Hello all,
 
I am using Sonar X3, a Fireface UC and a M-Audio Sputnik Mic.
When recording vocals that have a lots of volume variations (Whispering > loud singing > Whispering > normal singing for example) i have to manualy change the input gain on the audio interface all the time in order to avoid either " clipping " or low signal.
 
I'd like to know if there are any solutions (beside managing the distance between the singer and the mix) to get rid of those annoyances ?
 
I was thinking about using a hardware limiter between the mic and the card but i was told those hardware weren't fast enough to handle the signal variations accurately.
 
Any idea or hardware suggestions ?
 
Thanks in advance.

Ideally, you do adjust the interface's gain as needed, and avoid combinations of whispers and shouts in the same take. But if you are both singer and engineer, it's not possible to pay adequate attention to both tasks at the same time, in which case a hardware compressor/limiter can help immensely. Their speed is not a limitation.
 
Where you get into trouble is depending entirely on a compressor/limiter to take care of it all. You'll end up with accidentally over-compressed takes, or record too quietly because you assume it's all automatic now. Even with the compressor on you still have to remain cognizant of the interface gain and adjust it as needed.
 
I read about an interesting technique that was used to record a David Bowie tune in which he wanted to sing a wildly-dynamic vocal track in one take. They set up two microphones, one close, the other far, and recorded both of them simultaneously. In the mix, they crossfaded between them when transitioning from loud to quiet and back. I thought that was a clever solution, even if not practical for everyone (you need a very nice-sounding room to record far back from the microphone).
 

Thanks for your clever answer bitflipper ! ;)
 
I do 100% agree with you, as you said i have to be both the singer and engineer at the same time, which is why a piece of hardware between the mic and the audio interface would help a lot... :)
 
Although it's indeed a clever idea, because of the problem explained above, i won't be able to use two mics at the same time.
 
Any ideas about any hardware able to do the job in a nice and transparent way around $300-500 ?
 
Just saw those :
 
Golden Age Project Comp-54 MkII

 
FMR Audio PBC-6A

 
 
 
 

Separate takes is the most obvious solution, but to  add to bit's suggestions:
 
the first thing is the singer.  You can control a lot of dynamics at the source, simply by moving your head back and closer to the mic.  Yea, this can change the tone a little but is better than scorching the take.
 
If the room will allow, and you have a mic/preamp combo that can handle it, back the singer off.  I usually start out at about 18 inches normally, but one can do farther away.  Remember, distance is the original compressor.  Because sound energy is geometric, doubling the distance between mic and source drops the energy by 1/4. 
 
This is one of those things that should be noted when people wonder why so much is spent on equipment - hundreds (or thousands) on a preamp.  Where you can get a good compromise on a problem take using a lower-cost, integrated interface it will probably take more time and work.  A really nice preamp will have the gain to back the singer off without losing clarity or definition, it will grab the lower volume sounds yet not distort on the louder ones but go into a nice transformer saturation whereas the IC interface distorts or the converters give you digital hash.  Forget all the talk of "tone" etc., it is just that good tools will save you as an engineer.  I was recording one of the better local fiddle players.  She has a really nice violin and while talking shop she admitted it had saved her ... neck many times.  A good talent and tool work together.  Good tools make your job easier and less frustrating and let you survive your mistakes with reputation intact.
 
Long story short, try backing off the mic as far as the room and equipment will allow.

Thanks for your answer.

If you have an Omni pattern mic, or your mic can be switched from cardioid to Omni, then you should try that.  It will allow you to back off the mic without loosing tone (but you will add a little room "presence" as you back away).  A standard cardioid pattern adds more bass the closer you get to it, which sounds real nice for intimate close, quite vocals, but means your voice sounds thin if you back away from it.

Hi J,
 
To add to the great advice you got here, I'd go with a input signal of -12 to -6 and forget touching the input gain. As long as you're at 24.bit you have plenty of room to manipulate your sound after you record. Meaning, compress after you print the track qnd then just automate the parts you want louder or lower or heck, though I try to stay away from this, you can normalize low signals if need be. Years ago that method would never be mentioned by me, but in today's times, the only time you may get noise artifacts from this is if you record with an input signal that is super weak.
 
Personally, I'm a fan of recording individual tracks for individual situations. This way you get the parts exactly the way they should be and they will be closer to the vision you may have in your head. I do use compression going to disc but never in excess. So I don't think that will be as helpful as one may think. The reason being, it can mess up your dynamics if your not careful and there is no undo....you have to re-track.
 
In situations like yours, in all the years I've been doing this, the best fix for me was a happy medium on the input signal, compress in Sonar after, write in your automation and call it a day. Anything like whispers or what I would call "special" I just record on a separate track and process it the exact same way..... Compress and then automate to taste. Good luck!
 
-Danny

Danny beat me to it... With 24 bit you have a lot of leeway to play with in mixing as long as the signal is far enough away from the noise floor.

this:
 

Thanks all for you clever advices and feedbacks.
I agree with you all, i'd better have to manage distances and signal gain rather than putting some hardware i won't be able to " undo " after.
 
Thanks again.