Help explaining why is PEACE EQ showing clipping?

[Nay16]

Hi there! I'm trying to understand the reasoning of why PEACE shows in some songs that's clipping.
I went through Ending the Windows Audio Quality Debate thread to know how should I control my volume and ended up applying in PEACE pre amp -0.14dB. Wanna know why? Give it a read! It's very interesting.
However, in some songs the red box (clipping) in PEACE blinks and at other songs stays red for long periods, that'd make sense if I was using some EQ presets without the proper pre amp gain or if I was over 0 gain.
Clipping can supposedly cause damage if is loud enough, and there would be people reporting they've blown their speakers because they were listening Rammstein - Du hast too loud. So, what's going on?

Here's what I've experimented with windows, apple music and apple dongle:

PEACE 0dB pre amp, 100% volume apple music, any % volume in windows = clipping

PEACE -0.14dB pre amp, 100% volume apple music, any % volume in windows = clipping

PEACE -1dB pre amp, 100% volume apple music, any % volume in windows = clipping

PEACE 0dB pre amp, 50% volume apple music, any 100% volume in windows = No clipping

Songs used: Rammstein - Du hast, Slipknot - Sulfur.

Thank you in advance

 

[staticV3]

Are you referring to the red box in Peace?

Afaik EAPO by itself does not have such a clipping indicator.

Not sure how Peace's clipping indicator works, but songs can have inherently clipped samples stored in them. Look up inter sample overs for more.

 

[Nay16]

Are you referring to the red box in Peace?

Afaik EAPO by itself does not have such a clipping indicator.

Not sure how Peace's clipping indicator works, but songs can have inherently clipped samples stored in them. Look up inter sample overs for more.

Yes, I'm sorry I thought it was the same program, I think I fixed it now

 

[Joachim Herbert]

The first posts of the thread you linked recommend -4dB. Why not use -1dB and call it a day?

 

[Nay16]

The first posts of the thread you linked recommend -4dB. Why not use -1dB and call it a day?

That's not a problem. I can't hear the difference anyways but I want to know why that happens when theoretically it shouldn't

 

[Jimbob54]

That's not a problem. I can't hear the difference anyways but I want to know why that happens when theoretically it shouldn't

Have you disabled all effects in eapo/peace /windows ?

As static says there may just be illegal samples in your test tracks or something in the chain is adding something somewhere . Hard to know without walking everything through

Is there some setting in apple music perhaps ? Have you turned on any volume levelling / normalisation ?

If you can't ever track it down , just use the -1db preamp and be happy .

 

[Nay16]

Have you disabled all effects in eapo/peace /windows ?

Yes, windows sounds off, muted in mixed too

Look up inter sample overs for more

I will, at -1dB it barely clips

Is there some setting in apple music perhaps ? Have you turned on any volume levelling / normalisation ?

It's all off, besides it also happens at stock configuration, even -1dbs isn't sufficient to tackle down whatever is causing it

 

[Nay16]

This is at 0, after re-checking everything -1db fixes it, previously it barely clipped

 

[DVDdoug]

I've never used PEACE...

Of course, ANY digital boost (without an offsetting "preamp" attenuation) can cause clipping. Even if you are only cutting, sometimes there is ripple or a bump in the response that can push the peaks over 0dB. And it depends on the levels in the program material.

Almost all DSP is done in floating-point which can go over 0dB without actually clipping. But you'll clip your DAC if you play it at "full digital volume".

Note that a lot of MP3 go over 0dB so they can "show clipping", and again they can clip your DAC. MP3 can go over 0dB and the lossy compression process changes the wave shape making some peaks higher and some lower. So the highest new peaks can go over 0dB even though the uncompressed original didn't.

 

[Nay16]

I've never used PEACE...

Of course, ANY digital boost (without an offsetting "preamp" attenuation) can cause clipping. Even if you are only cutting, sometimes there is ripple or a bump in the response that can push the peaks over 0dB. And it depends on the levels in the program material.

Almost all DSP is done in floating-point which can go over 0dB without actually clipping. But you'll clip your DAC if you play it at "full digital volume".

Note that a lot of MP3 go over 0dB so they can "show clipping", and again they can clip your DAC. MP3 can go over 0dB and the lossy compression process changes the wave shape making some peaks higher and some lower. So the highest new peaks can go over 0dB even though the uncompressed original didn't.

I've used Peace to EQ to try different flavors of tuning, other than that is flat/stock.
I think what's happening is inter sample overs as mentioned by StaticV3, this is never ending!

Note that a lot of MP3 go over 0dB so they can "show clipping"

Even FLACs? isn't this something preventable by whoever is making the music? If you know something is going to be over 0dB why don't you make adjustments beforehand so it doesn't happen? I'm talking out of ignorance here

 

[staticV3]

Even FLACs?

Yes.

If you know something is going to be over 0dB why don't you make adjustments beforehand so it doesn't happen?

Either negligence or an intentional trade-off to achieve greater loudness.

