ClickRepair Best Settings

I adjust the setting on the fly. There is no one setting that is correct for all the songs on the album. I run with the 'Automatic' setting off or set to some very low number of samples. This means the software stops and lets me decide if the algorithm has picked up a real click or not. I choose whether to accept the suggested fix or leave it untouched. I adjust the threshold setting on the fly as I go through the album so as to be as high as possible without falsely identifying music transients as clicks. I will set as low as 5 for some music and as high as 60 (yes, 60) but only as a track is fading out to silence. This method takes a lot longer than just using it in 'Automatic'. About a half hour per album side, but it gets almost all of the actual clicks and pops and none of the music transients. I then follow up with a manual search (listen in real time with headphones) using an editor and manually editing out anything Clickrepair missed. People that aren't quite as anal as I am about it will get perfectly good results using the automatic mode. I wish I was more like that.....

Terry

 

No. I do as I see fit. I do what will help me recognize any differences.

 

Of course You do, I didn´t really mean You, I mean´t one can invert a file to look at differences; if there are only clicks as differences, there can be nothing else.

 

A click is part of a wave/music. Click removal programs always manage to take out some musical data. Now, whether an individual can hear it or not is another story. If you can't hear that, you're good to go. But, if you can, well, that's a problem.

 

Look, this is getting a bit silly. There is no musical information in a click, it´s a click normally. So leaving it is just a distraction. Removing the click will recreate a reasonable wave-form, but it´s not perfect.
What I´m saying is: If we look at the difference between a file run through a de-clicker, and the same file not run though the de-clicker; we will see/hear what´s been removed. If You don´t want to do this it´s fine by me.

 

You misinterpreted what I stated. If you transcribe a vinyl record, or any musical signal with an anomaly init, of course it's not music, but it's part of the music once it's there. Now, if you make an attempt to remove that anomaly, woy will also take out some of the musical content with it. Now, if you can surgically do this, you probably won't hear the effect. But, if you use an automated de-clicker, it will remove some type of musical information with that click, or whatever. Then it comes down to if you can hear the degradation or not, and it depends on the individual.

Let's consider noise reduction of, say, surface noise. Almost no one in this room will argue that the use of it removes musical information along with the noise, but for some reason, they think de-clickers, another type of NR, won't. The cold truth is that they both do. Again, it's a matter of how good the software is, and the skill and patience of the engineer. And, of course, it depends on the listener. There is no way you can eliminate the factor of the end listener.

 

Comparing clicks to noise is like comparing apples and oranges. By their “signature” sound, clicks can be detected and dealt with. Clicks are removed, and the sound that was “masked” by the clicks is interpolated from the music just in front and just after the click. This works fine because clicks have a short duration.

Noise is a semi constant sound and it was always hard to properly detect and remove it. Classic noise reduction doesn’t actually remove the noise itself, it only filters frequencies that are “used” by the noise, thus affecting also all other sounds that use those frequencies.

Maybe we’re getting better noice cancelling algorithms nowadays using big data analytics and statistics. But constant sounds are basically a pain to get rid of.

 

I use RX rather than click repair. One of the options in RX is to hear only what is being removed. Using higher settings, one can hear more than the click.

 

The nature of a click is that it is only a click, the level and duration varies; but during the time the click lasts the needle cannot read any groove information. I hope it´s obvious the needle cannot read both the cause of the click and groove info at the same time. What the de-clicker does is fitting this 'click-space' with reasonably correct information based on info before/after the click.

As for the surface noise, it´s something else completely, the freq range goes from a very low freq up to maybe 3000-4000Hz. It´s a layered info on the other info on the record, where higher freq is layered on lower freq. Remember there is only one specific wave where every info is layered.

I´m not really getting what You mean by the click is part of the music, the click is nothing more than a click, but as I said the level, the rise time and duration can vary a lot.

 

Noise is noise. Constant hiss is noise. Surface rumble is noise. Clicks and pops, and crackle: all noise. Technology is definately getting better at removing all of it, but the fact is that no matter what you do, you will take out or alter some kind of musical content. Again, whether the end-listener can hear it or not is another issue.

What we were talking about is the level of NR being used in Click repair. Frome reading the posts in this thread, it is clear that most of you are using way too much, and it makes me wonder what kind of results you are getting at numbers of, say, 5o, compared to the original, unprocessed file.

Click Repair is amazing, especially for the price, and I use it on virtually every transfer with clicks. But, as I said, it is rare that I need to use anything fast a setting of around 14. Anything else and I can manually remove bad clicks with one of several methods. This is why I do not use just one program for everything. And, using settings as high as the ones you guys are suggesting can create worse pops or clicks on some transients that will absolutely have to be removed by some other means.

The problem, as I see it, are those who look for shortcuts, want to be lazy, and just really don't want to take the time de-clicking requires. Then, it comes down to your goal. Do you just want a quick and dirty copy to lay on the go, or do you want a pristive transfer to archive? My goal is the latter. Yours may be the former.

A cool feature of Click Repair is that you can listen to the result in real time. You can do this with RX, too.

But, like I said, a lot of people want a quick and easy way out, and may have different goals.

 

The other methods are available but it's easiest for me to hear what I'm removing if only what is removed is playing.

It does appear that RX and ClickRepair have similar features. The intensity range for RX is from 0 to 10 in increments of 0.1. I typically use no higher than 1.0 (based on listening) and then repair anything that's left using interpolate. With that tool I can focus on as little as .001 sec. It does require more time but the results are superior.

Both de-click and interpolate remove the click leaving an empty space which is filled based on the data on either side of the space once occupied by the click. Hence the word interpolate. They do not remove the click and reveal what it was covering. However, we are talking about erasing and filling a space that is 1 to several milliseconds in length. Where we risk removing music is by using a setting too high. The de-click algorithm can mistake music for clicks.

Most times, if my lps are cleaned properly and are damage free, the low setting is all I need.

 

When I don’t use pitch protection I notice I lose some bottom end, anyone else hear this?

 

A cool feature of Click Repair is that you can listen to the result in real time. You can do this with RX, too.

But, like I said, a lot of people want a quick and easy way out.

My absolute favorite way to de-click clean albums with minimal clicks is manually. I open each file of an album in either or both RX and Audition in the spectral mode so I can see the clicks, and zap them in any number of ways, including literally editing it out. That requires sitting there going through each second of music of each track with high concentration, and repeatedly listening to the result to make sure what method you've chosen is transparent. And, if it's not, you try something until it works. I may even redraw or move samples. That can take hours or days! It helps to like the music!

I've even taken a tiny note from another part of a song and grafted it in the trouble spot. I'm so OCD!

 

It may be OCD, but turning a poor sounding recording into a good one without sacrificing the content is worth it.

 

My goal is to simply remove the clicks and pops that shouldn't be there. They bother me so I want them gone, but I don't want to damage the intricate sound.

 

No, it's not really another issue. It's not an issue ever, really.
Before de-clicking = click in the music.
Afer de-clicking = no more click in the music.

Done.