ClickRepair Best Settings

Perhaps you can set this in the preferences?

 

I looked but couldn't figure out how. Will revisit it when I get back home.

 

I just checked, and it is indeed in the Preferences. It should automatically save the file with the original file, but if not:

Click the radio button /Users/Softpedia + "-cr" and then click the Select folder button to choose your save location.

 

Try keeping the raw file in a folder and not just placed on desktop. The CR should place that new copy in same folder.

 

I'd gotten to that page, but I cannot access the desktop from it.

 

I never had that problem before. The program would always create a new file and save it to the desktop.

 

So create a folder for it on your desktop.

 

I think it's probably a glitch in the legacy Java and the new Mojave operating system. When you hit "open" on ClickRepair it actually shows the Desktop folder with the files saved on it, but they don't actually appear on my desktop unless you manually drag them there. They are also, apparently, not "searchable" until you move them. Weird. I'll never understand computers.

 

but you are having it now. So....

Might be a java thing with Mac.

 

Which is what I just postulated in my last post.

 

Plus this keeps popping up.

 

Actually, I no longer think Java has anything to do with it. It's a problem with Mojave. I can no longer drag images off the internet and have them copied, and saving files from Sound Studio also seem to go missing into the ether.

 

Here is what I would do. One pass in reverse (at 14 or 15) FIRST. Do the reverse first I said. Then another pass in non-reverse, and then listen to the file and note any ticks, clicks, or pops, and go remove them by hand (with any other sw that let's you zero in on each one. This should be easy as there won't be many at all.

I would not keep running the file through CR repeatedly.

 

I did Jefferson Airplane "Volunteers" the quadraphonic mix from a CD-4 LP, near mint looking loaned to me. Just recorded it in stereo 24/96.

Click repair in each direction, reverse / forward at 14. There were about a dozen ticks missed after all was said and done. Rather than raise my CR slider higher, or sending it through another pass, I simply listened to the entire 24/96kHz recording and made notes on where any click appeared. Then hand erased them (about 12 or so)

A lot of work, but this mix of the album is not coming out on disc or so it appears. And it's a unique mix compared to the common stereo.

 

I'm just getting started using ClickRepair, so have been going through all kinds of experiments. One of my first tests was to see what would happen when running the software against a source known to not have any clicks — a CD, which in theory doesn't have any audio anomalies typical of vinyl playback. My goal was to find the combination of setting that would yield the fewest number of changes (in theory, the lowest number of false positives. (Given, this was on a single CD source and a single style of music, so far from a thorough examination)

DeClick values of between 12 and 25 using the Simple method and Reverse turned ON all yielded about 12-13 false positives combined between the left and right channels. These, on examination, were really minor and very deep in the mix. They could be viewed as correcting flaws in the original recording and mastering process which frequently end up on many master recordings.

Using that as a starting point I've been using a value of 24 for LPs I've digitized myself where I know the original album has very little wear. For used albums where I have no idea of how well the album was cared for, I typically look at the waveform in Audacity, and — if I see lots of spikes — begin with a value of 36, which I adjust up or down depending on what the Noise waveform looks like. On that note...

I find it VERY helpful to look at and listen to the waveform of the noise identified and eliminated by ClickRepair, and make adjustments based on what has been removed. You can generate a waveform by creating two stereo tracks in Audacity — the original file, and the file generated by ClickRepair. Select the waveform of the ClickRepair file and Invert it. When played back or exported, you effectively get a diff of the two files, leaving only the Noise. So, export the project as a WAV or AIFF, then open it again in Audacity and you can play with the entire Noise file all at once.

 

Some have claimed that other sound not ticks or clicks have at times been removed as well. Touches of ambiance or something. I have not yet noticed it. I love CR thus far.

 

50 is way, way, WAY too much! 20 is way too much! Of course, the settings will depend on the condition of the vinyl in question, but I use no more than 12 or 13, if even that much. The general guideline is to go as low as possible. You can work wonders with as little as a setting of 3 on a record in good condition.

I also use reverse mode, Wavelet 3, and use the percussive setting.

 

I remove some pops manually. For hard cases, I use iZotope RX. In Sound Forge, I can redraw a wave. In Audition I can move samples.

My thing is that I will not leave any clicks. None. If I have to spend a whole day on one song, so be it.

I would never dream of using EQ to replace what a declicker took. If a program has you EQ, it's not a good program.

 

It does affect the music. All declickers affect the music. I can even hear it with settings as low as 3 through the speakers.

 

This should tell you what posters advice to heed. If you don't care if your audio is degraded, Stone Turntable is your guy. If you care about the sonics, he's not.

 

Try the experiment suggested above by John Purlia. Use a DAW to look at the original file and the file after Click Repair has been used on it. Invert one of the files and add the two together. You will see the result is exactly zero except the few milliseconds where a click or pop has been acted on. This means the two files are bit-identical except for those very brief times. The music is unaffected except where a click or pop occurs. Of course it is possible to set the threshold setting too high so that Click repair falsely identifies fast music transients as clicks and pops. This will start to affect the music. But so long as the threshold setting is used correctly, the music is unaffected. It is not like some sort of filter that affects the entire track. It only affects the samples that it identifies as clicks or pops.

As I explained elsewhere in this thread, Here are the statistics for a 20 minute side of an album :
Total Samples: 122,193,600 - Repaired: 1596 (left) 1690 (right) - (1/76562, 1/72304)
means it changed about 17 milliseconds (0.017 seconds) of the left and right channels. The remainder of the 21 minute track was not affected at all. A comparison of the bits before and after processing would show no change.

It changed 0.017 seconds out of 1260 seconds. Pretty good performance in my book. It gets rid of the clicks and pops and leaves everything else untouched. No difference. Bit perfect.

I like to use the method described and then play back the difference file. If you hear anything that sounds remotely like the rhythm or beat of the music, you have the threshold set way too high and Clickrepair is falsely identifying fast music transients as clicks or pops. You should hear nothing but the random clicks and pops that were removed from the needledrop.

Terry

 

John, something to keep in mind, the setting you choose from 3 to say 18, (to be effective) is not dependent on how bad the ticks and clicks are. The setting is for what type of music you are dealing with.

A soft wind quintet on an import classical label with severe clicks can still be handled with a low setting. A loud punk rock LP from the Ramones, or a Motörhead album with not too many ticks might need a higher setting do to the loud volume and texture of the music.

It is not about the condition of the LP where the setting should be. Imo

Now with that out of the way, I have seen CR miss correcting some clicks at a setting of 12. I did a fresh pass at 13 or 14 and presto! It came out perfect.

When you need to do a redo on a track that just did not come clean enough, up the setting by one or two numbers, and go back to the fresh (untouched) track and redo. Delete the failed attempt and redo the original track over again.

The only times I send a track to get another pass is when I am going in the opposite direction. First pass in reverse, second pass in normal forward.

I use simple, no decrackle.

 

It´s easy to see if it affects the music.

 

Yes. A/B tests.

 

No, you do as said earlier in this thread. You upload to some software, e.g. Audacity, where you can invert one of the files. Then look/hear at the difference between before/after CR. I can see no way it will affect the music in normal use.