I edited a microphone recording in Premiere and took it to Audition to remove the microphone hiss.
I managed to overwrite the original MP3, so now I only have the noise reduced version: https://www.dropbox.com/s/3cargfz8j7hi1x6/nonoise.mp3
Does its sound alright? I can still hear the hiss, and it stops each time he stops speaking (aka the places where I put the cuts).
I was gonna re-export the original noisy audio from Premiere, but that project file screwed up with the "This project contained a sequence that could not be opened" error. When I used the Notepad work around, I opened the project and it's now empty!
+ Reply to Thread
Results 1 to 8 of 8
Sounds fine to me. No audible noise and no loss of details.
Greatly depends on what you intend to do with this.
If this were being done for an educational video (I'm guessing by the content), particularly one which will be COMPRESSED using Lossy compression, you are fine and the gaps where it actually goes to full silence will compress nicely (giving more bit quality to the remainder). I would, however, have done a slightly more gradual fade-in/fade-out to silence on each segment, because it is right on the verge of being noticeable as is.
If this were being done for cinema or TV broadcast, this would be bad form and likely noticeable. In those instances, the correct thing to do is to fill in the gaps with appropriate & innocuous roomtone (also with crossfades).
If you don't like the hiss, you can still run a dehiss filter on it. Remember, you now are working with a compromised source (mp3 compressed 2nd generation), so tread lightly & give it somewhat greater bitrate to compensate.
Overwrote you original...
Personally, I'd prefer to have constant hiss rather than hiss that pumps with the speech/music/whatever. But the pumping is bearable in this clip.
Thanks for the tips. It's for a couple of videos, and there'll be music, so hopefully it won't be so noticeable?
Cornucopia, there are no fade ins/outs yet, I was going to add them afterwards, cos doing that pre-noise removal can make it trickier to remove the noise.
Ok, so say I'm stuck with this audio file, what would be the best option?
Cornucopia, are you suggesting fading his speech in and out to avoid the sudden cuts from hiss to silence?
Yes, I would strongly recommend fading in/out for each segment (yes, this will take some effort if you have lots of segments). I would consider it a noticeable difference. Less so if you have music under, but still there.
NOT his speech, btw, but the "air" around it. Of course, if you haven't already given yourself enough space around the actual speech signal , there's not a lot you can do. Then the choice is to try a quicker-than-optimal-but-still-workable fade of the actual speech signal, OR to do what jagabo & I suggested and drop in roomtone segments to fill the gaps.
But, if you don't have roomtone, or can't get a good segment of it from in-between speech bursts, etc., your options are more limited.
Note: it is preferable to insert the roomtone into the open silence rather than just underlay it, because the latter will add to the noise floor when the "air" from inside the speech clips is added to the roomtone's "air". And the "pumping" will still be there. With the former, that is not true.
Back at the last post-production house I worked at, I had a macro I created for ProTools that would go through and add dissolves (? 1/10 second?) between each speech segment and its adjoining roomtone clip (which had previously been gain adjusted to match the level of the "air" underlying the speech). Could do the whole timeline in less than a minute or 2. Make the computer work for YOU.