I did discuss the issue with MC:

ADI-2 defines “Over” as a single sample hitting 0 dBFS, for three samples in a row a counter had to be involved, which RME wanted to avoid.

0 dBFS, if showing up more than very occasionally, is in fact “Over” done to the music (so is constantly showing -0.2 dB or such).
No matter how soft, hard or whatever method of clipping was used during production, the music is distorted.
This must not necessarily be audible to everyone on every reproduction system, but to me it’s annoying here quite often.
I mean audibly, I am not listening to music watching the meters usually.

In the recoding studio we are metering this way nowadays, so - NOT doing it like this would simply be wrong or hiding the (sometimes bitter) truth.

In fact we even check for “Inter-Sample-Peaks” or read “True Peak”, which would be much more annoying to you to see!
I would not ask ADI-2 to do that, it’s a complicated issue.

The Pre-FX Meter shows the reality of what comes in, and can and should not refer to what happens at ADI-2’s output.

For the Post-FX Meter, = output, especially for headphones, ADI-2 already has highly sophisticated metering, the “Impedance based Level Meter” that shows the true overloads of the analog amplifier (DAC manual page 36).

What happens if you drive 100 and you speedometer shows 80?
You collect tickets.

I have a hard time to understand why you ask ADI-2 to kind of ignore 0 dBfs?
Especially as you are a colleague of mine, and and should be used to this.

I see you dislike the word “Over”.

As you mention, and of course not only am I aware of the points you made, but use limiting for masters too, 99% of all music is “distorted”, kind of.
Isn’t this “Over”?

Showing 0 dBFS as “Over” – or not – is both based on an “opinion”.
All DAWs I know have the “opinion” it is Over.

https://www.delamar.de/mastering/true-peak-level-14815/

As ADI-2 Pro doesn't seem to be able to count samples at 0dBFS or to support TPL according to ITU (Paper BS 1770) // EBU recommendation R128 it might be the "better/safer" approach that it highlights every sample at 0dBFS as OVR compared to tell nothing at all "because 0dBFS is a valid value".

One sample at 0dBFS would not be a problem, but (according to this article) the majority of music seems to have a problem by having much more consecutive samples at 0dBFS.

The article also tells "Initially there were nine samples, but today it is known that even two samples do not give sufficient overload protection against so-called shoot-throughs or interleaved sample overs.

All this indicates that signals at 0dBFS are "critical" and every sample counts.

I think
- this is what KaiS means here with "What happens if you drive 100 and you speedometer shows 80? You collect tickets." and
- we can live well with the current implemantation and documentation in the manual as we are on the "safer side" by it.

Additionaly it comes to my mind that values over 0dBFS theoretically may not exist and the ADI-2 doesn't do oversampling. so it can only warn at 0dBFS. I hope this makes also sense to you.

The manual documents it this way in ch 15.1: OVR (Over) are displayed in red color at the upper end. For digital I/Os, Over appears when the signal is at the maximum possible level. As usual in digital audio, Over is already displayed even if there is no real overload..

And .. RME must have thought of something for this implementation

I’m not insinuating RME did not put any thought on the subject. But the fact they changed opinions along the road means the subject has not a definitive answer. And my goal here was to combine both approaches by making propositions.

The issue here is that we take one single indicator — OVR — for two different things — 0dBFS reached, and unit overload.

Going the True Peak rabbit hole for thinking about this problem is not a good idea. For at least two reasons:
- inter sample peaks (ISPs) are handled by DACs, no distortion happen in the analog domain. The early days of digital are way behind us since a long time
- ISPs can reach +3 or even +6 dBFS. Which means that it’s not because your audio signal is peaking under 0dBFS (sample peak) that it will not generate inter sample peaks. Thus, an “OVR” indicator that would appear when a sample has reached 0dBFS, cannot be used as an indicator of potential ISP (let alone that there is absolutely no norm concerning the so called ‘True Peaks’, but that’s another subject).

I am here to make propositions to “reconcile” RME with itself (if I may say so) and provide a solution for the two approaches about the OVR indicator RME have had in the past.
One single indicator cannot do both.

Having an OVR signal popping up for a valid signal is just a nonsense, plus it is very annoying.

Ok, got also your point.

