Vocals: 7K-10K grainy/distorted - why?

[Scott Collins]

Greetings!

I'm getting a grainy kind of distortion in the 7k-10k area when recording vocals*, most apparent around 7.5K or so (but sometimes this area seems to change a bit). This isn't an in your face super-obvious thing (i.e. the casual listener might not notice, though I'm sure it would impact their experience anyway), but I hear it and it can get emphasized when processing the vocals (eq/compression). ("Distortion" also may not be the right word - it sounds kind of like something is resonating in this area.)

Equipment:

Studio Projects C1 (with pop screen)
Studio Projects VTB-1
RME 2496 PAD sound card
Cubase SX2


Things I've tried:

Lowering input/output levels on VTB-1 (I don't think they were too high to begin with, but I tried anyway)
Turning the tube drive off on the VTB-1
Changing the mic cable
Recording with an SM-58 (I found essentially the same issue when using this mic)
Changing mic position and distance



Any ideas how I can fix this? Is this, perhaps, a problem with my particular VTB-1 unit? Could this be a soundcard issue?

Any ideas/suggestions will be appreciated!


*(I think I've found the same issue on some acoustic guitar tracks, but I haven't gone over these as completely yet.)

[Alan Hyatt]

Hard to say...

Be sure you set the meter switch to input and set the gain until you are flashing red occaisionally

Switch the meter switch to output and set the output to +12 to the card...then you are set right. If this distortion is still occuring, let us know, but it could be a lot of things.

The VTB-1 can be used to generate two distinct outputs by using its Insert Jack as an additional output. The insert point is after the HPF, but directly before the Blend Control. The signal feed from the insert jack is before any tube circuitry, so the additional output will always be a SS only signal. The insert cable needs to be wired as a “Borrow Cable” by tying the Tip and Ring connections together.

This may help to eliminate the tube circuit as the culprit. If you try it and the distortion is still there, we should look at the card, if the distortion is gone, we should look at the VTB1...

Keep us posted!

[Scott Collins]

Alan,

Thanks a bunch for the troubleshooting idea. I'll get to work on this and post back what I find.

-Scott

[cellus]

You know I've been having this same issue with my recordings although my path is a bit different. I've actually given up on this mic because of the slight distortion and began using the AKG 414EB. I like the C1 mic; I just do not know why the signal is being distorted around those frequencies. I've even tried to have the artist back away from the mic thinking they might be too close, but that wasn't the issue cause now they just sounded faint. I've also tried to bypass my compressor and that wasn't problem either although the compressor did bring the distortion out a bit more. Below is my signal path:

1. pop screen
2. C1 Mic
3. Vintech Dual72 (PreAmp)
4. DBX 160A (compressor)
5. RME ADI-DS (DA/AD Converter)
6. Frontier Design Apache (Digital Patch Bay)
7. RME HammerFall DIGI 9652 (SoundCard) / Steinberg Cubase SL (DAW)
or
7. Fostex D2424 (Digital Recorder)

If you find out something, could you please post it.

thanks,
cellus :)

[hargerst]

Ok, 7-10kHz is where a lot of LDC diaphragms are tensioned to. It could be you're exciting some resonances in that area.

Try placing a pencil vertically in front of the diaphragm, using a couple of rubber bands to secure it.

Try singing into the mic slightly from the side.

Try placing the microphone forehead high and aimed at your mouth.

Lower the mic and try singing across the top.

Report back if any of these suggestions: 1) made the problem lessen or disappear, or 2) made the problem worse.

[Scott Collins]

hargerst:

Thanks for the tips - I will try them all.

I've hesitated to report back yet as I haven't found my experiments to be conclusive. I tried Alan's approach of bypassing the tube section of the pre-amp. I noticed some small difference, but the bottom line was that the problem doesn't appear to be there.

I actually think your point about exciting some resonance could be on track... like I think my voice itself may have a little grainy or problematic sound in this area on certain passages (e.g. note/dynamic/etc).

Even though I tried to sing "natural" during my clinical-type recording test, evidently I wasn't. I didn't get the resonance so much then, but when I went back to cutting more vocals... there was the problem. (When the problem is bad, it really bites.)

I think the overall problem may be due to a combination of factors. (This is something I'm becoming more sensitive to in audio matters... I tend to look for the one, fundamental, slam-dunk issue... but sometimes it just isn't like that.) I think it could be a combination of vocal technique, mic placement (possibly a big factor here), the particular reaction of the VTB-1 (a little bump in response around 10K or so?), and possibly mixing techniques.

