Harshness Begone - Going down for 3rd time

[mike gouthro]

I've posted this elsewhere without much luck. My apologies for the redundancy.
**********************
One last try before accepting that there are no insights to be obtained for my dilemma.
*******************

I've used Rode NT1 mics for three years. They work well for my acoustic guitar but have always resulted in harsh vocal tracks with female vocalists singing fairly loudly at 8" to 24" distance. I auditioned Rode NTV and CADE200 mics to compare. The CAD was no better than the NT1 and The NTV only slightly less harsh. The mics go into a DBX386 and then via s/pdif into a Roland VS1680 (MTP mode).

Over the last three years, my approach was to back the DBX386 input trim (drive) off until no flashing reds occur on the peak LED (-3db according to the DBX manual). Then, adjust the DBX output control so the DBX meter frequently illuminates the –4dB yellow meter light, but never high enough to illuminate the 0dB red meter light. I did this on the assumption that it would achieve maximum signal to noise ratio without overloading or distorting.

I then adjusted the VS1680 digital input fader so the 1680 meter ranged between –12dB and –4dB with regular peaks to -4dB and occasional peaks hitting –3dB or –2dB. I kept the levels below or at –4dB, not out of concern for overloading or distortion on the track, but to preserve headroom when later mixing multiple tracks down to master tracks.

This method has resulted in consistently harsh vocals. Initially, I suspected the Rode mics, then the analog inputs of the VS1680, later the DBX386 and finally (and increasingly irrationally) the VS1680 digital recording technology. I know quality recordings frequently have wonderful sounding loud vocals using similar technology. Several queries on the internet and locally made it clear to me that no one else was hearing the intolerable vocal harshness that threatened to torpedo most of my recorded song projects (which sounded fine in all other respects).

Well, in the last week, I finally achieved significant relief from the harshness. The strong female vocalist is positioned at 9” from the Rode NT1 with a pop screen. I set the DBX output control to the unity gain position. I then adjust the DBX input trim (drive) so the –8dB yellow meter light comes on occasionally. The –4dB yellow meter light never comes on. I set the VS1680 digital input slider to unity gain. The 1680 meter ranges from –16dB to –8 dB with occasional peaks hitting -6dB.The Rode NT1 still sounds a little too bright for my taste and lacks warmth, but the result is definitely usable and bearable.

I can certainly live with the NT1 for a while. It seems very accurate and detailed. But solving the major problem has now given me a taste for something as accurate and detailed but less bright and dry, and a little warmer.

To see if the NT1’s brightness could be reduced, I then recorded two NT1’s, one with the DBX386’s 20 dB pad on and the input “drive” cranked; the other without the 20 dB pad and the input turned down as described earlier. The clear winner was without the 20 dB pad. The pad did nothing to reduce the brightness but added a noticeable and unwanted reinforcement of the fundamental frequencies and some loss of detail.

I’m a happier camper now. But I would value any comments in case I have missed a better approach that would avoid the low meter levels that characterize my solution.

Mike.

[dobro]

The NTK is about twice as good as the NT1 - I've used both. The NT1's got detail, but the NTK's got detail and smoothness and beauty that the NT-1 can't approach. If a vocal sounds harsh through the NTK, then I'd guess it was just a matter of the wrong mic for that voice. Or the wrong vocalist for that mic.

[warlock]

cross-posting is a big no-no.

Keijo

[mike gouthro]

I should have read the rules more closely. It won't happen again. I guess most members tend to read many forums. My original post and a "to the top" yielded no responses for several days in the first forum I tried so I began to think I was targetting the wrong group. I know better know.

Thanks.