Dealing with "essing" distortion.

[firekat]

I am very much a novice running this program but over the years that I have had it (I don't get a lot of chances to play with it all the time) one of the things that escapes me is dealing with "essing". Since I purchased a record cleaning machine I have found much less occurences of this problem. I have found that I need much less use of the CNF and the thresholds are very different than what I have utilized in the past. From my understanding "essing" can be introduced by certain filtering (Impulse?, CNF?)or by poor recordings. What I do find is that after processing only the very highest volume and pitch "esses" are effected creating a high pitched "schhsss" sound that I cannot seem to remove. I have used the Dynamics processor to some effect though I have yet to set it up to get all the essing without knocking out too much dynamic range.

When I record the original I usually have it set so that the VU's will be tickling the red in the highest peaks, though I have yet to get the clipping indicators to light up. Is this too high for the high end "essing"? should I back it down a little? What I try to attain when I record is that after processing the file and utilizing the gain normalize function I get very little change. I hope this is the point.

I saw something posted on some other messages when researching this phenomena that related to setting gain to a lower percentage say like 95%. Would this help in eliminating the "essing"? Is it possible to set the percentage gain increase in DC5?

Do I have to set either the CNF or the impulse filters up differently?

Thanx all!

[Malcolm]

Re: Dealing with "essing" distortion.

I also am having great difficulty de-essing esses and tees: they are almost ear splitting. I recorded from a 1958 RCA stereo disc with excellent frequency response, but with poor female vocal microphone technique. Recording to the hard drive accentuated the problem a little. I have tried (on a zoom-in basis) ALL the de-essing presets I can find, tried various low-pass filter settings and tried various equalizer settings. The result is either the selected area sounds totally "dead" or the essing deteriorates into a rasping noise like blowing through a burst balloon! In either case the result is no good.
I have tried the "cures" before and after INF (INF accentuated the essing further in all modes regardless of recorded volume). I tried HF roll-off, Paragraphic equalizer with various notch settings, tried spot amplitude reduction.
The waveforms mostly have sharply pointed peaks and intermixed with some that have tiny flat tops or tops with angled flats (neither high enough amplitude to be considered clipping).
Has anyone else run into this problem, so severe?
malcolm

[Craig Maier]

Re: Dealing with "essing" distortion.

I have run into this problem, but perhaps not to the degree that you are expressing here. I have had good success using the de-esser contained within the dynamics processor. I do not use the presets, but adjust the frequency, threshold, etc. controls manually until I get the desired result. Another alternative is using the top band of the Punch and Crunch in compression mode. Put all of the other bands of the Punch and Crunch into a position where they are always active and move the thresold of the top band til the essing is attenuated. You may want to experiment with the attack and release times as well.

[ 12-28-2002, 09:07 PM: Message edited by: Craig Maier ]

[Malcolm]

Re: Dealing with "essing" distortion.

Craig, thanks for your prompt suggestions. I will experiment further after reading up on the DNF -- I find all its variables a little daunting, being relatively new to restoration work. The Punch & Crunch didn't even cross my mind as a possible method. Thanks again, and a Happy New Year to you.
m

[BigAl]

Re: Dealing with "essing" distortion.

You may want to check this thread to see how I used the CNF. With a bit of practice, you can pick up the sibilance fairly easily. Don't be shy to up the markers in the CNF, especially the higher frequency ones. For S-sibilance you're only really hammering about 90ms in a normal tempo tune.

Things to note when using selective filtering with the CNF:
* use a selection > 100ms (I always extend the selection before the sibilant passage. You then replace the "good bits" which have now been squashed with the file conversions filter.) To set the time selection, I have the cursor postion meter (View/Time display) parked off in the top right hand corner of the screen.
* the new smoothing function helps tremendously
* the CNF does not work on individual channels
* make sure you zoom in far enough using the new 2x zoom to see the curve and move the selection boundaries to a cross-over point on the x-axis in both channels (otherwise you add impulse artifacts at the boudaries). This applies to most of the DC filters.

This is time consuming, but it has worked very well for me.

[Craig Maier]

Re: Dealing with "essing" distortion.

Big Al,

I think that I may have found part of the problem with your use of the de-esser. In your statements in that earlier thread, you mentioned a gain control set to zero in the De-Esser. The Dynamics Processor does not have a gain control, but the Dynamics Noise Filter does. I think that you may have been trying to use the wrong filter entirely. The Dynamics processor is found under the Effects Menu and should do what you need to accomplish. The Dynamic Noise Filter has no way to deal with overly Siblant "esses." Am I onto something here? Please advise.

[BigAl]

Re: Dealing with "essing" distortion.

Sorry Craig, I see the control I was referring to as "gain" is actually called "output" in the Dynamics processor.

[firekat]

Re: Dealing with "essing" distortion.

Bringing up the Punch & Crunch feature reminds me of a tip that I found on Cnet from a review of one of the versions of the DC product. For us novices a little more enlightenment on just how capable this function may be is certainly in order. When I get a chance I will be experimenting with it some more. Anyway from Lemonade's review I excerpt the following:

"Hiss & more surface noise can be removed from any recording using the 'punch & crunch' feature using HEAVY EXPANSION/BALANCED SPECTRUM & adjusting the output below clipping, (set attack to lowest 1 ms & release to 0.50 for best results) a real find that feature!"

