Clipping?

[Stefano_Milano]

A question for the experts.
I bought the new London Symphony Orchestra recording with Vaughan Williams Symphonies 4 and 6 conducted by Pappano.
The sound seemed to me to have something annoying.
So I opened one of the most aggressive tracks with Audacity. Below you see the image.
How do you rate the graph? Does it have any clips?
Can I do something with Audacity to fix it?
Thanks so much
Stefano

 

[RayDunzl]

How do you rate the graph?

Has, like everything else, some dynamic limitations.

Does it have any clips?

Audacity can check to see if it has any 0dBFS samples. Menu Analyze -> Find Clipping

If the engineer has limited the values to less than full scale (common) it will not detect "clipping" even if the waveform is modified at the extremes by compression/limiting.

Can I do something with Audacity to fix it?

Probably not.

---

If it sounds good and you like the material, play it from time to time.

If not, well, whatever.

---

Zoom way in on some of those peaks and see what's going on there. Waveform look smushed against the limits?

Example:

Not a single clip as defined by a full-range sample or two, but obvious "engineering" throughout to push the limits of loudness, and dynamically constrained.

It won a grammy.

"2014 - Best Engineered Album, Non-Classical"

 

[Katji]

You might like to try Similarity analysis for that.

 

[Stefano_Milano]

Has, like everything else, some dynamic limitations.



Audacity can check to see if it has any 0dBFS samples. Menu Analyze -> Find Clipping

If the engineer has limited the values to less than full scale (common) it will not detect "clipping" even if the waveform is modified at the extremes by compression/limiting.



Probably not.

---

If it sounds good and you like the material, play it from time to time.

If not, well, whatever.

---

Zoom way in on some of those peaks and see what's going on there. Waveform look smushed against the limits?

Example:

Not a single clip as defined by a full-range sample or two, but obvious "engineering" throughout to push the limits of loudness, and dynamically constrained.

View attachment 128297

It won a grammy.

"2014 - Best Engineered Album, Non-Classical"

Thanks for the reply.
Too bad because the other LSO Live recordings I have are great, with wide dynamics.
Among other things, it is a 192khz and 24bit FLAC ... what's the point of limiting the dynamics ...
What a shame

 

[Stefano_Milano]

You might like to try Similarity analysis for that.


View attachment 128298

View attachment 128299

I will try.
Thanks!

 

[RayDunzl]

Among other things, it is a 192khz and 24bit FLAC ... what's the point of limiting the dynamics ...

When 24bit files appeared some years ago, I thought "Wow... They've extended the loud end."

Wrong.

The extra 8 bits replaced/subdivided the smallest bit of 16, as implemented.

If your DAC outputs 2V on 16bit, it outputs 2V on 24bit. Same "dynamic range" for practical purposes. The difference is buried in whatever noise the recording has already.

Does a 24bit file not play with (essentially) exactly the same volume level as its 16bit counterpart?

My thought was "Oh well."

 

[RayDunzl]

If your DAC outputs 2V on 16bit, it outputs 2V on 24bit. Same "dynamic range" for practical purposes. The difference is buried in whatever noise the recording has already.

Of course, as sometimes happens, my initial intuition was wrong.

If implemented as imagined:

16bit DAC would output 2V (as a basic standard) as the maximum signal in the recording.

So a 24bit DAC would output 502.4V - 48dB stronger

Oh.

---

Practical Opinion:

More dynamic range is possible in playing back any recording.

The volume knob stretches and contracts the perceived decibel value of the bits in the recording as it gets to the air on its way to your ear.

 

[DonH56]

Of course, as sometimes happens, my initial intuition was wrong.

If implemented as imagined:

16bit DAC would output 2V (as a basic standard)

So a 24bit DAC would output 502.4V - 48dB stronger

Oh.

I'd go with your intuition on this one. Most DACs, or components in general, aim for a typical output level that is useful for most preamplifiers and amplifiers. Lately it seems to be about 2 Vrms, and chances are a 16-bit, 20-bit, 24-bit, or whatever-bit DAC is going to target 2 Vrms full-scale output. That means you are extending the low level, not boosting the highest amplitude. I am not up on the latest products but strongly suspect a consumer audio DAC that outputs 500+ V or a preamp or amplifier that would accept such a high voltage does not exist.

A 24-bit DAC should have about 146 dB SNR. I have not seen one of those, either... Fortunately for my ears of clay 100+ dB is (more than) sufficient.

 

[mansr]

I am not up on the latest products but strongly suspect a consumer audio DAC that outputs 500+ V or a preamp or amplifier that would accept such a high voltage does not exist.

I have worked on devices that put out 1 kV. Not audio related, though.

 

[DonH56]

I have worked on devices that put out 1 kV. Not audio related, though.

Yes, that is why I thought to limit to "consumer audio" before someone pointed out the HV drivers (I piddled with one years ago for a 25 kV system, don't miss the HV stuff!) But chances are somebody will prove me wrong anyway...

 

[mansr]

Yes, that is why I thought to limit to "consumer audio" before someone pointed out the HV drivers (I piddled with one years ago for a 25 kV system, don't miss the HV stuff!) But chances are somebody will prove me wrong anyway...

In my case, it was photo flashes. Those things were probably lethal. Thankfully, nobody had put it to the test.