|
Post by Johnkenn on Jan 25, 2020 17:03:25 GMT -6
I’ll preface this with I probably don’t know what I’m talking about...but why hasn’t anyone designed any type of plug that can identify esses by their waveforms? From everything I see, they’re distinct waveforms. Maybe a tab to ess thing or something. That would be really cool.
|
|
|
Post by MorEQsThanAnswers on Jan 25, 2020 19:47:54 GMT -6
I’ll preface this with I probably don’t know what I’m talking about...but why hasn’t anyone designed any type of plug that can identify esses by their waveforms? From everything I see, they’re distinct waveforms. Maybe a tab to ess thing or something. That would be really cool. Airwindows deess!!!
|
|
|
Post by Johnkenn on Jan 25, 2020 20:21:24 GMT -6
I’ll preface this with I probably don’t know what I’m talking about...but why hasn’t anyone designed any type of plug that can identify esses by their waveforms? From everything I see, they’re distinct waveforms. Maybe a tab to ess thing or something. That would be really cool. Airwindows deess!!! Is that how it does it?
|
|
|
Post by matt on Jan 25, 2020 20:38:51 GMT -6
The Fabfilter DS shows something similar in it's UI. One thing I've always wanted in that plug is an extended frequency range so that any nasty waveforms, at any frequency, could be suppressed. It works very well, set and forget most of the time.
|
|
|
Post by Johnkenn on Jan 25, 2020 20:55:09 GMT -6
The Fabfilter DS shows something similar in it's UI. One thing I've always wanted in that plug is an extended frequency range so that any nasty waveforms, at any frequency, could be suppressed. It works very well, set and forget most of the time. Yeah, but nothing seems to work all the time.
|
|
|
Post by nick8801 on Jan 25, 2020 20:55:22 GMT -6
I think ozone does as well.
|
|
|
Post by matt on Jan 25, 2020 22:23:08 GMT -6
It works very well, set and forget most of the time. Yeah, but nothing seems to work all the time. True. Sometimes I'll edit the waveform directly.
|
|
|
Post by Johnkenn on Jan 25, 2020 22:29:59 GMT -6
I think ozone does as well. Ozone has a deesser? Besides the dynamic eq?
|
|
|
Post by Johnkenn on Jan 25, 2020 22:30:47 GMT -6
And really, that’s what a deesser is I guess. A dynamic eq
|
|
|
Post by nick8801 on Jan 26, 2020 0:15:46 GMT -6
I think ozone does as well. Ozone has a deesser? Besides the dynamic eq? Sorry, izotope RX does...I have the bundle so I mix them up sometimes. I think Slate has a great de-Esser too. The eiosis one. Pretty sure these both work in the way you’re describing.
|
|
|
Post by M57 on Jan 26, 2020 5:04:06 GMT -6
I probably don't know what I'm talking about either but it seems to me that when I draw automation to mitigate esses, the curves do not necessarily resemble a traditional attack/release form. The middle of the offending wave form usually doesn't seem to be what bothers me. Surprisingly, I've found a lot of the problem happens toward the end of the wave form. Given that, slow attack and fast release might seem to be the best setting on a dynamic eq, but sometimes just looking at the transient makes me think otherwise. Thinking about it conceptually with the understanding that we're talking about an artifact of close micing, the goal shouldn't be to mitigate the esses so much as it should be to make them sound more natural, and I wonder that a traditional compression oriented solution isn't necessarily the best solution. Perhaps that's why so many of us use multiple tools like de-essers, automation and clip gain simultaneously to tackle the problem.
|
|
|
Post by M57 on Jan 26, 2020 5:19:54 GMT -6
..piggy backing on my last post..
