Thimeo Perfect Declipper

[Dro]

This is B.Y.O.B. with default Declipper and Natural Dynamics, -7.7 dB gain, -13.7 dB RMS and "DR12":

Almost no peaks go above -3 dBFS, so it is as much "DR12" as some 80s master with one freak peak and everything else at -5 dBFS is "DR15". To me the result here is more DR9 than DR12.

DR9 can already sound excellent, given a good mix and EQ. I dislike the mentality to scoff at anything below DR12, when the truth is that most pop, rock and metal does not need any more than DR8 to DR10 to sound good. The Beatles remasters all come in between DR9 and DR11, most of the albums are DR10. I would not dare complain about their dynamic range.

When it comes to declipping, I would rather go for less, as there is very little value in leaving in those few very high peaks, but potentially they come with noticeable channel imbalance or other off-sounding quirks.

 

[Tool]

This is B.Y.O.B. with default Declipper and Natural Dynamics, -7.7 dB gain, -13.7 dB RMS and "DR12":

Almost no peaks go above -3 dBFS, so it is as much "DR12" as some 80s master with one freak peak and everything else at -5 dBFS is "DR15". To me the result here is more DR9 than DR12.

DR9 can already sound excellent, given a good mix and EQ. I dislike the mentality to scoff at anything below DR12, when the truth is that most pop, rock and metal does not need any more than DR8 to DR10 to sound good. The Beatles remasters all come in between DR9 and DR11, most of the albums are DR10. I would not dare complain about their dynamic range.

When it comes to declipping, I would rather go for less, as there is very little value in leaving in those few very high peaks, but potentially they come with noticeable channel imbalance or other off-sounding quirks.

I don't want to say that higher is better but it all comes to subjective feel, how you like the sound... And I can assure that there is a big difference between sound of this before and after Declipping:

 

[Dro]

My point was not that processed does not sound better than the original. The original pretty much sounds like garbage and given its simple straight clipping, you can expect a very good result with Stereotool default settings. The Stereotool result here will sound better for sure.

My point was that it does not sound great because now it says DR12. There is virtually nothing DR12 in the result. It sounds better because it gained around 4 dB of actual dynamic range. Truth is that I could take this song, make it a DR10 without sounding any different, just by running it through a steep highpass filter. It would peak 5 dB higher, measure DR9 or DR10, but has the same actual dynamic range as before.

Stereotool is not some magic tool that easily adds 8 to 10 dB dynamic range to a song. 3 to 5 dB is much more realistic. And it does not need to add more than that. A proper DR8 to DR10 track, with no freak peaks or weird EQ to get there, sounds highly enjoyable already.

Don't bother trying to get high DR numbers, because the DR database says only DR12 or higher is "green" or some vinyl rip with unknown processing has DR14 somehow. Chances are the modern vinyl masters are the same as digital masters, with some highpassing and then rumble filtering bloating up its measured dynamic range. What I look for when improving compressed music is how much makeup gain is used and where the peaks during loud passages are. If that indicates around 4 dB more dynamic range, I'll be happy.

Since you already run a VST host, I can highly recommend building a small DSP chain. I use Stereotool (declipper), Ozone Elements (limit/compress away freak peaks and channel imbalances), Stereotool (Natural Dynamics, True Bass), Ozone Elements (EQ, highpass, limit/compress again) and then UV22HR (dither to 16 bit). You can get much more consistent results like this, and it will be much easier to get something good sounding on very problematic clipping. Infestissumam by Ghost is a real challenge to declip nicely.

 

[Mike27]

Hi folks,

Late to this thread. Recently I had a little time to run a few thousand tracks through the VST version of Perfect Declipper. Some clean up real nice, others still sound “crunchy.” Of course one never knows what other distortions are deliberately part of the mix.

I got 8-9 dB added peaks from this ZZ Top track. It doesn’t sound super clean but a lot better. The raw version just sounds broken.

 

[HumbleDeer]

I use it too for some tracks when I'm in the music studio. I also have more complicated chains of other things, as well as some proprietary things.

Stereo Tool / SA Perfect Declipper is a lot more real time and more advanced than the other options, but beware that sometimes declipping introduces more artifacts than the clipped master does because clipping can be managed in a mastered track. Some tracks don't sound clipped with more than 100 samples flatlined. When declipping those, you get crackling and noise artifacts in the side channels. Those ARE very hard to get out with normal tools on the market.

 

[guitar sam]

guitar sam here. i know i'm late to the game however here is a technic i discovered that will declip an entire digital album in 10 seconds flat and the results are in my opinion as good or better than perfect declipper with 100% of the original audio untouched. using foobar2000 (free) https://www.foobar2000.org/ and this dsp IIR filter (free) https://hydrogenaud.io/index.php?topic=84544.0 simply install foobar2000 and open the dsp filter and it will install itself. double click iir filter and select allpass in the dropdown menu, leave the default frequency at 400hz set gain to zero
https://i.postimg.cc/FsGHCH6P/www.jpg

 

[guenthi_r]

Evaluating this Declipper, the results are mixed.
For example, declipping an Portishead Song, with this tool i clearly hear added distortion, that does not exist in the original file.

