My wav file had 3 big drops in Har-Bal at about 90 Hz, 120 Hz, and 190Hz regions. After moving up those parts I discovered very bad hissing in my strings.
So I reloaded a small part of the wav file containing only strings. I shelved down all frequencies above 200 Hz, moved up the 3 deficient regions, and saved the Eq'd file.
In Acl (Audio cleaning Lab) I created a noise print file with the Eq'd file in the denoiser window and applied the noise print to my wav file and voila ! The hissing part was almost removed at all. Some bass portion was also lost but the deal was great.
If the Eq cursor could work in a more narrowed band Har-Bal would be a terrific tool for such manipulations.
SergeD
A Nice way to cut hissing
SergeD,
I'm happy to hear that you had success with your unusual application.
On the subject of the narrow-bandedness of the parametric EQ tool, it is basically restricted to it's realistic workable range for the low frequency region. It would be possible to realize filters with sharper notches in the HF region but I would need to change the way in which filter files are specified.
At present, a filter is defined by 1/6 octave points on a cubic spline which nicely corresponds to the 1/6 octave spectrum smoothing applied to the spectrum estimation. If you want better resolution than this you'd strictly have to reduce the level of smoothing proportionately from 1/6 octave to soomthing smaller and increase the length of the FIR to match. On the other hand having more resolution is not necessarily a good thing because you then get clouded with too much information. I choose 1/6 octave because that is on the limit of about how small you need to go to have a filter control point control a bandwidth that is entirely within a critical band. Having the ability to adjust with a resolution substantially less than a critical band in width doesn't add much except in very specific cases (ie. supressing single tone noises and the like).
I get the gist of what you're after but as of now I don't think I've got a good solution. I'll need to think about that a bit more.
Thanks,
Paavo.
I'm happy to hear that you had success with your unusual application.
On the subject of the narrow-bandedness of the parametric EQ tool, it is basically restricted to it's realistic workable range for the low frequency region. It would be possible to realize filters with sharper notches in the HF region but I would need to change the way in which filter files are specified.
At present, a filter is defined by 1/6 octave points on a cubic spline which nicely corresponds to the 1/6 octave spectrum smoothing applied to the spectrum estimation. If you want better resolution than this you'd strictly have to reduce the level of smoothing proportionately from 1/6 octave to soomthing smaller and increase the length of the FIR to match. On the other hand having more resolution is not necessarily a good thing because you then get clouded with too much information. I choose 1/6 octave because that is on the limit of about how small you need to go to have a filter control point control a bandwidth that is entirely within a critical band. Having the ability to adjust with a resolution substantially less than a critical band in width doesn't add much except in very specific cases (ie. supressing single tone noises and the like).
I get the gist of what you're after but as of now I don't think I've got a good solution. I'll need to think about that a bit more.
Thanks,
Paavo.
Oups... my topic should better be in the unusual and creative use of Har-bal area.
I understand your point. Perhaps it could be possible to drop specifically the smallest part without sloping other near parts. Something that could look like an hole rather than a slope. Anyway I consider this way of using Har-Bal as a nice bonus.
SergeD
I understand your point. Perhaps it could be possible to drop specifically the smallest part without sloping other near parts. Something that could look like an hole rather than a slope. Anyway I consider this way of using Har-Bal as a nice bonus.
SergeD
SergeD,
Actually, I think you've nailed it on the head. A different spectrum adjustment tool may be the solution to the problem.
Thanks,
Paavo.
PS - If you could email a bit more detail (maybe screenshots) of the type of edit you wanted to acheive (but had some difficulty) it would help in figuring out exactly what a solution should be able to do.
Actually, I think you've nailed it on the head. A different spectrum adjustment tool may be the solution to the problem.
Thanks,
Paavo.
PS - If you could email a bit more detail (maybe screenshots) of the type of edit you wanted to acheive (but had some difficulty) it would help in figuring out exactly what a solution should be able to do.