Audio Test Feedback and advice required

T

trueindie

Active member
*Edit: I meant 'requested' not 'required' in the subject ;)

Hey guys,


Can you take a listen to these two files in this link and let me know which one sounds better, or if they both sound horrible? I'm doing this as a test to show someone to get some audio/video podcast work, so any advice and help will be truly appreciated.


http://torontopicturecompany.com/temp/
Audiotest1 doesn't start until like 10 seconds.


Also if any of you guys have some advice on how to best treat sound for the web (youtube/vimeo), that would be greatly appreciated.


Best,
Aveek
 
Last edited:
Both versions are well within the range of generally acceptable for your typical informal blog or podcast.

Certainly #1 is crisper and more "present" as it has a higher dialog/ambient ratio.
But it seems to have some slight clipping or distortion on peaks.
Sounds like it was close-miked like an earset mic or something(?)

#2 has significantly more "room tone". It would sound a bit "off-mic" for a formal production.
It sounds more distantly-miked like a lav or a table-stand mic(?)

The way to avoid the hollow room sound is to either use close-mike techniques (earset, etc.) or to acoustically treat the surrounding surfaces (walls, ceiling, floor) to be less reflective at voice frequencies.
Acoustical treatment varies widely depending on resources (time, budget, expertise) and expectations (permanent, temporary, etc.)
 
Comparitively, the first sounds better. Realistically, even the first has some issues.

The room in which you recorded is reflective and has a ring in the midrange that is distracting. For the second version, which sounds just like a processed version of the first, the compression (and EQ?) make it sound crunchy, thin, and harsh. The gate takes out the ambient noise floor between lines, which really just ends up spunding like a poor dialog editing job.

Podcasts have no standards. Have you listened to The Nerdist? He has a single, cheap mic that keeps getting passed around the room (to the point that handling noise is very audible) so someone is often speaking off-axis from the mic. On the flip-side, there's Serial which has a polished, produced sound (brought to you by, and to the standards of, NPR's This American Life). So, for audio podcasts, there's a very low bar. This isn't a reason to half-ass it, but really more motivation to hit the Serial end of the spectrum so that your work stands out.

Video podcasts have the same "all over the map" spread of quality, but it's the ones that maintain a professional quality that stand out. Honestly, treat it like you would for a paying, broadcast client. Same for YouTube and Vimeo. Make it look and sound the best you can, and don't lower your standards just because it's for the Internet.

Back to your samples, can you describe the room, mic, and recording setup? The space should be treated, or abandoned for a better space that is fairly dead. Some processing (compression and EQ) may be fine, even necessary, but used much more carefully. The processed version sounds like all the processors were pushed much too far.
 
Last edited:
Guys,
Thanks for the replies. I just put up a third file audiotest3, since Richard's comments. Please take a listen to that and let me know if you find that satisfactory.

1. So it was recorded in my living room studio in the manner shown below with an Avantone CK1. Microphone is above his head in front of him.

Tim1.png


2. So audiotest1 was rendered in early January before I had begun to play around with plugins and knew what de-essing was. So after I got the chance to talk to this person who may be looking for podcasts for partners in his firm, I decided to give it another go. So I ran the audio through a compressor, limiter, eq, de-esser and some reverb.

3. After Richard's comment, I basically posted audiotest3, and got rid of the reverb. To me it sounds crisper than audiotest2, and it's also de-essed, which audiotest1 is not.

Goddamn... I'm like ... oh my lord, do I have to go back to the original file and do all the work all over again to make it sound good, or will one of these do as a test???

- These guys want to do little videos of their partners and put it up on vimeo. It's a decent, but small corporation, but they deal with well to do clients, so the audio work can't be just any run of the mill podcast. I'm hoping audiotest3 is decent. Please have a listen and let me know your thoughts.

Best,
Aveek
 
Wow Alex. I thought the gate was doing that, getting rid of the noise floor, and the noise was rushing back when he spoke. But I thought "What the hell do I know? I'm just an audio tadpole. And this other guy was talking on youtube about how fantastic gates are. So maybe I should put it in."