I have tried this and it does seem to work a bit. In my workflow I tried using it after attempting as much noise reduction as possible, as Punch & Crunch is listed under effects I figure it would be best utilized post-noise reduction work. Is this a safe assumption?

Thanks All & Good Luck!

[Craig Maier]

Re: Dealing with "essing" distortion.

The Punch and Crunch feature can be viewed as either a filter or an effect, depending on how it is used. Those using it to modify the dynamics of a recording would probably classify it as an effect. Those using it for noise reduction would look at it as a filter. Nonetheless, it can be used to reduce hiss, as you have discovered. It can also be used to de-ess a recording. I would tend to use it at the end of my processing sequence, since it modifies the dynamics of the recording thereby making it more difficult for the other filters to do their job properly.

[weskola]

Re: Dealing with "essing" distortion.

The Punch and Crunch works extremely well in reducing background hiss as well as de-essing. I use "light" rather than "heavy" and it does a great job.

I have an album that has several areas in which there is a great deal of distortion with a guitar in particular but also several other areas as well. With the guitar distortion there is a great deal of background hiss when the distortion happens.

P&C in light expansion mode will almost totally eliminate this hiss and it helps tremendously in reducing the distortion. I've tried using the dynamics processor in de-ess mode but haven't been too successful in getting it to work to my satisfaction (I'll probably keep experimenting).

I've had a great deal of success in applying a low pass filter at 13KHz/12dB followed by a couple of Virtual Valve passes and finishing up with the P&C in light expansion mode. With the P&C I have to adjust the top two bands to bring back some of the clarity and higher frequencies.

[BobJ]

Re: Dealing with "essing" distortion.

I've too run into this problem (sounds like everybody has) and although I don't know if it is the same situation I found that backing off on the recording levels helps. I get the problem when restoring albums with big brass passages, such as Blood Sweat and Tears and Chicago. Although the rest of the recording is fine, when the brass sections kick in and the music crescendos (verb?), the overload light comes on and I hear the distortion. By reducing the recording levels this does not happen. The problem is that the entire volume of the recording is lower.

[BigAl]

Re: Dealing with \"essing\" distortion.

I've brought this one back to the top to thank those guys who shared their success in de-essing using P&C (especially Weskola for the hint on Light Expansion). I used the following settings (taken from the pst file as the filter does not give exact values for the threshold values):
[PunchAndCrunch]
Set=-129,-60,-129,-129,31,500,100,-90,1,0,0

The first four numbers are the threshold values (-12.9, -6.0, -12.9 ,-12.9) and the rest are from the "Light Expansion Bright" preset, except for output, which I upped to -9dB.

I used this on the offending Ess only and not the whole file.

[BigAl]

Re: Dealing with \"essing\" distortion.

From Craig:[ QUOTE ]
I do not use the presets, but adjust the frequency, threshold, etc. controls manually until I get the desired result.

[/ QUOTE ]
I've been playing around with the Dynamic Processor's De-esser again, and have got it to work at last by fiddling with the frequency and threshold as recommended by Craig. Questions arising are:
* what threshold setting do you start with in trying to pinpoint the frequency?
* also, the ratio setting plays a big role in the attenuation of the sibilance. What values should one start with on this? At what stage of the process should one start playing with the ratio setting?
* in my current project (Rod Stewart) on one track the sibilance is at two different frequencies. Chaining two de-essers together seems to knock out the cymbals. Anyone got any ideas for a workaround on this?

What seemed to work for me was to isolate one of the sibilances and then to fiddle with the sliders until the sibilance was minimised. I then zoomed in on a selection with a large number of these sibilances and fiddled with the sliders some more to try and bring back the cymbals without bringing back too much sibilance. I then ran the filter on the whole file. I'm not sure if this is how other folks use this filter, but would appreciate some comment.

[Ringmaster]

DE-Essing

Is De-essing the same as "Sibilance".

I am restoring four of "The Kendalls" LP's, and I have several areas where Jeannie Kendall has some distortion in the higher frequencies (at least I think it is the higher frequencies).

Big Al, in his last post under this thread, asked several questions, and I don't see any response to his questions. If someone would respond, I think this will help me with this problem; I have everything filtered, except the "---ssss", and I don't want to process anything but the remaining ---ssss.

Regarding BobJ's 1/12/03 post: The only way I have been able to reduce the brass distortion, is to reduce the levels of the Scratch & Crackle filtering, and run them several times at the same or "lesser" settings. This does make for more "Manual Interpolation", which is not a big problem for me.

I ran into this while I was restoring 63 jazz albums (after which I gained a better appreciation of jazz music). It became a "trade-off", do I want the brass (yes), or do I want to tolerate some light distortion. I have found that a lot of the light distortion is not apparent when played on my music system; hwever, you do have to be careful, because some of the "Dolby" play-back modes tends to emphasize the high frequency distortion; however, horn distortion is really noticeable when using the "Dolby" processing.

You have to be really careful, when you intend use "Dolby" play-back processing, because a lot of the high frequency distortion is emphasized, especially in the "Hall" and some of the "Music" processing.

Anybody have some suggested settings for only filtering the "---sss" distortion, that would have little effect on the rest of the recording?

[Craig Maier]

The reduction of Siblance or Essing (sss) is accomplished with the Dynamics Processor found in the Effects menu. Use the De-Esser function to reduce Siblance. You will have to adjust the threshold, frequency and ratio to obtain the desired results. Consider starting out with one of the 4 De-Esser Presets and take it from there. I hope that this helps.