When I select a segment on my MP3 as the noise profile and ask to get profile, the noise removal window closes. Then when I want to apply the noise removal by going to effect-noise removal, I see that the noise profile has is not stored in Step two. Basically the noise removal option only allows me to manually select noise features in step 2(db etc.) and I don’t see the actual information of the captured nose profile by audacity. Do you know how can I fix this problem?
I follow the two step noise removal, When I select the noise section and click on get profile, the noise removal window closes. Then when I want to apply the noise removal, I go to effect again and click on noise removal, there is no information saved on the section two (which supposed to be the specific db etc of the noise profile taken by Audacity). Do you know what is wrong. Basically the noise removal option does not work in my Audacity (latest version)
I deleted your duplicate post. Please only post the same topic once.
Please read the pink panel at the top of the page and give us the missing information (all three numbers for the Audacity version).
Audacity supplied by us ( http://audacityteam.org/download/windows ) does not show the dB level of the noise profile. The controls in step 2 are for you to use to apply the level of noise reduction, sensitivity and so on that you want - the values start off at default values then Audacity remembers the values you last entered.
When you see that the OK button in Noise Removal is no longer greyed out, that indicates that Audacity has captured the noise profile.
I have Windows 7 on my computer. The version that I have is the latest you provide, which is Audacity 2.0.6
I follow the two step noise removal as suggested in the manual but I think the noise profile of background sample segment is no being captured.
Thanks. Yes the OK button is active but the four measures in step two (Noise reduction, Sensitivity, frequency smoothing and attack/delay time) are always set at the same lowest setting number (0, -20, 0, 0). I thought when we capture the noise profile in step one, these four measurement show show what is captured in step 1 and then we should be able to choose how much noise to remove.
My recording is too long to attach it here as a file. Can you please let me know how can I copy a small segment of it and save it as a different file and send it to you. I do not want to lose my recordings.
Thanks. So since I do not know anything about the four features in step 2 to change them manually, I would only need to click OK on step two to remove the noise captured in the noise profile. Am I right?
The problem is that my recording has lots of sounds that are not clear at all. I am trying to get rid of the background noise in order to understand the conversation.
I have attached a few seconds sample of the noise.
Noise Removal only works on well-behaved noise. It does OK with microphone hum or computer fan noise in your “studio,” but it can’t do much for traffic noises, jets going over, conversations, TV set or other noises that constantly change.
We may be able to help a little with the constant background hiss, but you’re stuck with those pieces of wood somebody is throwing around.
Another restriction is drag-selecting the profile from an area that’s Noise Only. If you get some of the voice in there by accident, Noise Removal will try to remove the voice, too. It’s a pretty stupid (simple) tool.
You can start with:
Attack 0.15 sec.
Keep doing UNDO and bumping up the reduction until you get closer. If you never get there, then it’s probable you’ll never get there.
It’s not a forensics tool. We can’t drag clear conversations out of the mud.
Probably the easiest solution would be to re-record and ask the person to speak more clearly and closer to the microphone.
Please note that we do not offer an audio forensics service for uncovering hidden audio or recovering seriously damaged recordings. For that type of job, look for an “audio forensics” service: http://tinyurl.com/p85tf3t
My recordings have nothing to do with forensics. These are recordings of scientific speeches in different auditoriums that often give low quality recording, specially when people sit far from the recorder. I have to transcribe and make pdf files available from the talks and the discussions. I have most troubles at the end of the sessions (Q&A), when the person who is asking the question does not have a microphone, his/her voice looks like a whispered voice. I can not transcribe the questions when the quality is so low. I was hoping to be able to understand their questions by Audacity.
Unfortunately, if the quality is that bad then there is probably not much that can be done in Audacity to recover it.
CSI frequently gives the impression that a single voice can be pulled out from a sea of noise and made comprehensible, but that is fiction. The “Telephone” preset in the Equalization effect can sometimes help a little, but without specialist equipment and a great deal of knowledge and experience regarding voice analysis, there’s not much more can be done.
If you have access to the lecturer before the show you might suggest they repeat the question before answering it. If the recording didn’t get the question, then chances are good the rest of the hall didn’t get it, either.