Damnit. Alright. Definitely getting rid of the gate. And you're correct. Audiotest2 is just a processed audiotest1.
Let me know kindly if I answered your questions. The room is of course far from ideal, but it's the only one I've got at present.
And yes, the idea of easing up on the processing sounds like a great idea. Thank you sir.
 
I would draw about same conclusion as Richard.
Clip-2 sounds more compressed. I also hear some distortion in both, but not a serious issue and typical of podcasts.
If it's going to be encoded to MP3, AAC or AC-3, leave a few dB headroom (from full-scale) for the encoding process.

Otherwise, spoken word PCM files <.wav & .aif>, normally need not be dual-mono. A single-channel mono file will have the same quality and playback attributes at half the file size (all other factors being equal).
Note: This quality/file size convention does NOT apply to data compressed audio files.
 
So I assume you were using the Avantone CK1 with the hyper-cardioid capsule?

So this is a comparison of different post-production processing options from the same raw track?

I grew up listening to radio in the 1950s-60s, a bit after the "golden-age", but where they still maintained "pristine" audio quality.
So even now, any hint of un-motivated "room tone" just screams "sloppy amateur-hour" when I hear it.
But I have managed to suppress my autonomous cringe response here in the age of YouTube where anybody with a video camera can (and does) post videos.

Perhaps only a rhetorical question, but what is the motivation for hiding the mic off-camera for a video blog?
Especially one ABOUT production? Unless the topic is microphone selection and placement?
 
I listened to the third sample, and the gate is still distracting. You have, however, confirmed that's what I am hearing and I think taking it out of the mix will be helpful. Ditching the reverb was also helpful... as highly-compressed as the second file is, it was hard to discern the presence of reverb as it all just sounded like a squashed dynamic range.

As for your room and setup:

That shot has tons of headroom, which tells me the mic was a lot farther away than it could have been. As Richard pointed out, moving the mic closer will decrease the amount of background noise relative to dialog. Tighten the shot and get the mic within 8"-12".

For the boxy sound and that ringing midrange, you might start out by putting something absorbant in the corners. Also, the large, flat surface of drywall behind him may be part of the problem. Hanging some sound blankets on stands out of frame may help a little, but since you are in a living room you really want to break up the flat surfaces wherever you can, and make sure there is some absorption around the room as well.
 
trueindie said:
"And this other guy was talking on youtube about how fantastic gates are. So maybe I should put it in."
Click to expand...

Just out of curiosity, can you link to that? I'd love to see just who was advising this, and in what context.

The Internet is a free-wheeling place with absolutely no editoral control or discretion, and as such there are lots of tutorials out there full of bad advice.
 
trueindie said:
...this other guy was talking on youtube about how fantastic gates are. So maybe I should put it in."
Click to expand...
Gates are fantastic as a repair tool or a special-effects tool. But nothing beats actually getting a good recording to begin with.
As Alex mentioned, the interweb is full of information of all types, ranging from immutable gospel truth, to laughable rubbish.
A growing number of "demonstrations" and "reviews" and even "tutorials" are made by people who don't have a clue what they are talking about.
 
I never had much use for noise gates, especially for dialog. A good quality expander can be useful and less noticeable, though they're much more difficult for a newb to tune and the default presets will rarely be right-on.
My favorite plug-in type expander is the Ultrafunk. I don't know if that's even available these days from Cakewalk or the original Sonitus.
 
Last edited:
Re: Gate tutorial and youtube link of offender

If my recollections are correct, I am the offender. I think that dude was just showing how to get low drum noises out with the gate, and I extrapolated that idea into dialog editing, because I thought at that time, that I was a genius. In general, audio guys are usually precise with their advice. Most know what they're talking about. It was really me, muddling up the advice.

Re: Gate
I just changed the gate threshold significantly to let in the noise. Around the 1:22 mark, the streetcar noise begins, and you can hear it continue in the gap.

