Help Me Find a De-Esser Plugin For Spoken Word/Podcast

If it's purely a problem of the ess sound being too prominent and everything else being okay, then moving the mic won't help. For this it's either change the performance or use frequency-specific dynamic processing.

If the whole sound of the voice is too bright then moving the mic, using a different mic, or using eq, can help.

If it's a problem of the esses blowing into the mic causing a rumbling sound, then moving the mic or using a pop screen can help.

Which one of those scenarios is the case?
 
Man, thanks for the continued discussion guys. Sorry for the long absence. I was out of town all weekend and just got around to getting back over here.

I should clarify something I worded poorly in my previous post. I didn't originally have the mic pointed at the ceiling. I moved it to the ceiling to try the advice of changing the mic direction. I have my mic oddly positioned to begin with in the sense that it is set sideways, instead of upright, to make room for hands/papers/etc underneath the mic.

Has a sample been posted?
Regarding a sample, there's a link to our site in my signature, but in case that isn't working, it is Four TV Sports Podcast

I think overall the sound quality is pretty good, and substantially better than 90+% of podcasts. However, the "sss" sounds are the one thing that really seems to be standing out at me.

That said, if you have suggestions in addition to the de-essing problem, please let me know either here or in a PM.

Thanks again for all the helpful hints and criticisms.
 
Last edited:
I should clarify something I worded poorly in my previous post. I didn't originally have the mic pointed at the ceiling. I moved it to the ceiling to try the advice of changing the mic direction. I have my mic oddly positioned to begin with in the sense that it is set sideways, instead of upright, to make room for hands/papers/etc underneath the mic.

This is referred to as speaking off axis. If you sit with the mic directly in front of your mouth, place it about two inches below your chin and turn it 15 to 20 degrees to the right. This will help with plosives as well as de-essing.

Regarding a sample, there's a link to our site in my signature, but in case that isn't working, it is Four TV Sports Podcast

I think overall the sound quality is pretty good, and substantially better than 90+% of podcasts. However, the "sss" sounds are the one thing that really seems to be standing out at me.

Your last ep shows it was done in mono @ 96/44.1. This is fine. Your LUFS was a -22, your Loudness Range (LRA) was 5.6dbs and your dbTP was a -3.8dbs. The AES specs are between a -16 to -20LUFS, LRA 4db max with a -1dbTP.

View attachment 100645

That said, if you have suggestions in addition to the de-essing problem, please let me know either here or in a PM.

Thanks again for all the helpful hints and criticisms.

De-essers are nothing more than using an eq to cut or reduce certain frequencies. You can build your own and it is very simple to do. You would build one for each of your hosts and just make sure that you use the plugin in the proper order. Once you learn how to do it, you can quickly build one for any guest you may have.

Now with that being said, I have analyzed close to 7,500 podcasts and IMHO, you really do not have a problem with your sss and your audio sounds great! If I have time this morning, I will do a video for you using Reaper and show you how to sweep the freqs to custom build your de-essers.
 
Looking for guidance on a de-esser for spoken word/podcast. I know usually people would say to de-ess through editing, but for spoken word, and weekly episodes that amount to 1.5-2 hours of content each week, that's not really an option.

Is there a plugin that anyone would recommend? For some reason, the stock plugin isn't installed and I can't find it for the life of me.

I'm on a Mac using Cubase 4 (b/c I'm too poor to upgrade my DAW ha)

I'd prefer a plugin than an outboard gear. I'd rather be able to edit fully within the software. I also don't want to complicate the signal chain too much, so not having to do side-chains, sends, etc, would be helpful, as I'm not super familiar with that.

Thanks in advance! Any help is appreciated.

Also, I've done a lot of Google searching and haven't seen much on this topic that isn't 3-4 years old, or threads that recommend plugins that no longer work for Macs. I'd use SpitFish, but it's not Mac compatible anymore. Same story for a few others I looked into.