I was thinking that maybe the end of the ess is such a big part of the problem because of the way it is enhanced by reverb and/or any delay effects we typically use on vocal tracks. I wonder if it might be possible to isolate and copy all the esses, move them to a separate track, and then NOT process them like the vocal track. Instead, treat them separately; for instance put no reverb on them - or at least use the same reverb but carve out a swath of the offending frequency range.
|
|
|
Post by theshea on Jan 26, 2020 5:57:17 GMT -6
come on, don't be lazzzzzzzzy ... automation by hand and problem solved. no need for a de-esser though i do use waves de-esser sometimes. or tokyo dawn dynamic eq with smart function lately. but if it ain't working right away i automate the esses.
|
|
|
Post by M57 on Jan 26, 2020 8:40:06 GMT -6
come on, don't be lazzzzzzzzy ... automation by hand and problem solved. Not always, and certainly not for me always. Often straight automation too noticeably sucks the air out of the track for that brief moment and it's clear that a more surgically oriented frequency-specific remedy is necessary.
|
|
|
Post by jcoutu1 on Jan 26, 2020 9:23:53 GMT -6
..piggy backing on my last post.. I was thinking that maybe the end of the ess is such a big part of the problem because of the way it is enhanced by reverb and/or any delay effects we typically use on vocal tracks. I wonder if it might be possible to isolate and copy all the esses, move them to a separate track, and then NOT process them like the vocal track. Instead, treat them separately; for instance put no reverb on them - or at least use the same reverb but carve out a swath of the offending frequency range. Put a heavy handed deesser on the aux track before the reverb. This will help a ton.
|
|
|
Post by Guitar on Jan 26, 2020 9:43:53 GMT -6
Fab Filter DS
Waves Sibilance
I use the Waves since it's way cheaper than the Fab Filter option (I got it for free.)
They both seem to work perfectly well with minor adjustments, and they show you where the esses are in their real time displays.
|
|
|
Post by theshea on Jan 26, 2020 11:08:41 GMT -6
come on, don't be lazzzzzzzzy ... automation by hand and problem solved. Not always, and certainly not for me always. Often straight automation too noticeably sucks the air out of the track for that brief moment and it's clear that a more surgically oriented frequency-specific remedy is necessary. i thnk you can make automation on esses sound very naturally. it takes time though.
|
|
|
Post by M57 on Jan 26, 2020 12:05:17 GMT -6
Not always, and certainly not for me always. Often straight automation too noticeably sucks the air out of the track for that brief moment and it's clear that a more surgically oriented frequency-specific remedy is necessary. i thnk you can make automation on esses sound very naturally. it takes time though. I have spent many many hours automating esses. It doesn't matter how much time I spend on each ess. I'm quite sure I've spent more than a half an hour working on just one ess experimenting with different approaches (and on more than than one or two occasions). I've spent an equal amount of time trying to dial in the holy grail magic setting for my esses on maybe a half dozen de-essers. My best results tend to involve a combination of techniques, the most effective of which is clip gain. But clip gain is harsh around the edges with artifacts ..and that sucking sound, so it usually needs to be applied gently and tempered with automation and/or dynamic equalization. I'm not complaining about having to put in the work. Like JK, I'm just somewhat surprised that a more natural sounding turn-key and universally effective solution hasn't been developed yet in this age of AI computing. Oh, and can I add inexpensive? ;}
|
|
|
Post by notneeson on Jan 26, 2020 12:10:35 GMT -6
I’m fine with the stock PT desser.
Things that have helped me: assist the plugin by clip gaining standout esses. De-ess last, after all vocal processing is final. And don’t de-ess in solo.
|
|
|
Post by theshea on Jan 26, 2020 14:33:21 GMT -6
i thnk you can make automation on esses sound very naturally. it takes time though. I have spent many many hours automating esses. It doesn't matter how much time I spend on each ess. I'm quite sure I've spent more than a half an hour working on just one ess experimenting with different approaches (and on more than than one or two occasions). I've spent an equal amount of time trying to dial in the holy grail magic setting for my esses on maybe a half dozen de-essers. My best results tend to involve a combination of techniques, the most effective of which is clip gain. But clip gain is harsh around the edges with artifacts ..and that sucking sound, so it usually needs to be applied gently and tempered with automation and/or dynamic equalization. I'm not complaining about having to put in the work. Like JK, I'm just somewhat surprised that a more natural sounding turn-key and universally effective solution hasn't been developed yet in this age of AI computing. Oh, and can I add inexpensive? ;} yeah clip gain can be drastic. what i do is paint the automation in logic with at least 2 points. so i can make the reduction less steep or just as i want to. and most often 5db reduction is enough.
|
|
|
Post by rowmat on Jan 26, 2020 15:09:13 GMT -6
'Esses' are easy to identify once you've seen a few zoomed in.