Re: Mic position
Yes, going forward, I'll bring the mic in closer. Does the mic in the shot matter? probably not, but I was trying out the white background look. You can check out the video with the original audiotest1 sound here:
https://www.youtube.com/watch?v=jRdIZuPkdhU

What I'm now trying to do with the corrected audio is post this thing up on vimeo to show my prospective client, the new video with the new audio.

Rick,
Yes the audio will be compressed to an h.264 file. I was pretty sure I left 2DB headroom. Should I leave more?

You guys are just fantastic by the way. Unbelievable.
Thank you gentlemen,
Aveek
 
Last edited:
trueindie said:
...if any of you guys have some advice on how to best treat sound for the web (youtube/vimeo), that would be greatly appreciated.
Click to expand...

If you want the best sound quality you can get, it's not about "how to treat sound". Not about processing at all. It's mostly just about proper mic placement. Choose the right mic, put it in the right place, and do this in a decent room (not an empty concrete cube). Do this and nearly all of your troubles are solved before you ever get to post. As it should be.

The best post processing is as little as you can possibly get away with.
 
Bruce Watson said:
If you want the best sound quality you can get, it's not about "how to treat sound". Not about processing at all. It's mostly just about proper mic placement. Choose the right mic, put it in the right place, and do this in a decent room (not an empty concrete cube). Do this and nearly all of your troubles are solved before you ever get to post. As it should be.

The best post processing is as little as you can possibly get away with.
Click to expand...

My question was more in the vein of "What should the dynamic range be?" "How much headroom do I leave at the top?" "Should the sound be wet or dry?" that kind of thing.

Unfortunately, all I have available to me is a concrete box and a cheap capsule hyper-cardioid, as my semi-decent audio-technica got lost by the last guy I lent it to. Even the mic stand gets iffy when my subject moves around too much. Filming should also probably be done on film. But all I've got are my DSLRs.

Until we can afford real audio gear/personnel/rooms, some of us have to operate in the world as we live in it, right? I was just trying to learn how to make the best of my situation. :)
Cheers,
Aveek
 
trueindie said:
Unfortunately, all I have available to me is a concrete box and a cheap capsule hyper-cardioid...

Until we can afford real audio gear/personnel/rooms, some of us have to operate in the world as we live in it, right? I was just trying to learn how to make the best of my situation. :)
Click to expand...

I don't think what Bruce was saying requires much outside the scope of what you have at your disposal, at least to get from where you are to a place that is quite workable.

The room, as I said, can be treated. You don't have to drop a couple grand on acoustic foam treatments, but it would help to break the room up a bit with furniture and sound blankets. 90° corners are nothing more than bass amplifiers, so putting something in each corner to break that up will also help. And since you are trying for the infinite/seamless white background look, breaking up the three other walls as best you can will help.

If the mic stand gets dicey when the subject moves, you need to look at balancing and isolating it.

- Is it a tripod mic stand? Make sure one of the three legs is parallel to the boom arm.
- Is it a cast-iron base with a small footprint? Or is it a tripod that you have rigged with one leg parallel and still having issues? Go to the sporting goods store and pick up a couple of wrist/ankle workout weights to use like sandbags with the stand.
- Is there movement transmitting through the stand into the mic? This is the one place that may cost a little more to fix, but get a good isolation mount (Rycote INV-6). And you can also use a rubber kitchen mat under the stand base to help isolate it from the floor.
- Add a short (1.5') cable, looped for strain relief, between the mic and the mai mic cable to further decouple noise from traveling up into the mic.

I think we've addressed the wet/dry issue. Adding reverb just sounds weird and out of place.

Leave yourself 3-6dB of headroom in the mix. Final dynamic range, for the Internet, really depends on personal taste. You can squash it a little using compression and limiting, but if you don't know what you're doing with those plug-ins you can do more harm than good. If you don't want to do that, then don't. Afterall, the less (obvious) processing, the better.
 