I have tried some mic placement experiments, but not the ones you note so those will be next on the list.

If I could knock this problem out I think I could get pretty close to what the average listener would consider a great vocal sound. I've got some of the nice UAD compressors plugins to work with and some decent EQs, so we'll see...

Thanks again - still working on this one!
Scott

[hargerst]

And don't forget the impedance switch on the VTB-1 - try both positions; it won't hurt anything.

[Brent Casey]

Ok, 7-10kHz is where a lot of LDC diaphragms are tensioned to. It could be you're exciting some resonances in that area.

Try placing a pencil vertically in front of the diaphragm, using a couple of rubber bands to secure it.

Try singing into the mic slightly from the side.

Try placing the microphone forehead high and aimed at your mouth.

Lower the mic and try singing across the top.

Report back if any of these suggestions: 1) made the problem lessen or disappear, or 2) made the problem worse.


7kHz-10kHz would be an extremely high tension for an LDC. The tension on a C1 is more in the 1kHz-1.2kHz range.
A higher tension of ~7kHz may be more prevalent on a SD omni, for instance.
As far as the VTB-1 goes, there should be no emphasis anywhere in the frequency response. It is by design, a wideband amplifier with a ruler flat response across all frequencies of concern.
The same goes for the circuitry of the C1. It is quite linear as well. The capsule is another story. LDC transducers generally look like Kansas throughout most of their frequency spectrum and then start to look like the Rockies in the 2kHz-12kHz range. (Yes, I am generalizing while simultaneously exagerating. My wife can attest to this).
The C1 does have a presence peak around 10kHz, but there is not a lot of emphasis in the 7.5kHz area.
The SM58 in the 7.5k range should not be peaky at all. This is probably by design so as to keep feedback to a minimum.
A resonance in the ~7.5kHz range, would likely present itself as sibilance, although you have not described it as such.


Sincerely,

Brent Casey

[hargerst]

Brent,

Perhaps I didn't hear Stephen right. We were talking (actually, mostly he was talking) in general about LDC mic weirdness up around 7 and 8kHz, and I thought he mentioned that was due to the diaphragm tensioning. I must have misheard him.

[Nik Keefe]

Scott

Is there any chance of you sticking a clip of this phenomenon up? There are some highly trained ears here who might be able to give you some ideas of what your issues sounds like it could be (not me though!).

Nik

[Brent Casey]

Brent,

Perhaps I didn't hear Stephen right. We were talking (actually, mostly he was talking) in general about LDC mic weirdness up around 7 and 8kHz, and I thought he mentioned that was due to the diaphragm tensioning. I must have misheard him.


Was this regarding resonator or aperiodic capsules? The resonators are tensioned quite a bit higher than the aperiodic.

Brent

[hargerst]

Was this regarding resonator or aperiodic capsules? The resonators are tensioned quite a bit higher than the aperiodic.

BrentWe (he) was talking about LDC's in general, and then it was mostly about the newer U87's.

[Brent Casey]

We (he) was talking about LDC's in general, and then it was mostly about the newer U87's.


Yeah, he was probably discussing some 7k weirdness with LDC's, but I guarantee you he wasn't tensioning K67/87 diaphragms that high.

Brent

[Scott Collins]

Nik,

I'll try to post a few samples ASAP. I've had my testing delayed due to being sick last week and some other stuff that has kept me busy. But if someone could help me identify the problem that would help my trial/error process.

(Forgive the delayed response - I stopped getting email notifications on this thread for some reason.)

[Scott Collins]

OK - I just posted some files to my website at http://www.wscollins.com. I've posted some raw vocal files (24 bit @ 44.1) as well as some short clips of the tunes the clips belong to (showing them in context). I'm hoping doing it this way will help uncover whether this is a problem with the original files, a mixing problem, or possibly a combo of both.

I'm pretty sure I could help mask the problem I hear by cutting narrow somewhere between 7.5K and 10K, but I didn't do that on any of the tune clips 'cause I wanted the problem (at least as I perceive it) to come across.

Thanks!

PS: The tune clips are 16-bit, 44.1. I'd be happy to post smaller, MP3 versions, but I guessed these would be preferred.

[atma]

huh, that is strange. i definately hear what you're talking about just listening to the first vocal clip.. very crunchy in the top end.. my first thought was it might just be the grainyness of your voice, but after running a HP filter on it and listening to it a while it almost does sound like a kind of harsh digital distortion of some kind, almost like a bit-decimator, which is what i'd assume poor a/d convertors might sound like, only that i thought hammerfall cards are supposed to have pretty good convertors, so really i have no idea. :confused:

[Scott Collins]

atma,

Thanks for the listen! Right off the bat I'm relieved someone else hears this, too.

