VO-BB - 20 YEARS OLD!

Lapianoman · Posted: Mon Feb 27, 2012 2:25 pm Post subject: Noise Floor...right?

So, I used Audacity to analyze (using Plot Spectrum) 10 seconds of recorded silence on my AT2035, resulting in the following:

I then switched to my new RE320 & repeated the process, with this result:

Have I just reduced my noise floor by 8dB? Would this confirm that I now have a noise floor of -80dB, or is there another program I need to use? Is -80dB a decent enough noise floor for VO? I'm trying to achieve an acceptable level of sound quality & this is just another step...

SkinnyJohnny · Posted: Mon Feb 27, 2012 6:15 pm Post subject:

If that's a true picture of what's going on, then that's an AWESOME noise floor! I'm wondering about the frequencies below 1000Hz though. Also, when recording at the level you used, what are your peaks showing?
_________________
John Weeks Voice Overs
www.johnweeksvoiceovers.com

Bish · Posted: Mon Feb 27, 2012 6:36 pm Post subject:

If I got those results... I'd wonder if the mic was plugged in Laugh

As Johnny says, some context might help. With my rig set-up, I'm shooting for a recording level sitting nicely around -12 to -6dB, allowing for peaks scratching the -3dB limit, and shouty-bits dangerously approaching 0dB. With that, I'm doing just a tad better than around -60dB (-63 ~ -60dB) on the noise floor.

I would look at your settings in audacity. What is the Spectrum/Hanning measurement actually showing you? I really doubt that these are meaningful noise floor figures... I'd expect a higher level than you're showing from the mic's self-noise... or even Brownian motion! Wink

Cheers
Peter
_________________
Bish a.k.a. Bish
Smoke me a kipper... I'll be back for breakfast.
I will not feed the trolls... I will not feed the trolls... I will not feed the trolls... I will not feed the trolls.

Lapianoman · Posted: Mon Feb 27, 2012 7:31 pm Post subject:

I decided to fire up Sound Forge & watch the meter. Looks like the levels during silence flutter around -57dB. That sounds more realistic.

I just did a bit of recording for an audition & saw my levels falling about where yours are Bish: Most levels sit around -16 to -10dB for normal conversational lines. When I'm doing growly or "shouty bits", I'm just under -3dB.

Leslie Humble · Posted: Mon Feb 27, 2012 7:57 pm Post subject:

Wow Brownian Motion! I must be at a convention of Mensa Voiceoverists.
Maybe more technically Brownian Noise which is produced by Brownian motion but wow.

BISH WROTE:
"I'm shooting for a recording level sitting nicely around -12 to -6dB, allowing for peaks scratching the -3dB limit, and shouty-bits dangerously approaching 0dB. "

And that sounds perfect to me. Anyone wondering about your levels need only read this. You gotta leave a little head room and a little processing room whether for you or the studio. And though it's been said many times many ways do stay away from the 0db. You are in danger of transgressing and you are not going to suck up that much floor noise if you are around that -12 and normalize.
_________________
There are rules to the Universe. Learn them and prosper. www.HumbleVoiceover.com

cyclometh · Posted: Mon Feb 27, 2012 9:50 pm Post subject:

I can hit a noise floor of about -60db. I do about what Bish said with levels. Seems to work out.
_________________
Corey "Vox Man" Snow
http://voxman.net

georgethetech · Posted: Mon Feb 27, 2012 10:04 pm Post subject:

Put the Axis setting on "logarithmic" for a more clear picture of what's going on below 1000hz. That's typical to have a rumble in a home studio, hence the need to use a HPF in almost all cases. -80db would be as good as the best world-class studios, and better than most. -70 is exceptional, -60 is typical, -50 is OK, -45 and higher is a FAIL when it comes to SaVoa standards.
_________________
If it sounds good, it is good.
George Whittam
GeorgeThe.Tech
424-226-8528
VOBS.TV Co-host
TheProAudioSuite.com Co-host
TriBooth.com Co-founder

Bish · Posted: Mon Feb 27, 2012 10:24 pm Post subject:

Whenever I see that SaVoa -45dB threshold, apart from seeming pretty high, I can't help but think back to the early 70s when I was working on data transmission lines for British Telecom... I could set-up the copper from one end of the country to the other and give you a magnificent (and expensive) 2400 bps*. -45dB was my max noise. Also from the early 70s was this noize Wink