I have not yet found anything, plugin, hardware etc. that sonically beats this for de-essing.
1. Duplicate the vocal track
2. Delete all the audio between the 'esses' in the duplicated track just leaving the 'esses. (This method can also be applied to reduce hard consonants, plosives etc if required.)
3. Glue the 'esses' only track together forming a single wavefile. (The 'esses' only track must remain in sample perfect sync with the original vocal track)
4. Assign the 'esses' track to the same bus as the main vocal track.
5. Phase reverse the 'esses' track.
6. Adjust the level of the 'esses' only track to control the amount of phase cancellation of the 'esses' occurring in the vocal track.
|
|
|
Post by M57 on Jan 26, 2020 16:12:25 GMT -6
'Esses' are easy to identify once you've seen a few zoomed in. I have not yet found anything, plugin, hardware etc. that sonically beats this for de-essing. 1. Duplicate the vocal track 2. Delete all the audio between the 'esses' in the duplicated track just leaving the 'esses' and also hard consonants, plosives etc if preferred.. 3. Glue the 'esses' only track together forming a single wavefile. (The 'esses' only track must remain in sample perfect sync with the original vocal track) 4. Assign the 'esses' track to the same bus as the main vocal track. 5. Phase reverse the 'esses' track. 6. Adjust the level of the 'esses' only track to control the amount of phase cancellation of the 'esses' occurring in the vocal track. For some reason, I have had fair to poor results using this method. For one, the track has to be PERFECTLY synced to the sample level. One tick off and you're making things worse, not better. I do like the idea of re-glueing the esses. At least that way, it's all or nothing if something gets inadvertently moved.
|
|
|
Post by rowmat on Jan 26, 2020 19:13:24 GMT -6
The very reason it works so well is exactly because of the sample accuracy. I would check it, glue it and check it again. Then I suggest making sure both tracks are rendered back to the start of the timeline so you have a fixed starting point in case you move something untentionally. If you somehow manage to knock it out of alignment then of course it's not going to be any good at all. 'Esses' are easy to identify once you've seen a few zoomed in. I have not yet found anything, plugin, hardware etc. that sonically beats this for de-essing. 1. Duplicate the vocal track 2. Delete all the audio between the 'esses' in the duplicated track just leaving the 'esses' and also hard consonants, plosives etc if preferred.. 3. Glue the 'esses' only track together forming a single wavefile. (The 'esses' only track must remain in sample perfect sync with the original vocal track) 4. Assign the 'esses' track to the same bus as the main vocal track. 5. Phase reverse the 'esses' track. 6. Adjust the level of the 'esses' only track to control the amount of phase cancellation of the 'esses' occurring in the vocal track. For some reason, I have had fair to poor results using this method. For one, the track has to be PERFECTLY synced to the sample level. One tick off and you're making things worse, not better. I do like the idea of re-glueing the esses. At least that way, it's all or nothing if something gets inadvertently moved.
|
|
|
Post by MorEQsThanAnswers on Jan 27, 2020 1:04:19 GMT -6
It targets waveforms with high-slew rate like “esses” and can filter/turn them down separate from the rest of the source
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jan 27, 2020 12:12:04 GMT -6
come on, don't be lazzzzzzzzy ... automation by hand and problem solved. no need for a de-esser though i do use waves de-esser sometimes. or tokyo dawn dynamic eq with smart function lately. but if it ain't working right away i automate the esses. My god, do I hate automating S's. I'll go for a de-esser every single time as a first line of defense.
|
|