 

[Gyr8]

Morning all,

Re Perfect Declipper presets

I hope you you are all well and keeping safe. Would any of you be able to share ‘BrickwallHaters NEW Declipper PRESET’ please? It is a 9k .dcl file and the torrent is long dead and gone!

Thanks in advance and have a great day!

 

[sr.coca]

My point was not that processed does not sound better than the original. The original pretty much sounds like garbage and given its simple straight clipping, you can expect a very good result with Stereotool default settings. The Stereotool result here will sound better for sure.

My point was that it does not sound great because now it says DR12. There is virtually nothing DR12 in the result. It sounds better because it gained around 4 dB of actual dynamic range. Truth is that I could take this song, make it a DR10 without sounding any different, just by running it through a steep highpass filter. It would peak 5 dB higher, measure DR9 or DR10, but has the same actual dynamic range as before.

Stereotool is not some magic tool that easily adds 8 to 10 dB dynamic range to a song. 3 to 5 dB is much more realistic. And it does not need to add more than that. A proper DR8 to DR10 track, with no freak peaks or weird EQ to get there, sounds highly enjoyable already.

Don't bother trying to get high DR numbers, because the DR database says only DR12 or higher is "green" or some vinyl rip with unknown processing has DR14 somehow. Chances are the modern vinyl masters are the same as digital masters, with some highpassing and then rumble filtering bloating up its measured dynamic range. What I look for when improving compressed music is how much makeup gain is used and where the peaks during loud passages are. If that indicates around 4 dB more dynamic range, I'll be happy.

Since you already run a VST host, I can highly recommend building a small DSP chain. I use Stereotool (declipper), Ozone Elements (limit/compress away freak peaks and channel imbalances), Stereotool (Natural Dynamics, True Bass), Ozone Elements (EQ, highpass, limit/compress again) and then UV22HR (dither to 16 bit). You can get much more consistent results like this, and it will be much easier to get something good sounding on very problematic clipping. Infestissumam by Ghost is a real challenge to declip nicely.

Hi, it's already 2021, I was wondering if you're still using the same dsp chain, or if you've added new effects, so far I've only used perfect declipper with the default settings and I want to try other stuff to see if I get better results.

 

[Dro]

Hi, it's already 2021, I was wondering if you're still using the same dsp chain, or if you've added new effects, so far I've only used perfect declipper with the default settings and I want to try other stuff to see if I get better results.

I upgraded my RX license to replace the Declicker with the Decrackler, which works better at removing remaining distortion. It looks more like this now:

• (optional EQ to filter out high frequency that throw off the Declipper on some masters)
• Stereotool Declipper
• (optional EQ to revert the first EQ)
• Izotope RX Decrackler
• Izotope Ozone Maximizer (get peaks to a more uniform level and filter out channel imbalance between peaks)
• Stereotool Natural Dynamics and True Bass
• (optional Izotope Ozone Master Rebalance to bump up drums in a mix weak on drums)
• (optional EQ)
• Izotope Ozone Maximizer (same as first instance)

I am still only aiming for 3-4 dB increase in dynamic range. With punchy drums that are consistently coming very close to 0 dBFS, that is enough to sound good in my opinion.

 

[digitalfrost]

I have played a lot with this declipper and it makes beautiful to look at waveforms, however I cannot hear a difference. I used extremly overdriven tracks from my collection, but I cannot ABX both versions when adjusted for volume. Can somebody name tracks where the improvement is clearly audible?

Also. It seems the plugin will always cut off frequencies higher than 20khz, so if you run any high-res content through it, the high frequencies get lost.

e: A big strenght of this is that it doesn't care much about input level. So you can leave it always on and it will do a good job because the algo has its own heuristics to find the peak samples, even if they are down 12dB.

e2: I have to correct myself. The full Stereo Tool only goes to 20khz, but the standalone Perfect Declipper does no nothing to high res audio. I've now found some settings that leave most audio alone and only change stuff that indeed clips, and I have to say I like it.

 

[Tool]

I have played a lot with this declipper and it makes beautiful to look at waveforms, however I cannot hear a difference. I used extremly overdriven tracks from my collection, but I cannot ABX both versions when adjusted for volume. Can somebody name tracks where the improvement is clearly audible?

Also. It seems the plugin will always cut off frequencies higher than 20khz, so if you run any high-res content through it, the high frequencies get lost.

e: A big strenght of this is that it doesn't care much about input level. So you can leave it always on and it will do a good job because the algo has its own heuristics to find the peak samples, even if they are down 12dB.

e2: I have to correct myself. The full Stereo Tool only goes to 20khz, but the standalone Perfect Declipper does no nothing to high res audio. I've now found some settings that leave most audio alone and only change stuff that indeed clips, and I have to say I like it.

Hi, so I use Themeo Perfect Declipper all the time, works good for me. If you got a little time check out a little project of mine on YouTube, It's called Loudness War Under Control , here is the link:

https://www.youtube.com/channel/UCH59KSBVtVTLDzdnYUeF-cA

All the work is done with a little help of TPD.

Who knows, maybe it'll change your mind...