Help fixing underwater "tin room" sound

[davidlieb]

Hi Folks, I'm hoping you can help me.

One of my instructors recorded a full e-Learning class on a Blue Yeti. Later, we found that the microphone was having issues. The problem is she put a couple of days of work into it.

Here's a WAV snippet

It has a very underwater "tin room" sound to it. I'm wondering if anyone can offer advice to make it sound... well, passable. It doesn't have to be perfect.

I have a little bit of experience with sound programs, but am still a beginner, so please take it easy on me.

Any help would be super appreciated.

[ryclark]

I don't think that you will have much luck with improving this. The audio is too degraded and variable to be able to do anything to it I'm afraid.

[SteveG_AudioMasters_]

I have to say that this sounds as though somebody's already had a go at it - a mic fault wouldn't sound like that at all. This sounds like processing artefacts from Noise Reduction. Do you have a clean copy of the original?

[davidlieb]

It does sound processed. But, It's not. That is the original recording.

The gain setting on the back of the Mic was acting wacky, and then the mic stopped working all together, so I assumed it was the mic. I guess it could have been the computer or something, but everything else seems to be working. 

Thanks for thinking about this.

[ryclark]

It sounded almost as if it was being recorded from two different audio sources that weren't locked together somehow causing wierd cancellation effects. Or somehow the audio being fed back delayed out of the PC to mix back with the incoming mic signal. Or even some strange audio enhancement effect being applied by the Windows audio driver.

[SteveG_AudioMasters_]

davidlieb wrote:

It does sound processed. But, It's not. That is the original recording.

The gain setting on the back of the Mic was acting wacky, and then the mic stopped working all together, so I assumed it was the mic. I guess it could have been the computer or something, but everything else seems to be working. 

The problem here is that there's some rather indisputable evidence, I'm afraid. If you listen to the speech, and the silences between the words, the room tone clearly audible in the speech parts disappears in the silences. So at the very least, some dynamic expansion has taken place, and microphones simply aren't capable of doing this unaided. So something is processing the signal from that mic before recording it, and it isn't Audition - because Audition simply isn't capable of recording anything other than dry signals.

And the Blue Yeti doesn't do any signal processing except to amplify and digitise the three capsules, and output them as a 48k 16-bit stereo stream. The recording we have is dual mono...

So can you please explain exactly how this miraculous recording was made?

[davidlieb]

Huh. Thank you all for your expertise. This is proving very enlightening.

The audio wasn't originally recorded into Audition. I was just hoping to use Audition to fix it. It was originally recorded into Camtasia, which is a video screen capture program for eLearning. It does have a noise reduction feature, but my team swears they haven't turned it on, and looking at the file in Camtasia I can see that it's not on. It is possible, I guess, that noise reduction was applied and saved in some weird way.

I'm also trying to track down the laptop the recording was made on (we have several we check out when needed) to see if maybe there's a sound card issue.

So, I don't know about miraculous, but it sure is mysterious!

I am hearing correctly that you guys feel like there is no helping this file, though, right?

[SteveG_AudioMasters_]

davidlieb wrote:

The audio wasn't originally recorded into Audition. I was just hoping to use Audition to fix it. It was originally recorded into Camtasia, which is a video screen capture program for eLearning. It does have a noise reduction feature, but my team swears they haven't turned it on, and looking at the file in Camtasia I can see that it's not on. It is possible, I guess, that noise reduction was applied and saved in some weird way. 

Hmm... you need to look very carefully at the settings, I'd say. And get people to check the output quality of the recordings they make before proceeding too far with them...

I'm also trying to track down the laptop the recording was made on (we have several we check out when needed) to see if maybe there's a sound card issue. 

Won't be a sound card issue, because the Yeti doesn't use it - it's a USB mic that does its own digitising, and provides a stream that can be recorded directly. There's no other way it can work. And anyway, listening to the actual mic sound reveals that it's not actually a faulty digitisation - it's too clean behind all of that noise for that to be the problem.

I am hearing correctly that you guys feel like there is no helping this file, though, right?

That's certainly the case with the file as presented, and if there really is no unprocessed version then that's what you are stuck with, yes.

If you are just going to do mic recordings, you'd be better served by doing them directly into Audition anyway - it simply can't make that sort of mess of a recording, quite deliberately. This is because it complies with standard industry practice, which is always to record dry whenever possible.

Sorry if this is perhaps a little forensic, but that's the way it is!

[davidlieb]

And get people to check the output quality of the recordings they make before proceeding too far with them...

Yes, that's what people are supposed to do. Guess they didn't this time.

I'd love to record directly into Audition, but the recording needs to match up perfectly with the video screen capture.

Thanks for all your help. I love the forensic aspect! It is really obvious you guys are knowlegable professionals. The amount of time you put in to help people like me on forums should buy you some extra heaven points, for sure.

If any of you guys need to brush up on other types of programs, drop me a line at www.LearnItAnytime.com. Ask for David.  I'll get you hooked up.

[SteveG_AudioMasters_]

davidlieb wrote:

I'd love to record directly into Audition, but the recording needs to match up perfectly with the video screen capture. 

I'd have to look into it, but I'm sure there has to be a way of doing this. In terms of timing errors, they would be completely insignificant, even over a long recording period - it's not phase-aligned music you are recording, after all. I know that you can import content into Camtasia, but how this would apply to audio alone, I'm not sure. I think that there ought to be a way, though. I might download the trial version and have a look - there are one or two tutorials I was thinking of making, and I'd like to know what the actual limitations here are.

[davidlieb]

The problem is that it does have to almost exactly match up to the video.

Instructors point at things and talk through what they're doing, and

whatnot. Would love to know if there is a better option...

On Thu, Feb 20, 2014 at 2:06 AM, SteveG(AudioMasters) <

[ryclark]

Steve was talking about music alignment where accuracy would need to be within a sample or two. You don't even need lip sync for your purpose ie. frame accuracy of 1920 samples. You only need to have the audio within about a 1/4 of a second of the picture ie. 12000 samples.

lCamtasia should be able to produce better results than yours.  But if it won't you should be able to record the audio onto a standalone audio recorder whilst capturing the video with Camtasia and lay the sound back against the video in Camtasia.