping in red at playback, it didn't make sense anymore. Now it appears that selected audio tracks are also clipping red at playback, which doesn't really make sense as it's not an indicator of clipping at A/D or D/A stage, but clipping in the 32 bit floating domain.

Not a biggie though...

Red means "you clip,man, and its either at the AD (record) or DA (output) level, so that's rue clipping that's gonna ruin your work if you don't take care of it".

That's what we were discussing.

Select an audio track that has already been recorded. It will show clipping in red. Stop deselect it and it will show clipping in orange.

It's the same signal flowing through the FP engine so it's probably a visual bug and like I said a side effect of that track being in a "record ready" state.

Orange means "you clip, man, but its in the 32-bit float domain, so no big deal"

I wouldn't call that clipping. Over is probably a better term.

it appears that selected audio tracks are also clipping red at playback, which doesn't really make sense

Exactly.

Let me see if I understand this now:

• Orange: This signal is "hot" enough that, if it went to an output, it would be "too large to fit" and therefore would be clipped. However, it isn't now being sent to a final destination where clipping would actually occur. (All of its immediate destinations are other internal processing stages, which are floating-point with no applicable numeric capacity limits. However, "you have been warned!" If you're seeing clipping on the output, this track alone would be big enough to do it, if not compensated-for "downstream.")
• Red: This "hot" signal is being sent to a final destination, where it is "too large to fit" and therefore will be clipped.

Yes?

Eric & Mark (multispace), sorry for having misunderstood the original question. Now I better understand. Probably a visual bug, as Eric puts it.

I'm not currently writing form my Logic computer so I take note to test something : you can select a track and while it stays selected, disengage record-enable. Now will it "clip" red or orange?

Anyway, for MikeRobinson, the most important thing to remember is : Never clip when you record (or on your inputs, more generally) and never clip on your output buss(es). That's a the golden rule. Digital clipping when crossing one way or the other between digital and analog is a lethal beast to be avoided by all means. For the rest, you should not really have to care (nor care whether it's displayed orange or red).

Yeah, I can understand how audio capture is different from audio "results." (For lack of a better word ...) And, I can see how a "red indicator" would be used in the case of capture, since the bright-line interpretation is the same: "red == the-data-is-gone." ("The audio came in, but it's so loud that there was no place to put it in the audio file, and unfortunately there's no other place to put it.")

I would guess, then, that "orange == the-data-might-be-gone unless you do something about it." ("It's too big to store, but it's not stored yet.")

(Oops ... my use of "==" probably reveals me to be a software-geek.)

I'm not currently writing form my Logic computer so I take note to test something : you can select a track and while it stays selected, disengage record-enable. Now will it "clip" red or orange?

The selected track doesn't have to be record enabled for it to display the red indicator.

Logic acts in a "record ready" state when a track is selected even if it isn't record enabled. There is a visual indication of this, the R on the button will be red but the button stays grey.

Yeah, I can understand how audio capture is different from audio "results." (For lack of a better word ...) And, I can see how a "red indicator" would be used in the case of capture, since the bright-line interpretation is the same: "red == the-data-is-gone." ("The audio came in, but it's so loud that there was no place to put it in the audio file, and unfortunately there's no other place to put it.")

I wish it was that simple.

In reality the indicator shows that there is 1 or several samples (I'm not sure if the meter is programmed to show clipping for anything > 0 samples. This is often a parameter that you can set in other digital equipment.) registered at 0.0 dBFS.

This doesn't theoretically mean that you have clipping but it's a good indicator that you probably have.

I'm not currently writing form my Logic computer so I take note to test something : you can select a track and while it stays selected, disengage record-enable. Now will it "clip" red or orange?

 

The selected track doesn't have to be record enabled for it to display the red indicator.

Logic acts in a "record ready" state when a track is selected even if it isn't record enabled. There is a visual indication of this, the R on the button will be red but the button stays grey.

Eric, thanks for this additional piece of info. Clearly, as you guys, I don't see the point in this behavior (red indicator simply because the track is selected). Much probably a visual bug of some sort.

Regarding Mike's concept of "data is gone", it's not that bad an approximation. The clip indicator will clip if more than one consecutive samples are encoded (in an integer format, i.e. 16-bit or 24-bit), with the maximum available value for that format (for instance, at 16-bit, a value of 65535). Horizontal lines at maximum level indicate that the signal was truncated which even if done on purpose (square wave at max level) introduces harmonics (most often unwanted) which the system cannot reproduce properly, or, much more often, shows that part of the signal information was lost, as shown on the image below. Truncated => clipped.

For those who may read this post later, a simple graphic illustration of this:

You're right Eric in saying that you can reach 0dBFS without having actually clipped anything, if your signal just touched the ceiling. But if for more than one sample in a row (I think I remember somewhere that Logic's clip indicator lights up if it's more than one in a row, but its only a recollection here), probability is very low that the actual shape (and harmonics) of the original signal is what the system is able to reproduce. So in really life let's conclude that I'm in trouble if my signal clips red when entering or exiting the digital domain.

Another thing, as you guys know: CD factories will test this and will reject as flawed a master that clips, so even if nowadays CDs may quickly become a dinosaur of the past, it's another very good reason why no one should take the chance of letting clipping occur at digital output of one's system.