|
Post by mcirish on Apr 1, 2019 9:30:55 GMT -6
I'm mixing a project right now and have run into kind of an odd situation. I found a solution but wanted to see if others have found similar problems and solutions. I'm working with a singer with a huge vocal and dynamic range. During the mix I noticed the vocals had some extreme harshness on the louder/higher parts. Just doing a gentle cut in the 2.5k-4k range helped a bit but made the performance dull. After struggling for a while, (ad doing some reading) I decided to see if there were any odd resonances. I looped the most annoying areas and then did a very tight Q sweep (soloed the band) and listened for anything that jumped out. I did find a couple frequencies that rang out into an oscillation when I hit them. The worst were 3.2k, 4.1K and another around 9.5k. I did a super tight cut of around 6-12db at those frequencies and it completely cured the problem without destroying the overall tone of the voice. From doing a bit of reading, it may be due to overdriving the capsule on the louder parts. That would make some sense, since the lower register of her voice sounded perfect with very little EQ. Have any of you run into that while mixing? The day I was trying to find a solution, I ran across the article about Adele's Rolling in the Deep and the cuts they made to her EQ on the higher parts of the chorus. That's what initially got me thinking about this idea. I normally wouldn't want to do anything drastic as I hate overly processed vocal sounds. But, in this case, it worked and solved the problem. I put this EQ in the first slot and I believe that cleaned it up well enough that the rest of the processing came out much better, since it was not having to deal with these huge resonance peaks. I did try the plugin "Soothe" and it did help but by the time it picked up the worst issues, the tone was just lost in the voice. Also, Soothe is a gigantic CPU resource hit as well as GPU hit. I could have used it off-line but i felt the tight EQ cuts sounded better. Attached is an EQ from the Adele session I mentioned.
|
|
|
Post by Johnkenn on Apr 1, 2019 9:36:29 GMT -6
Yep. All the time. I find I have to notch ringing overtones with acoustic guitar a lot. Usually bell tones around 1khz. Try Radix Surfer EQ. It has been a godsend for a lot of vocals for me...not notching, but it follows problem resonances around.
|
|
|
Post by Johnkenn on Apr 1, 2019 9:38:22 GMT -6
I kind of assumed it was the room. Also - I’ve actually been standing farther back from the mic these days. It’s such a temptation to get close for the proximity effect...but I guess a lot of this could be attributed to bad mic technique.
|
|
|
Post by jcoutu1 on Apr 1, 2019 9:38:25 GMT -6
I usually use dynamic eq on resonances like that.
|
|
|
Post by svart on Apr 1, 2019 9:43:17 GMT -6
Vocals yes, all the time. Generally I've found that the singer needs to step back a little more and it'll even out.
|
|
|
Post by mcirish on Apr 1, 2019 10:02:29 GMT -6
I usually use dynamic eq on resonances like that. I was using Waves F6 in dynamic mode but these resonances were not quite so dynamic. They were there a lot. Funny thing is, the tight cuts sounded more natural to my ears, but I might give a dynamic EQ another try on the next song. I usually do use it, but more for a wider dynamic cut. BTW, the EQ I used for the cuts is Crave EQ. It's surprisingly good. I sort of stumbled across it a couple months ago and just really like the workflow and the sound. Being able to solo a band was extremely useful.
|
|
|
Post by trakworxmastering on Apr 1, 2019 10:04:06 GMT -6
I usually use dynamic eq on resonances like that. Yes! Dynamic EQ or multiband compressor will kick in only when needed. Much better result in most cases. Also, I've often found that resonances come from the singer's voice rather than anything to do with the mic.
|
|
|
Post by mcirish on Apr 1, 2019 10:08:11 GMT -6
I'm not sure a multiband compressor would have worked in this situation. I tend to think of those as a gentle control over a few bands. In this case, it was some pretty sharp cuts at a few frequencies with a super high Q. I just wanted to notch out that resonant ringing at those particular frequencies. It wasn't a matter of subtle control. It was a 12db cut, but so tight as it only hit the ringing and nothing else around it.
