How would you tweak the audio track in this clip?
 

Cut3

The music is from SonciFire Pro, and the voice was recorded with a shotgun (Azden SGM-1000) just off-camera (an HVX-200). (I'm sure using a shotgun contributed to the slight - but certainly detectable - echo).

Am particularly interested in tweaking the vocal.
While I have Adobe Audition as part of the CS2 suite, I know little about tweaking audio. And, while I could experiment all day with settings and effects, time doesn't allow me to learn through trial & error.

The speaker's voice is of a generally limited range, with what I would describe as a bassy quality, i.e. doesn't hit highs often, if at all. The client likes his delivery, i.e. somewhat serious and slow - in order to be understood, and while he sounds naturally "muffled" in this piece, I have heard other work done by him but recorded by a real audio engineer (someone who does only that for a living) who I know used some compression and/or limiting, and maybe even some sort of gate and maybe other filters/effects. (I've listened to an early demo of his from a pro studio in NYC, and it's so much smoother). Unfortunately, I don't have much of a grasp of those tools and am appealing here for help.

It may also help to know that at my age (something I've yet to accurately divulge in this forum), my ears and hearing (among other things) aren't what they used to be.

Any advice on how to improve it, or re: specific effects or settings to use, would be greatly appreciated.

Thanks.

Hi Denis,

Without hearing any samples it's almost impossible to give advice about improving this, similarly to giving advice to improving a video when you haven't seen it.

Can you post a mp3 or wav sample?

If, as you describe, you are dealing with a "muffled" track, then this is like a wrongly mic'd situation - lavlier, radio mic somwhere other than where it should have been. As you say the shot gun.

So, we are dealing with a limited range in terms of frequency (you say it was bassy) and perhaps not limited dynamic range. So Expanding or Gating is not likely to do the job here as this by your description is not a level related issue but one of a limited freq spectrum. Except the echo, a little soft gating can probably eliminate this.

First, I would say this is damage limitation. You cannot restore what is not there! Not by magic, potions, rack gear or what not. But you can probably improve on what you have.

This were me? I would try boosting some selected bands of mid to hi frequency eq. First using a broad Q (wide frequency bandwidth) then try and "tune" this with a narrower (smaller) Q. I don't have Audition (I use Soundbooth) but I'm pretty sure you should have that in your Audition package.

Also keep in mind attenuating (reducing or filtering) the lower bands can achieve the same thing, but without introducing system noise as invariably boosting the top end will. Though this will be significantly less due to the audio being recorded in the digital domain - so you have a fighting chance here - perhaps a combination of lower frequency filtering and higher frequency boosting.

It will probably never sound like it should, but it could turn an issue into your client calling you again :)

Regards to Mid Town: I worked for a number of major broadcast audio equipment manufactures around the world a decade back and was frequently down W49th.

Claire B.

A link to see/hear the clip is
 

supposed to be showing in the first line of my posting. Click on the "Cut 3".

Here is the same link:
Cut3

Please let me know if it's not working.
(Also, thanks for taking the time to offer the benefit of your experience. Seems lots of folks have viewed my posting, but I've had no feedback at all about it.

Hi Denis.

Sorry, should have followed your link.

From what I hear:

This guy is in a box. There is a lot of low energy coming back at you.

There is nothing wrong with the spectrum as I hear on my compatible monitoring. Keep it simple:

Roll out the bottom end at about 150Hz at about 12db per octave (this is the slope of the attentuation). In other words, just bring down the energy level at the lower end.

Then apply a little gating. This is a trigger point whereby you have - and you have not. There are parameters relating to this called Threshold - it can also be frequency dependent triggered.

Go for the simple things first like LF filtering (or roll-off). The simple things are often the best.

But all round, you should be able to hog-tie this guy (well his sound that is).

What I hear is room mode - it sounds like he's boxed in and his voice resonance is actuating a natural resonance in the room.

Additional: treat this as two issues: frequency related, and as the short term "echo" bouncing back. The eq tweeking should take care of the frequency issues, and the gating should take care of the "bounce".

Claire B.

Wow, you're good.
 

You certainly picked up on the kind of environment/set-up where this was recorded. (Can you also tell me what color the room was? ;)

By noticing there is "...low energy coming back at me...", I think that's what I meant - and maybe described poorly - about the "discernible echo", something I know is a threat when using a shotgun in an enclosed space. (Would a lav have eliminated that)?

Just a couple more questions, if you're willing:
1) there are a few discernible lip-smacks and breathing noises. Is it usually "best" to simply select those and mute/cut/remove them?
2) there is also noticeable sibilance. When De-essing, should that be done to an entire track, or only where - and for - the brief moments that the sibilance occurs?

I'll start experimenting with your advice later today. And yes, Audition does have all the tools you mention, and I've read lots of the help pages, but my difficulty is in determining the frequencies involved.