Hmm, as the ADI-2 Pro does no oversampling, nor is counting the number of 0dBFS samples
it might be better (and a real alternative or "way out" in this discussion)
to only highlight / display what the device is really able to show and this is that 0dBFS has been reached
but not more, to prevent making potentially false claims about OVR's when it's only based on "assumptions and likelinesses".

Maybe the 0dB mark could be made red and a little bit bigger so that red light at 0dBFS would be easier to see.
Plus documenting this change / the intention in the manual.

And if you need a more granulary view on this then better use DIGIcheck or other tools ...

And std disclaimer ... @MC to decide of course Has to be in alignment to their other products to give it a unique shape.

EDIT: question to RME, what is the reason for not counting the number of consecutive samples at 0dBFS ?
Would such a quite simple counter add a certain / unwanted delay ?

What is surprising - to say the least - is that in the DIGICheck help, RME does explicitly state that:

• methods counting consecutive 0dBFS samples (including count=1 like in the ADI-2) and showing OVR based on this count are a thing of the past (I’m slightly paraphrasing but that’s pretty much it)

• those methods are unable to differentiate between valid and truly distorted signals (sic)

• the way to go is to display peak values after Oversampling, without showing any OVR indicator

I say Amen to all of that! So I’m not sure why RME decided to come back to those methods of the past, which are not meaningful, nor scientific, nor useful at all, and misleading...?

semantics here and my paraphrase was not about counting but about the claim that it’s a thing of the past
edit: DIGICheck allows to set a count=1 by the way

I don’t read German unfortunately, but this is exactly what I was saying to Ramses in my first answer to him above

I personally am not asking this question. I was just wondering why RME recently in this firmware came back to an implementation they consider NOT the way to go.
If displaying “OVR” based on the number of consecutive samples reaching 0 (be it 1, 3, 9, or 20) is a thing of the past and tells nothing about whether the underlying signal is valid or truly distorted [=all what I said in this thread before reading the DIGICheck help] why having (re)implemented it in this firmware..? (rhetorical question, I’m not waiting for an answer)

KaiS said: ADI-2 defines “Over” as a single sample hitting 0 dBFS
A single sample hitting 0dBFS has absolutely no meaning regarding saturation, clipping, distortion, nor Inter Sample Peak value.

Yeah, I know what 0dBFS means in terms of sample value.

Now okay... maybe you are not familiar with how music is made, how music is mixed, and how it's mastered; the very one music you listen to and you like.

Here's a bad news for you: clipping is EVERYWHERE in the music you listen to, and it has been for decades. The same way analog clipping has been taken advantage of by using it as a creative tool, digital clipping is also being used as a creative tool at the production stage, at the mixing stage and at the mastering stage. Did you know that drums are clipped, basses are clipped, vocals are clipped, hole songs are clipped, etc.? It's there, but you don't notice it in the music you enjoy listening to. (Of course I'm not saying it's there is 100% of music that is made, but in very large number -- KaiS said 99%, I don't agree with that number but for this sake of demonstration let's take this number).

Now that you know that, will you suddenly throw away all the records that you love and that you find beautiful sounding and good feeling?? Just because now you know there's clipping?... sic. Of course not. Would you dare to say that 99.x% (your number) of 99% of music (KaiS's number) is a problem?

With your reasoning and the reasoning of the 'clipping police', genres like rock'n'roll would have never existed, because according to you " distorted guitars are wrong, it's clipping all over the place! ".

Please do not go fighting something you don't know about, and which ironically is everywhere in the music you listen to.

Have you ever analyzed electric guitar tracks created from a real DI passing through a digital guitar amp? Please don't because you'll get a heart attack just by seeing those ugly clipped waveforms... Did you know that synthesizers use square wave oscillators, i.e. what you can find worse in terms of clipping? Did you know that some music is made with that? Did you know that mastering engineers clip their converters, and have been doing so for decades? And I can go on with the list.

You know why you are fooling yourself with the theory saying "Pre-FX OVR flashing -> problem with the input stream in 99.x% of cases"? Because if I tastefully and transparently clip and limit a song at the mastering stage and lower the output by just 0.1dB (yes, hundreds of mastering engineers can do that), the 'OVR clipping police alarm' will never switch on on the ADI-2, not a single time. And you'll never notice any clipping or limiting and you'll enjoy the song and you'll be happy. But if I didn't lower the output, then the 'OVR clipping police alarm' would flash and *Suddenly*, the same song that you loved and which had no issue, becomes a problem.