Of the lot #3 sounds the best to me. I agree about the gate, it's really not very useful for voice unless you are trying for an effect. As a general thing you don't want to be over -3 dB however I have had projects where the producer insisted on pushing it to 0.1 dB (which is really not a good choice). It's not so bad on the web because the distortion it may cause is going to be right in there with the part of the sound most trashed by heavy compression so I'm honestly not too sure it makes much difference in that situation.

I was thinking you were doing an audio podcast, in which case the mic is never in the shot and you would want it a lot closer. For video without a boom op you might get away with a lock down IF you stay on your mark and don't move, other wise I would use a lav. If you don't move a lot a wired lav will work and not cost an arm and a leg. If you have to move more a wireless is almost essential. Being close though one of the new WiFi sets may work really well and they are pretty affordable.

There is only so much you can do with recorded sound. Cleaning it up is subtractive so the better the original the better the finished version. You can usually make things sound "Better" but horrible to fantastic only happens in fictional wolds. Serious "fixing" also takes some tools that cost more than a better mic setup. The better version of Rx (the cheapest quality noise reduction suite around) costs more than a decent wireless kit, even the cheapest version costs about the same as the WiFi wireless set ups.

So "fixing it in Production" is faster, better, AND cheaper that "Fixing it in post".
 
I hear you guys. Yes, post is always a bad idea, even in video, and I did plunk some cash down on RX advanced version. I got it into my head that I need it for my feature. I still think that I do.


No the mic stand is fine. I was just talking about how the sound changes when the subject moves around, and I can hear it in my headphones.

I have Sennheiser lavs, but I was afraid of using the lavs, as I didn't want to take the chance of introducing scratches when the subject moved around. I thought the Avantone would be the safer bet.

I re-uploaded audiotest4, in case you guys still want to take a listen. I had uploaded the wrong version before (same as version 3). It's the de-gated version this time. I'd love to know if you think it's any better than v3.

I appreciate all the input. If I get the gig, I'll send you guys e-cards or maybe even postcards :)
Cheers,
Aveek
 
"the audio will be compressed to an h.264 file. I was pretty sure I left 2DB headroom. Should I leave more?"
- MP4/H.264 video typically would have AAC audio. Three dB below FS is usually adequate. The lower the resolution encode, the need for more headroom. (a downward spiral) 320kbs usually sounds pretty decent. I would check for clipping after encoding.


"I have Sennheiser lavs"
Which one(s)? The ME series that came bungled with the G2/3 wireless are almost 'usable' for walk and talk stand-ups.
- I have never used an Avantone. Most folks prefer a LDC for narration type work... but if that's what you got, that's what you got. I'm sure most webcasts are done with much worse than that.
 
Rick R said:
- I have never used an Avantone. Most folks prefer a LDC for narration type work... but if that's what you got, that's what you got. I'm sure most webcasts are done with much worse than that.
Click to expand...

Rick, I think we're really discussing matters related to talking heads on camera, as evidenced by the video link from the original audio sample psted in the OP's comments HERE. Yes, a large-diaphragm condenser might be warranted for narration, but not for on-cam.

The Sennheiser systems may be a better bet, and it doesn't really matter in videos like this if the lav can be seen or not.
 
I have a couple of pairs of G3s. If I do get the gig, if they do give it to me and not some big shop as it seems they'd like to, I'd still have to shoot it at their offices, and there they'll talk about their company and what they do for their clients. Even if I'm hired, the situation will not be ideal audio settings. I'm sure there'll be printers running and phones ringing. I'll probably try to use both the lav and a microphone. If I actually get it, I'll borrow the microphone.

Anyway, I don't think I'll get it. This is the first time I've applied for video work, where they specifically mentioned a requirement for good audio. I'll do my best. I'm not planning to be an expert by tomorrow. I'm pretty sure the gig is going to go to a professional outfit.

Cheers.
 
You must log in or register to reply here.
Back
Top