|
|
kcatthedog
Temp
Super Helpful Dude
Posts: 15,058
Member is Online
|
Post by kcatthedog on Apr 1, 2019 10:16:56 GMT -6
|
|
|
Post by mcirish on Apr 1, 2019 11:18:00 GMT -6
Yep, I did try Soothe. I had that in my first post. It's a cool plugin but has a very high CPU/GPU demand. I'd probably have to use it off-line on individual tracks and bring the audio back in to the DAW. With the track count I have and the amount of CPU hungry plugins already in the session, it was a no-go. Higher settings in Soothe get a little odd sounding. I think it could make subtle changes pretty nice. I also tried bx_refinement. That seems to be a wideband EQ that just pulls down 2k-5k. I think I can do that with an EQ more precisely. TDL has DeEdger, but I found it to be way too subtle to even make a difference.
|
|
|
Post by svart on Apr 1, 2019 11:24:34 GMT -6
I try not to cut resonances on solo vocals these days. Sometimes they aren't heard in the mix and cutting them might make the vocal sound strange in a mix context..
|
|
|
Post by theshea on Apr 1, 2019 13:25:39 GMT -6
rokyo dawn nova GE works great for this kinda problems too. i am using it all the time.
|
|
|
Post by mcirish on Apr 1, 2019 13:49:07 GMT -6
In the past, I would always use dynamic EQ, either Nova or F6 at the start of the chain to get rid of any of the harshness. In this very specific case I couldn't get it to work well enough. It wasn't a general 3K reduction on peaks. That wasn't working. The cuts I needed were super narrow. I may have even cut 18dB at 4.1k (Q=12). I almost wonder if there is a problem with the mic tube. I recently sent my Wunder CM7GS/M7 to Mike for some maintenance and he found the tube (very low hours) to have some odd resonances in the lower mids. That mic was not used on this recording but I do know it also had a Telefunken EF800 tube in it. Maybe it has some issues too. I know we replaced the tube at some point of the recording because it developed noise. Now I wonder if it was the old tube or the new tube that might have issues. Or, maybe there just are not any more good NOS tubes around and what we end up buying is all rejects? Honestly, if the tubes didn't sound nice (normally) I'd go with FET mics. Much less maintenance. :-)
I should drop Bowie an email and get his take on it.
|
|
|
Post by the other mark williams on Apr 1, 2019 15:03:05 GMT -6
Not sure if you tried this, mcirish, but if it's just on the louder/higher sections, you may find it helpful to cut those sections from the main vocal track, paste them into a new track, and do the extreme EQ treatment there. That way, you can avoid the extreme cuts on most of the vocal track. Just a thought.
|
|
|
Post by jeremygillespie on Apr 1, 2019 18:43:38 GMT -6
Not sure if you tried this, mcirish, but if it's just on the louder/higher sections, you may find it helpful to cut those sections from the main vocal track, paste them into a new track, and do the extreme EQ treatment there. That way, you can avoid the extreme cuts on most of the vocal track. Just a thought. You can also automate the EQ on and off. At least in PT you can. I’d assume you can do the same in other DAWS.
|
|
|
Post by notneeson on Apr 1, 2019 19:32:06 GMT -6
Not sure if you tried this, mcirish, but if it's just on the louder/higher sections, you may find it helpful to cut those sections from the main vocal track, paste them into a new track, and do the extreme EQ treatment there. That way, you can avoid the extreme cuts on most of the vocal track. Just a thought. You can also automate the EQ on and off. At least in PT you can. I’d assume you can do the same in other DAWS. Audiosuite sometimes feels faster to me, like for filtering plosives, I’ll just jump from trouble spot to trouble spot and print it. Always after duplicating the playlist.
|
|
|
Post by Johnkenn on Apr 1, 2019 19:57:55 GMT -6
Used soothe 2 (or whatever the newest one is) on a sibilant vocal today and it worked better than a deesser. Very cool.
|
|
|
Post by chessparov on Apr 1, 2019 20:44:50 GMT -6
Great thread! Chris
|
|
|
Post by the other mark williams on Apr 1, 2019 23:56:09 GMT -6
Not sure if you tried this, mcirish, but if it's just on the louder/higher sections, you may find it helpful to cut those sections from the main vocal track, paste them into a new track, and do the extreme EQ treatment there. That way, you can avoid the extreme cuts on most of the vocal track. Just a thought. You can also automate the EQ on and off. At least in PT you can. I’d assume you can do the same in other DAWS. Absolutely. Great point. Though of course if it’s a latency-inducing EQ, you may have to look for an internal bypass to automate, rather than the DAW plugin shell’s bypass.