Cheers
Peter

*Yes, I do know the difference between bauds and bps, but I'm keeping it simple
_________________
Bish a.k.a. Bish
Smoke me a kipper... I'll be back for breakfast.
I will not feed the trolls... I will not feed the trolls... I will not feed the trolls... I will not feed the trolls.

cyclometh · Posted: Mon Feb 27, 2012 10:59 pm Post subject:

I feel bad sending something out with a -63db noise floor. I can't fathom -45db, heh.
_________________
Corey "Vox Man" Snow
http://voxman.net

cyclometh · Posted: Mon Feb 27, 2012 11:00 pm Post subject:

And Bish- mad props for knowing the difference between baud rate and bits per second. Wink

_________________
Corey "Vox Man" Snow
http://voxman.net

Jen Gosnell · Posted: Tue Feb 28, 2012 12:50 am Post subject:

Dayo · Cinquecento Joined: 10 Jan 2008 Posts: 544 Location: UK

To be honest I've always been a bit confused how people measure noise-floor. I mean it's all relative - right?
_________________
Colin Day - UK Voiceover
www.thurstonday.co.uk

Bish · Posted: Tue Feb 28, 2012 7:17 am Post subject:

Ignore the following if you don't like anything tech/geek/math Wink

This is a minefield... but it's a minefield that I know my way around reasonably well. I started working on data/line transmission back in '72 and continued working in the field until '04. This was back in the day when all the data was being pushed around on the existing speech network, and these circuits had to be conditioned and set-up to squeeze the maximum data-rate out of them. Most people of a certain age remember their old dial-up modem which was designed to get the best out of the existing speech line... part of my job was to set-up & fine-tune the private lines that would out-perform the standard speech line by an order of magnitude while using the same basic infrastructure.

Colin is absolutely right when he says that the dB is a relative measurement, and should only really be used when referring to a s/n ratio. If we want to talk in absolutes, then we should be using standard references to give us a dBm figure that is an absolute measurement of power. There are obvious variables, but with signal transmission we always used 800Hz into 600ohms as a standard reference. We use decibels because bels are an unwieldy large unit (hence the "10 log")

N=10 log (P/1mw) where N is the power in dBm and P is the measured power.

The whole dB/dBm/dBw/dBu/dB(A)/etc. nomenclature can be a nightmare, and is also very industry & application specific. One industry's standard 800Hz into 600 ohms, pink noise weighted, may be another's pork pie.

In the real world, we tend to take some shortcuts when it's anything overtly technical. We drop the bits that confuse (or are irrelevant), and as long as we are careful, this can simplify the issue. If we assume that we are using a maximum input level of 0dBm (albeit choosing to work around 6dB lower than that), then the dB/dBm become interchangeable for all practical purposes.

The thing for the non-techs to remember (because of the logarithmic nature of d cool

is that -48dB is half the power of -45dB. A 3dB drop will halve the power. But, this does not mean that it will sound twice as quiet...

Comparing a -50dB studio with a -71dB studio ... the noise power level is over two levels of magnitudes less (1/128th), hence the initial skepticism of seeing nothing on the graphs poking up above -90dB.

Leslie... I'm guessing that I should have really referred to the Johnson noise (caused by the molecular Brownian motion in conductors above absolute zero) and Gaussian noise (more generally associated with noise generated across semiconductor junctions) Laugh

I really wish I hadn't started this. It's a painfully gross over-simplification of the issue, and now my head hurts Rolls Eyes

Cheers
Peter
(Exit, pursued by bear)
_________________
Bish a.k.a. Bish
Smoke me a kipper... I'll be back for breakfast.
I will not feed the trolls... I will not feed the trolls... I will not feed the trolls... I will not feed the trolls.

Lapianoman · Posted: Tue Feb 28, 2012 8:00 am Post subject:

I appreciate all the replies! At this point, it's probably best for me to post some unprocessed sample audio & let the experts determine if I'm getting a decent enough sound. Hopefully I can upload something tonight.

As a side note, that Slade video brought me back to the early days of MTV:

http://www.youtube.com/watch?v=bHoPYLQvnQM&feature=related

Dayo · Cinquecento Joined: 10 Jan 2008 Posts: 544 Location: UK

Hi again Peter

Thanks for the in depth explanation.

Mind if I ask a supplemental? If 3dB lower is half the power, how come it's not half the volume? Is that a perception thing?

Colin
_________________
Colin Day - UK Voiceover
www.thurstonday.co.uk