Page 1 of 1

scanning for corrections based on length

PostPosted: Wed Jan 19, 2005 3:07 am
by ramiller
Along with finding corrections based on magnitude, I would like to search based on the time width of the inserted correction also. The vast majority of corrections that can be improved by manually adjusting them are the ones that use the longest duration of inserted correction.... for some reason these corrections have a slight drop in amplitude that results in an audible artifact, almost always they can be fixed by shortening the duration of the correction. Most of the time these are the high amplitude clicks shown in red in the correction it's not a big deal on good condition records, but I have seen even amplitudues of 8 or so cause these long duration corrections and when I search through all clicks of 8 or greater it takes a while, especially on records in poorer condition.

I want to say how happy I am with this program....I'm astonished at how good a job it does. It's better than any other click software I have used (and I've used several) in automatic mode, and if a person is willing to take just a little bit of extra time and manually adjust the large clicks, scan at low correction rates when opening the file (1 or 2), manually remove corrections caused by musical transients and block scan at higher correction rates in low volume passages where soft clicks are audible, you can make a better cd from an lp than many cd releases from major companies of lp era recordings. Now that I'm familiar with the program (about 30 lps processed) I can usually do all this in about an hour per album side.

A very satisfied customer
Randall Miller

PostPosted: Wed Jan 19, 2005 10:06 am
by Derek
Yes this is a good idea. These very long corrections arise because there is a limit to the width of correction that the program can produce. When the click detector indicates a very long duration click, the program puts in a partial mute but this, as you say, is usually slightly audible.

I think the easiest way to (nearly) implement your suggestion would be to override the magnitude parameter whenever one of these very long corrections occurs; in other words, ensure that all these very long corrections are magnitude 15 or more. That way they will stand out in the Correction List and be easily searchable for. Does that sound OK?

scanning for corrections based on length

PostPosted: Fri Jan 21, 2005 8:48 pm
by ramiller
I think that would be fine, that would reduce the number of corretions to search through to find these by a large percentage.

PostPosted: Mon Feb 28, 2005 5:50 pm
by Glenn

I include myself, along with Mr. Miller, as a very satisfied customer. Wave Corrector is by far the best click detection and correction software I have used, regardless of price. As an audiophile, my approach to click removal in the past has been the 'hunt and peck' method, correcting the noise by redrawing the waveform with a pencil. Wave Corrector has made this approach obsolete, but one can only begin to appreciate how good it is when it comes to 'fine tuning' the results.

On a level 3 scan, it provides acceptable results in a matter of minutes. However, on closer examination of the sound I noticed that percussion transients are often smoothed over, removing the dynamics. Also, instruments such as saxophone can sound dull, lacking the 'edginess' of the live instument. Applying a level 2 scan leaves too many clicks in the recording, forcing one back to the 'hunt and peck' method to remove them.

What I have noticed is that in most instances where the click detector is over-shooting, there will be a large number of sequential corrections (either 0 or .001 seconds apart) of varying magnitudes in the list. My current approach is to find these sequential corrections and remove them as a block. Some means to help identify these sequential clusters would be helpful to me. Perhaps they could be indicated by a different colour, or some other identifying marker.

I'd like to second Mr Miller's request for some means of reducing the correction width. However, I envision an approach that would adjust all the corrections across the board. Something like an option to 'tighten corrections' in the autoscan options, along with the ability to vary the amount of adjustment that would be applied. In this way, one could fine tune the corrections to a level of audibility that would be appropriate for the listener. I don't claim to understand the principles involved to say whether this is workable or if it would even provide the desired results.

I feel the best way to improve the already excellent engine that is the program's core strength, is simply to provide more tools to accelerate the final editing that is inevitably necessary for producing the highest quality results that Wave Corrector is capable of.


PostPosted: Tue Mar 01, 2005 9:54 am
by Derek
Yes, I am aware for the problem of the sequential clusters. It has been brought up before. The correction signature option was an attempt to identify these but it is not very effective. I'm actively looking at ways to overcome these problems so your suggestions are very welcome. Maybe a signature in the main window rather than the overview window would be more useful as it would allow you to zoom in to the problem areas.

Discriminating between clicks and music is very difficult. When you consider the vast range of the musical repertoire, it is perhaps not surprising. Added to this is problem that some records are near pristine whereas others are in a dreadful state with ingrained dirt in their grooves.

Please be assured I am working hard to find a solution to this.

PostPosted: Thu Mar 10, 2005 1:19 am
by Glenn
Hi Derek,

Regarding your suggestion:
Maybe a signature in the main window rather than the overview window would be more useful as it would allow you to zoom in to the problem areas.

Yes, this would work for me. I would have prefered an indication in the corrections list, but the main window is probably better, especially if I could delete all the corrections in the block, rather than just the ones in the list. After all, it is better to take a look at what's going on...