When you say, "like a kind of harsh digital distortion of some kind, almost like a bit-decimator" I think we're on the same page. That is a good description of the way I hear it, I suppose. Your point about this possibly being the sound card would make sense, and my guess would be that it is a problem with my particular card (rather than the model).

[kenjkelly]

Hi Folks,
Did anyone ever find out why this vocal grainyness at 5K-7.5Khz is happening? I have exactly the same problem and have gone through the same steps.

HERE IS A DUMP incase we have suspect equipment or software in common:

I tried 4 different microphones besides my AT-4050 (I tried an SM57, SM58, C02 small condenser and a DBX RTA mic). I tried Emu 1820m mic-in, DBX576 pre, Yamaha 02R96 mic-in.

I sometime run my vocals through an auto-corrector (internal VST or external AVP-1), compress a little(1.2:1) to even out the vocals. My deep voice combined with proximity effect requires that I roll off quite a bit of low end. Then the highs start to show the same broken GRAINYNESS you discribed. If I sweep an EQ between 5K and 7.5K the grainyness area can be pinpointed in that range. I tried recording at 88.2K but that did not help.

I have a PETSAFE wireless dogfence transmitter which I discovered transmits at 18.7Khz (and is easily picked up by guitars and other unsheilded inputs), but I don't think that is related. (If you have a PETSAFE - let me know as it could be...)

Scott: Do you use vocal correction and could that be causing this(it happens with it)? Could it be the room (I tried close micing (2") to test/avoid that, but it didn't help).

I though it was the multiband compressor, but compressing is probably just making the grainyness more noticible. I noticed that squarewaves passed through the multiband VST compressor, when NO compression was happening, were seriouly distorted. Other forums said that was normal(Phase due to X-overs) and after passing squarewaves through my DBX Quantum II I confirmed it happens there too. I got a sample multiband from WAVES and their linear-phase multiband does not have the distortion problem BUT I can't try it on my production system because I am running Windows XP x64 (WAVES doesn't support x64 yet.). If it was that kind of problem more folks should be seeing it.

I am using Roalnd RSM-90 Monitor speakers and have had the problem with Cubase, Traction and Audition.

It's driving me nuts, trying to get a silky smooth vocal sound,

Ken Kelly

[Scott Collins]

In my case, I actually wound up attributing a lot of this to my vocal technique (or lack of it). The C1/TB1 came into play because they are considerably more responsive north of 5K than the old SM58 I was using (not knocking the '58, just an observation).

Once I zeroed in on the area I started becoming sensitive to this area when I'd listen to myself sing (i.e. not while recording). I started trying different things to help and, finally, I've begun to get something that appears to work pretty well. I'm no vocal pro, so I can only do my best to describe:

It seems like the problem was that I was pushing my voice too hard and the harshness in that area was resulting. So what I do now is 1) don't worry about volume - ONLY be concerned with tone (even in rock - the volume of the voice isn't so important in my opinion - it's the vibe) and 2) hard to describe but I work on generating the sound internally rather than trying to project it. It's like I'm letting the voice resonate more - almost visualizing it in the back of the head/throat area.

I'm going to be trying this new approach out in the next few weeks with some new recordings, but when practicing I can already hear an improvement.

ALL that said - I think processing was/is a factor, too. One thing I'm suspicious of is the 1176LN compressor (UAD1 card). I think this is a great compressor, but I have read about its "high end distortion" and I think if it is pushed too hard with the wrong voice it might increase the problem. (I wonder if the 1176 concern I've read about is simply due to its super-fast attack/release times?) I still plan to use it, but far more gently and I'll hit the LA2A harder as it is likely to have a smoothing/calming effect due to the way it works.

Also - I've purchased the SP mic cable (and pop filter). I haven't tried it yet, but I was thinking it could help mitigate any genuine distortion that still might be present.

In your case, if you've already eliminated the main problem I had, then I'd suggest you look at 1) compression settings (you might want to skip the MB for a few tests and try a standard comp) and, specifically, attack/release times, 2) your cables/signal path prior to recording. (It sounds like you've already eliminated a number of other factors.) To your other question, I don't think room was an issue for me, as I record fairly close to the mic.

Let me know how it goes!

[kenjkelly]

Thanks!,
I have a new singer who will be doing some of my songs. She's starting after May 11, so I'll know then if it was my voice or not (assuming she doesn't have that same problem).

I'll let you know,
Ken