Stereo Tool
https://forums.stereotool.com/

Understanding "TILT". Hans .. can you demo w. a video?
https://forums.stereotool.com/viewtopic.php?f=27&t=30318
Page 2 of 2

Author:  hvz [ Thu Jul 15, 2021 3:22 pm ]
Post subject:  Re: Understanding "TILT". Hans .. can you demo w. a video?

Quote:
Quote:
Hi, sorry for the slow response. What tilt basically does is, it tries to de-tilt the audio before detecting clipping. As you can see in many files, if a tilt has been introduced after clipping, then detecting clipped audio is no longer a matter of just finding samples at the same (and highest) level, but instead you need to tilt the audio first to make those peaks flat.

It's a bit hard to explain, the default settings were created based on some worst-case recordings. But there may be even worse out there that we haven't seen.

To get an idea what tilt does, you can use the RC slider for the FM output and check what it looks like on clipped audio or on square waves. The de-tilt filter basically tries to find the best RC value match for every piece (a few dozen ms) of audio.
Hey Hans, in the Metallica Album Death Magnetic, if you zoom in on the waveform, the signal is not only heavily clipped, but you can also see a very clear tilt downward of around 15 degrees in the flat peaks across say 100 samples. Is this what you are talking about? If so, when I get back home I can upload an image of it. BTW, What the heck plugin would have caused that in the mastering stage when they made that fatiguing album?
Yes that's exactly what it is, and in fact I tweaked the settings on that album. No idea what caused this, it looks like they first clipped it and then highpass-filtered it. But it's a mystery why anyone would do that, for one, it would have been both louder and with less distortion if they had done it in the opposite order.

Author:  MrKlorox [ Thu Jul 15, 2021 4:53 pm ]
Post subject:  Re: Understanding "TILT". Hans .. can you demo w. a video?

The band Baroness appears to use a similar tilting clipper, though not with as much loudness. However wow the distortion is so extremely fatiguing, even with your declipper. Though I haven't tried that band recently with my updated knowledge and plugin. 😬 It's definitely an artistic choice. But WHY?!

Oof. That emoji does not scale.

Author:  beneyto [ Thu Jul 15, 2021 11:09 pm ]
Post subject:  Re: Understanding "TILT". Hans .. can you demo w. a video?

Talking about clipping... Do you listen NDA by Billie Eilish?

Author:  Dro [ Sun Aug 01, 2021 12:35 am ]
Post subject:  Re: Understanding "TILT". Hans .. can you demo w. a video?

Quote:
Quote:
Quote:
Hi, sorry for the slow response. What tilt basically does is, it tries to de-tilt the audio before detecting clipping. As you can see in many files, if a tilt has been introduced after clipping, then detecting clipped audio is no longer a matter of just finding samples at the same (and highest) level, but instead you need to tilt the audio first to make those peaks flat.

It's a bit hard to explain, the default settings were created based on some worst-case recordings. But there may be even worse out there that we haven't seen.

To get an idea what tilt does, you can use the RC slider for the FM output and check what it looks like on clipped audio or on square waves. The de-tilt filter basically tries to find the best RC value match for every piece (a few dozen ms) of audio.
Hey Hans, in the Metallica Album Death Magnetic, if you zoom in on the waveform, the signal is not only heavily clipped, but you can also see a very clear tilt downward of around 15 degrees in the flat peaks across say 100 samples. Is this what you are talking about? If so, when I get back home I can upload an image of it. BTW, What the heck plugin would have caused that in the mastering stage when they made that fatiguing album?
Yes that's exactly what it is, and in fact I tweaked the settings on that album. No idea what caused this, it looks like they first clipped it and then highpass-filtered it. But it's a mystery why anyone would do that, for one, it would have been both louder and with less distortion if they had done it in the opposite order.
Clipping into high-pass seems par for the course for loudness war victims. The tilt from the high-pass often leads to a second stage of clipping, one straight (caused by the high-pass), one tilted (originally straight clipping that got tilted by the high-pass).

FYI, I still don't think the detilt feature as we have it right now works correctly though, I get much better detection of tilted clipping with it turned off and then tweaking the default tilt, as linked in my post on the first page.

Author:  EliteData [ Thu Aug 19, 2021 6:42 pm ]
Post subject:  Re: Understanding "TILT". Hans .. can you demo w. a video?

here is a short video that visually demonstrates what the input "Tilt" function does.
https://youtu.be/PuMKjOMszDE

Author:  Bojcha [ Thu Aug 19, 2021 8:44 pm ]
Post subject:  Re: Understanding "TILT". Hans .. can you demo w. a video?

Quote:
here is a short video that visually demonstrates what the input "Tilt" function does.
https://youtu.be/PuMKjOMszDE
That is NOT Tilt, it's just bad DC offset and it can be fixed with HPF.

Author:  EliteData [ Thu Aug 19, 2021 10:13 pm ]
Post subject:  Re: Understanding "TILT". Hans .. can you demo w. a video?

Quote:
Quote:
here is a short video that visually demonstrates what the input "Tilt" function does.
https://youtu.be/PuMKjOMszDE
That is NOT Tilt, it's just bad DC offset and it can be fixed with HPF.
its DC offset corrected by input tilt, a subset function of tilt.
HPF can correct a DC offset *from* the input but not *on* the input itself like input tilt does.
as demonstrated in my video, i am discussing the input tilt, not the input tilt for the de-clipper or the input tilt for the FM output but the function and behavior of tilt is the same regardless.
the demonstration video using input tilt correction is more illustratively and visually pronounced using a DC offset.
input tilt is really overlooked for what it can do aside from de-clipper tilt and FM output tilt.
here is another example using 40hz square wave instead of a DC offset on the input tilt.
https://youtu.be/xyGjw45yE90

Page 2 of 2 All times are UTC+02:00
Powered by phpBB® Forum Software © phpBB Limited
https://www.phpbb.com/