|
|
|
Post by WKG on Apr 2, 2019 2:27:49 GMT -6
I usually use soothe but PA's DSM can be useful. I'll use the capture function to take a snapshot of a section closer to what I want and use threshold and ratio to apply it to the difficult areas.
|
|
|
Post by Ward on Apr 2, 2019 6:13:56 GMT -6
This thread is yet another awesome source of useful info here.
Thanks everyone!
|
|
|
Post by mcirish on Apr 2, 2019 7:54:34 GMT -6
I really appreciate all the advice. I have to say I'm still struggling a bit with it but with a combination of approaches, I think I will get this one done. I do believe I will have to automate an EQ a little so I don't cut the life out of the rest of the vocal track. Who would have thought that the human voice could be so hard to make sound right. Such a huge dynamic and timbre range!
I may even just buy Soothe and do a bit of off-line processing if I can't get what I want any other way.
|
|
|
Post by Ward on Apr 2, 2019 8:32:16 GMT -6
I may even just buy Soothe and do a bit of off-line processing if I can't get what I want any other way. I was thinking the same thing, actually. Sometimes I 'burn in' processing to avoid any slowdowns, bogdowns or CPU latency issues.
|
|
|
Post by Vincent R. on Apr 3, 2019 21:49:43 GMT -6
I’d be curious to know what your signal chain was. Changing the mic on the way in is probably the best route. I know that’s not always an option. It really takes a lot of sound pressure to overdrive the capsule. Trust me I’m freaking loud and my wife is louder. The loudest the human voice gets is the mid 120s dB Level, and that’s opera singers. I guess it depends on what your mic is rated for. Most distortion would be from a tube or transformer in a circuit somewhere. My guess is it’s just a frequency in her voice that is prevalent in her top range. Not uncommon. On a big voice distance can be your friend. 1 - 1 1/2 ft is usually where I start. Part of it could be enhanced by The proximity.
That said, I had a coloratura soprano in the studio singing through my U87ai and when I went to mix it I found is it really harsh when she went up into her top range. I ended up using IK Mic Room to make the U87ai sound like a TLM170. That actually helped a lot. I had EQed the vocal to death and it got dull and ugly, and I could’t stand it. Digital remiking did the trick. From there it was mostly simple EQ moves.
|
|
|
Post by mcirish on Apr 4, 2019 10:14:15 GMT -6
The signal chain was real simple. Miktek CV4 with a BN K7 capsule into a John Hardy M1, Lynx Aurora 16. Super simple and clean. Nothing was pushed. The particular mic has always been one of the smoothest sounding mics I've ever heard or used. I'm really thinking she just got up on the mic a bit too much in the moment and the capsule added some harmonics that were not flattering. There was a pop filter, but I wouldn't be surprised if she was 4" from the capsule. It might have been fine had she backed of on the higher parts. But, hindsight is 20/20. At the time, the tracks sounded just fine, but I hadn't put much in the way of processing on it. I really needed this vocal to sit in front of the mix, so there are a couple compressors. Nothing heavy handed though.
After all the EQ messing and notching I did, i felt like I needed to hear what was originally there so I setup a couple channels with a duplicate of the vocal take, so i could compare easier. All my notches in EQ helped on the loud, upper range parts but I did notice it took some of the life and character from the vocal, especially in the lower range. I went back to my usual plugin chain and started over without the EQ doing any notching. Just some subtle EQ and a couple deEssers to hit some of the harshness. I also used a dynamic EQ in some of the louder areas to tame it, but nothing too drastic. Being more subtle brought the vocal performance back to life. I still need to do a bit more work in the loud areas but I'm going to process them separately so it doesn't mess with the lower vocal tone.
Funny thing is that I've used this mic and preamp on this voice over 100 tracks. Just this last round had these issues. Still makes me wonder if the tube I put in to replace a noisy one (hiss) has some problems. Half art, half science, and half bewildering... Yes, I know the math doesn't add up. :-)
|
|