THANKS so much.

I think you're stuck. As Claire mentioned, what you're hearing is room slap, the indirect reflected sound from the walls, ceiling and floor arriving at the mic slightly delayed from the direct sound. Part of the fix is the get the mic away from the camera and closer to the subject. Even shotguns need to be within about 2 feet of the speaker. The mix of direct and reflected sound and the poor behaviour of shotgun mics towards reflected sound in general results in that hollow "ringy" sound. What you hear is the end result of a combination of room reverb and comb filtering within the mic. Sad to say, there's not much you can do about it after the fact except to apply the reshoot filter, preferrably using a hypercardioid boomed just out of frame or a lav on the talent rather than a shotgun mic. Equalization and compression can't fix those issues.

Hi again Denis.

In response to the mic'ing

The shot gun is hypercardoid (on speed), though I am not familar with the make you quote, but the directional element should do it's job providing it's kept tight.

Solution for the future:

If you use this artist again, then try a tighter mic'ing technique, rolling off at about 120hz (12 or 18db per octave). Alternatively, try getting the artist to "constrain" his delivery - not easy and often defeats the purpose. He is projecting (as he should), thereby pumping out energy which in turn causes a resonance in the room. So it is the room + the artist + the mic technique = what you have.

I have monitored this using Genelecs and on a pair of Sennheiser HD201s. These low budget but excellent headphones are very good at hearing the issue.

Good luck Denis.

Agree this Steve, and thanks for coming in on this - but Denis could use selective gating (not compression) as mentioned. But it is thought for the future - get that mic tighter.

Claire B.

Claire, while you are correct about the shotgun's pattern, the issue is that the mic is a line gradient/interference tube design and they are very poorly behaved where there is a high proportion of reflected sound versus direct sound hitting the mic. "Hypers" such as the Schoeps MK41 do not use the interference tube principle to get their pattern and so do not suffer from the comb filtering effects within the interference tube that classical shotguns experience indoors. Interference tube mic patterns are very uneven with lobes of higher sensitivity in the rear-side positions, precisely the directions from which reflections are most likely to arrive, and also become almost omni at the lower frequencies, precisely the frequencies that are giving Denis fits. And finally, Azden is not a brand with, shall we say, the most sterling of reputations.

Gating might help, as you suggest, but I doubt it could fix it completely. The slap is sitting on top of the desired sound, offset by a few cycles, and I think it's going to be almost impossible to cut it without damaging the desired sound. Echo, where there is clear separation between the desired and the undesired sounds, would be another matter but I think these are too tightly intermingled for it to help much. Certainly worth a try though.

Thanks, Claire and Steve.
 

(Boy, this is great getting feedback; it helps me learn both: parts of the jargon of the art, and the art of audio-in-post itself).
I appreciate the feedback and the lesson.

FWIW, the mic was just out of frame and wasn't much farther than 2"...certainly less than 4"...but considering even 3" = a 50% increase from 2"...that makes a big difference.

I know I shouldn't even think this - yet alone do it, but at this point, the client hasn't mentioned anything about the audio...in fact initial feedback indicates they like it, but waiting for final word; it's more me being picky and knowing it could be better. Getting the guy in front of the camera again soon would be not be easy, and would delay the job, (low-budget as it already is). I'll give Claire's advice a whirl and see what I end up with. BTW, and out of a strong curiosity, what's your best quesstimate of what a job like this is worth, i.e. starting from polishing a script, recording with an unknown subject, keying/editing, etc?
Big THANKS to both of you. You're both obviously experienced at this.

Denis:

It sounds fine for the nature of the work. Sure it could be better, but in this case it just dosent matter.

It looks like you ran out of images as the last one seems to be on-screen a lot longer than the others. I hope you have usage rights for the images or in this case I think you will get burned.

PS: They have near zero chance of attracting VC based on that drivel. My day job is managing a portion of a major ARRA 09 solar project.

No lesson given Denis - just sharing the knowledge, sharing the experience.

As you say, the client has not mentioned the audio - in fact I wouldn't have made a big issue of it myself, though it depends upon the context - working for say, ooo.... MTV (ask Marc Repp) SNL (Bob Paladino) or Bob Dixon out at NBC Olympics - I mention these guys as they are in your area (apart from Marc), they would probably pick this out as second nature...

But in the context of the package, is it that critical? In news, documentaries and such programming this would not be an issue. So, learn to ease back on the peddle and see it as the overall effect. If you can improve it then do so, if not, then try to learn from the previous gig and move on - simples!

Claire B.

How much time is it taking you and what do you see your time as being worth? Add in a prorated allowance for your hardware acquisition and replacement. Figure in an estimate of the economic benefit to the client of having the video, his potential return on investment?

Claire, your name rings a bell. Do you post on some other forae?

@Steve,

Glad we could all help here. Dropped you an email.

Claire B.