This is a serious lack of knowledge.

NOW, the clipping which is happening inside the DAC -- because of DSP, because of Inter Sample Peaks, etc. -- is something interesting to have an OVR indicator for, for the user to know that some clipping has been introduced by the DAC unit to the sound he/she is listening to, which was not there in the first place. This makes absolute sense. (1)

"OVR" in the digital world came initially from the fear of Full Scale; in the very early days, clipping was considered bad because we just came from the analog world where waveforms were continuous. But the same way some people one day decided "screw the police" and overloaded a guitar amp and made that sound a thing (whether you like it or not, distorted electric guitars became a thing), some people started to exploit digital clipping as a creative tool and say "screw the police", and started to clip drums, basses, vocals, synths, whatever sound source, engineers started to clip their converters, etc. etc. as I said the list can go on.

A famous sound engineer and musician (Dan Worrall) recently made a video where he created a music where every single sample was at 0dBFS. Yes, every-single-sample. lol, he just won the loudness war. And there's absolutely no distortion (I quote Dan Worrall: "To those saying its distorted: there is no distortion. Its a clean, high fidelity recording of two pulse waves with complex modulation.").
=> Who are you to say this is bad? Who are you to say this is not music? Who are you to say there's an issue in the input stream? Who are you to limit creativity? If you don't like it: just pass your way.
Here's the link: https://youtu.be/s_ANEQu5Lto (beware of the heart attack, I warned you )

(1) you can read this sentence again because I think we agree on that point

@soupiraille,

I am familiar with how music is produced for many decades now (both analog and digital recording, mixing and mastering), no need to lecture me.

Clipping or whatever as an artistic choice of course is not to be debated at all.

And sure, lowering the level just a bit on material that is crushed to death (like Metallica's St. Anger, an album that I  love just because of its production style -- and for the music, of course) will remove OVR's and still be as damaged as before (and still produce IS-overs which may or may not be rendered correctly by a DAC.

All of this doesn't not affect the intent and the usefulness of the OVR indicators. I might repeat that I considered them rather useless as they don't sport an adjustable (up to infinite) peak hold.
That's another reason why DigiCheck should be used for a thorough review of the audio stream... or just check things right in the DAW/editor.

Possibly OVR is from the audio source ......

First, let us understand +7dBu +1dB what? dBr?

Second, what is that 0.999834? from where you caught this?

Third, does it happens in every piece of music?

You can try dual meter to observe it, could be your pre-amp input should be a bit lower.

1. What model of your amp?

2. Did you try another album?

3. Is the "EQ" you mean is not set to on at the RME? (use Auto-ref is better)

I streamed (Lana Del Rey - The Next Best American Record) from Apple Music, no problem at all.

No problem with your amp.

If you set RME auto-ref on, PEQ/BT will not overload, but you said post-fx, it sounds a bit strange to me .... do not know if it is Foobar problem but it seems not reasonable ...... wish other members can assist you in this matter.....

As of here on my ADI-2 Pro / latest firmware, the OUTPUT meter shows Over if I set any Volume level above 0 dB.

The Pro FS and all DAC models use a DA-chip that has 2.5 dB headroom ABOVE 0 dBFS, so no need to worry.
This headroom is there to handle InterSamplePeaks, but can be used to cleanly reproduce “normal” mixes at a Volume setting of +1 dB.

If you do so, of course the output is 1 dB louder than the original.
If the Over display is disturbing you, just go back to a Volume of 0 dB.
This doesn’t change anything on the sound.

Oh it doesn't bother me at all as long as I know I'm doing nothing wrong here. Indeed going above 0db is the result here. Thank you for giving me peace of mind

Ar .... he set the output 1dB higher than the source .... thought he is set at 7+1 = 8dBu ....

Exactly, at first I thought the 1dB you mean is 1dBr (my bad)... YES, I tried, and it did OVR.

Now you can enjoy the music, even with your setting, still safe for the amp, at the manual it said input max. is 7V ....

On the DAC (not Pro), latest firmaware, for the Line Out: no OVR is displayed on the post-FX meter if I set the Volume above 0 e.g. at +1. OVR shows up when I put the volume at +2.5
I guess this is a discrepancy between the DAC and the Pro. Why? I don't know

But the headroom for the Phone Out is actually +1dBFS.
Line Out and IEM (DAC) have indeed a +2.5dBFS headroom.