I was considering buying the Waves plugin, but wanted to ask around first.

waves plugins are severe resource hogs. here is some that you can use: Free VST Plugins: The Best VST De-Esser Effects...

I would also find out what vst engine version. Because some vst plugs require VST ver2.4 and I don't know what version you are running. So you need to test drive the plug just to make sure its going to work with Ye Olde DAW....
 
An unpopular opinion, I'm sure, but the difference between a dynamic mic and a condenser is really mostly a static EQ curve. I mean, yeah overall sensitivity, but once you normalize for that, even proximity effect is just an EQ adjustment.

That's why things like Anteres Mic Modeler can work as well as they do. It "undoes" the curve of the mic you recorded with and then applies the one from the mic you want to fake. Maybe a little more sophisticated than just messing with your EQ, but if you're not trying to emulate anything specific, you can get pretty decent results.

If a dynamic actually helps, then it's probably because your condenser is giving you too much high end in general, and probably less low end from proximity. Shelves on either end can go a long way toward just solving the issue. It's a heck of a lot easier to EQ a condenser to sound like a dynamic than the other way around. You can always add a deesser after, if you find there actually is also a dynamic aspect to your issue. Since you've already got the condensers...
 
Just upload a sample mp3 here on the forum so I can download and analyze it.

This is their latest audio file from their website. It is a five min clip with a ton of sss's. I tried to upload their entire show but it is exceeding the upload limits for the forum. I hope I am not overstepping any boundaries, but they may not know how to upload audio files to the community or may be restricted due to their post count. View attachment 100673
 
Looking for guidance on a de-esser for spoken word/podcast. I know usually people would say to de-ess through editing, but for spoken word, and weekly episodes that amount to 1.5-2 hours of content each week, that's not really an option. Is there a plugin that anyone would recommend? For some reason, the stock plugin isn't installed and I can't find it for the life of me.

I did this video using your audio from your last EP. If your DAW has an eq and a spectrum analyzer, this is how you can build your own. It's simple. I also gave a shout out to this community as well as your show. I have one vid with over 11,000 views and this vid may send some listeners your way as well as potential members to this community. If anyone has any objections, just let me know and I will edit out any references to either site.

Just so you know, regardless of whether you use a free DE-esser or one that cost hundreds of dollars, they all operate on the same principal as explained in the video and you still need to tell the plugin what freqs you want to duct/reduce/cut or however you want to phrase it. If you have any questions, just post them back to this thread.


 
Just so you know, regardless of whether you use a free DE-esser or one that cost hundreds of dollars, they all operate on the same principal as explained in the video...

Not that it's all that important, but there are two main ways de-essing is done, band limited compression or full band compression with band limited detection.

The method I use is somewhat different from both: frequency sensitive compression, where only the frequencies that exceed a threshold curve are affected.

ReaFIR de-esser.png
 
Last edited:
Not that it's all that important, but there are to main ways de-essing is done, band limited compression or full band compression with band limited detection.

The method I use is somewhat different from both: frequency sensitive compression, where only the frequencies that exceed a threshold curve are affected.

View attachment 100691

I agree with you 100%. The point I failed to make is that the freqs must first be detected and than dealt with for either free or payed de-essers. I have always been taught that the less feqs affected on each side of the notch, the better the results. I think I will look into your method in more detail. It does make perfect sense that if compressed they would sound more natural maybe.
 
Hi,
Thanks for your reply.
Yes, I did see them as I even verified the block size.

Now I redid the split and join. The result is the same.

I re-read the memory 3 times, all have diff when verifying:
For 1st read,
diff: 013da000
diff: 10b7b000
For 2nd read,
diff: 10b7b000
diff: 122af000
For 3rd read,
diff: 013da000
diff: 10b7b000
diff: 122af000

Does it mean the memory chip has some kind of problems?

Thanks again.
Phil
 
Back
Top