Removing hiss/noise (1 Viewer)

[RED(tape)MENACE]

I'm pretty sure Tower do head cleaner tapes - 60 minute casettes are tougher to find.. you'll have to do a bit of searching there. If you use a lot, I guess you could internet them.

 

[ryanego]

get a decent parametric eq, theres probably one in at least one of your programs, set the bandwidth as narrow as it will go, cut it as much as you can and drag it across the frequency spectrum. stop when the hiss goes away and note the frequency its at. You'll probably have to decrease the amount of decibels you cut as it will have a big effect on the sound depending on what frequency the hiss is roughly at, but just try to find a balance and see what works. thats how i do it anyway, hope that helps

 

[Rogue]

get a decent parametric eq, theres probably one in at least one of your programs, set the bandwidth as narrow as it will go, cut it as much as you can and drag it across the frequency spectrum. stop when the hiss goes away and note the frequency its at. You'll probably have to decrease the amount of decibels you cut as it will have a big effect on the sound depending on what frequency the hiss is roughly at, but just try to find a balance and see what works. thats how i do it anyway, hope that helps

"Danger Will Robinson!" "Don't cross the streams!" etc...

This technique should only be used to get rid of discrete frequencies, booming etc - noise, by definition, occurs across the spectrum at varying statistical levels - there's almost no instances where I'd do this willingly or as anything other than as last ditch measure. Hum noise is usually not limited to a single frequency, the harmonics spread across the spectrum.

If you've got enough noise to be noticeable as noise then removing individual bands will always be a bad compromise where you almost always lose more and more signal to get rid of noise. If you can hear the noise the signal to noise (SN) ratio is too much to the noise side, pulling frequencies will reduce both the signal and the noise so you don't actually affect the SN ratio. If you're lucky, the bands you remove are perceptually more noise than signal and you get an improvement. If not, the end result is usually deadening of the track by removing signal brightness.

Just what i've got from the heartbreak of hours of running a parametric eq across a track to no avail. Find a half-decent noise remover plugin and experiment.

Grrr or any other pros, feel free to correct me

 

[kraster]

"Danger Will Robinson!" "Don't cross the streams!" etc...

This technique should only be used to get rid of discrete frequencies, booming etc - noise, by definition, occurs across the spectrum at varying statistical levels - there's almost no instances where I'd do this willingly or as anything other than as last ditch measure. Hum noise is usually not limited to a single frequency, the harmonics spread across the spectrum.

If you've got enough noise to be noticeable as noise then removing individual bands will always be a bad compromise where you almost always lose more and more signal to get rid of noise. If you can hear the noise the signal to noise (SN) ratio is too much to the noise side, pulling frequencies will reduce both the signal and the noise so you don't actually affect the SN ratio. If you're lucky, the bands you remove are perceptually more noise than signal and you get an improvement. If not, the end result is usually deadening of the track by removing signal brightness.

Just what i've got from the heartbreak of hours of running a parametric eq across a track to no avail. Find a half-decent noise remover plugin and experiment.

Grrr or any other pros, feel free to correct me

That's pretty much spot on. As Igor said, most noise is broadband in nature but the noise in the upper part of the frequency is most apparent because that's where the ear is most sensitive ie. 1khz to 5 khz. Pulling out frequencies in that region will lead to a dulling of the material and make it less intelligible. Also a Q that high on an eq will lead to phase shifting.

 

[C24]

I've had to do quite a bit of this - and as someone else mentioned the best I came across was the waves stuff. And, hopefully not entangling myself in semantics (!), reducing noise is the only thing you can do - it's impossible to eliminate without degrading the signal. There are other pc based ones availabl for cheap - but they're all shit.. Also the use of eq when removing noise from 4track cassette stuff ain't so stoopid... There's shag all in the top and bottom of those anyway - so filtering out some high and low probably won't lose you much signal wise - (everything above say 15k as a start and below 60hz with a super steep q).. I guess I'd try some filtering and then some subtle noise reduction on the result.. Noise reduction really is all trial and error - And ALL the harware units ain't worth a damn bar the cedar stuff...

Hope this helps

David