 

[Nay16]

Either negligence or an intentional trade-off to achieve greater loudness

That's disappointing. Doesn't that cause just noise and distortion? This might explain why some songs didn't jump in quality when I entered in this hobby. Not that I can tell lossy from lossless but when I went from using spotify with regular cheap bluetooth buds to lossless, dac and IEMs some songs changed completely and some got even worse, I literally can't listen to them anymore. Songs listed are an example.

to achieve greater loudness

May I ask what's this? I've seen it being mentioned but I don't understand the logic behind it if at the end it'll sound bad.

 

[AnalogSteph]

You may be in luck regarding Du hast, according to the DR Database there's a 2023 Anniversary Edition of the Sehnsucht album with improved dynamics and sound. Brickwall limiters back in 1997 couldn't have been overly fantastic, those got a lot better with the adoption of the oversampling variety around 2012.

As for what kind of overs one may be seeing, you may want to check out this thread:

True Peak may be a matter of definition...

I have been playing with the options for "True peak scan" in the Foobar2000 ReplayGain scanner lately, which allows you to use any resampler DSP installed for the oversampling. This made me notice a marked difference in results between "auto 4x oversample" - which I presume uses the filtering...

www.audiosciencereview.com

In any case your playback chain has too little headroom in general and seems to be lacking in volume normalization (e.g. ReplayGain). Output levels across a diverse collection would be all over the place to an annoying degree. What is your player of choice?

 

[Nay16]

You may be in luck regarding Du hast, according to the DR Database there's a 2023 Anniversary Edition of the Sehnsucht album with improved dynamics and sound. Brickwall limiters back in 1997 couldn't have been overly fantastic, those got a lot better with the adoption of the oversampling variety around 2012.

Woah! I'll definitely check it out.

That's impressive, why then isn't a standard configuration having all devices at -3db to avoid all of this?

In any case your playback chain has too little headroom in general and seems to be lacking in volume normalization (e.g. ReplayGain). Output levels across a diverse collection would be all over the place to an annoying degree. What is your player of choice?

I'm using windows 11 latest version, apple music and a combination of apple dongle when I'm on my laptop and topping dx1 when I'm at my desktop. AM is capped at lossless 24 bit 48khz, same at windows. Followed the instructions by the linked thread since I was using 1 streaming at a time and all else muted. -6.3db pre amp made max my volume on the AD DAC so that's why I backed to -0.14 but now I'm at -3dbs all across.

 

[Auditory Cortex]

Oh you also noticed that? yeah, when I was using PEACE EQ in the past I also noticed that weird behaviour, I also made sure there wasn't anything else messing with it but I concluded that is just PEACE because I wasn't hearing any clipping. I am just using EAPO and Voicemeteer now, the meters doesn't show any clipping when I adjusted the levels to hit 0dB. I guess you could post a ticket or question to the developer of PEACE himself, he is active in the Discussion section of its Sourceforge page where the software is being downloaded.

 

[Nay16]

Oh you also noticed that? yeah, when I was using PEACE EQ in the past I also noticed that weird behaviour, I also made sure there wasn't anything else messing with it but I concluded that is just PEACE because I wasn't hearing any clipping. I am just using EAPO and Voicemeteer now, the meters doesn't show any clipping when I adjusted the levels to hit 0dB. I guess you could post a ticket or question to the developer of PEACE himself, he is active in the Discussion section of its Sourceforge page where the software is being downloaded.

This is not a bug, AnalogSteph linked a thread showing many examples of songs going above 0db.

Look up inter sample overs

As I was looking for information about it, I found a recent video from "Audioholics" Intersample Clipping - A Problem with Most Digital Playback Systems? and at 53:27 his guest talks about audio myths and says that clipping occurs in the digital reconstruction filter due to insufficient headroom not because is caused by incorrectly produced recordings. His next point says most CDs reach +1dbFS, Few CDs reach +2dbFS and less than 1% exceed +3dbFS.
But if it was correctly produced we wouldn't be having this problem in the first place?

 

[staticV3]

But if it was correctly produced we wouldn't be having this problem in the first place?

Correct.

 

[levimax]

With so many variables with computer based playback I always run a 0 dB sweep file generated in REW and watch the levels and listen. Digital clipping is very easy to hear on a sweep. This is most likely very conservative but at least I know I am "safe". The only problem with this is I usually find I have to use ~-8 dB of attenuation to prevent all digital clipping (even though the computed filter anaysis say -3dB is enough). This incenerates a lot of headroom.

 

[kchap]

... at 53:27 his guest talks about audio myths and says that clipping occurs in the digital reconstruction filter due to insufficient headroom not because is caused by incorrectly produced recordings. His next point says most CDs reach +1dbFS, Few CDs reach +2dbFS and less than 1% exceed +3dbFS.

That is an amazing statement. Has the recording industry informed the manufacturing industry, particularly the DAC manufacturers of their unilateral decision to change digital audio standards. Oh by the way, that DAC stage which outputs 2.828Vpk-pk (1.0Vrms) now has to output 7.1Vpk-pk. My experiences are similar to @leivmax, so I have assumed 8dB.

As well as changing standards on the fly, the "guest" seems to be unaware of broadcast standards. For example, an EBU Quasi-Peak Programme Meter for digital broadcasting expects the nominal 0dB (100% modulation level) to be -9dBFS. Peaks will occasionally hit 0dBFS, but it should never exceed it.

 

[PeterNL]

Very interesting.