Subject:Help

Posted by: HeyJude

Date:11/13/2013 10:19:23 AM

I've been narrating for a year now and recently took a contract with a new company. I've had my first work rejected. This is what they said and the solution they proposed. ... I've spent an entire day reading/watching/researching ... trying to figure out how to do what the suggested fix is. I can't figure it out. Can anyone decifer this and tell me please what to do to fix my files.

I usually narrate at 50% volume of my noise reduction mic. I go through looking for pops and things that need the noise gate to erase and I do that. Then I equalize the whole file. then I run the whole chapter through volume and increase it to 197%. This was the rejection I got and how to fix it

"The files have an RMS level (average level) of -41dB. We ask that all audio submitted measures between -23dB and -18dB RMS, with peaks hovering around -3dB. Your noise floor should fall between -60dB and -50dB.

Solution:

I recommend using RMS normalization to -18 dB and limiting your peaks with a limiter, threshold set at -3dB. This should bring your audio to the required range."

When I used the Normalizer on the file... I tried setting it to -18 but there was a lot of white noise laced behind my narration. Is there a next step that I am missing?

what do I do? Please help

Subject:RE: Help

Reply by: rraud

Date:11/13/2013 11:10:00 AM

Don't know what you mean by a "noise reduction mic" and 50% volume".
Furthermore I don't understand "197%" either.
In any case, if needed after your processing [u[peak normalization, not RMS, possibly were the noise is coming from due to it's compression scheme. I recommend using a compressor/peak limiter combo to even-out levels some. The Wave Hammer plug-in works good... if not used to excess. Your workflow of removing pops, noises and EQing is logical and SOP. However If a noise gate is needed to reduce background noise, you have other issues to correct, probably hardware related or your room/recording environment.
BTW, if your final peaks are around -6dBFS or so, that would be sufficient for post... as long as they never exceed 0dBFS (clipped). Many professional clients and broadcasters I encounter still require a calibration tone... you may want to put that at the head or in a separate file. Normally -20dBFS in the USA.

Subject:RE: Help

Reply by: HeyJude

Date:11/13/2013 12:04:11 PM

The microphone is advertised as having a noise reduction function. I have a home constructed sound booth to keep things quiet. I set the microphone to record at 50% volume on the computer. Then after I am finished part of the process is to run the file through the volume function on the process tab to increase the volume 196.79% or to 5.88 dB.

The back noise comes after normalization process is completed. I use the noise gate to take out the breaths that are distracting.

In any event the thing I need to do is apparently this
:"I recommend using RMS normalization to -18 dB and limiting your peaks with a limiter, threshold set at -3dB. This should bring your audio to the required range."

How do I do that please?

Or in your suggestion. What is a wave hammer plug in? Where can I get one?

Subject:RE: Help

Reply by: roblesinge

Date:11/13/2013 1:54:03 PM

What specific make and model of mic are you using? I think you're recording too low, but without hearing a sample, that is just a guess. (BTW, linking to a sample might help people here give you suggestions).

This is what I do with the voiceover I edit every day. We usually record at a decent level to begin with. I would set yourself up to record where you don't peak above -6dB. That should give you a pretty good level to start with.

In post, I usually follow these steps. First I do a Graphic EQ highpass at 120Hz to remove any rumble that snuck onto the recording. This will also help with plosives. Next, find a noiseprint, or silent spot in your recording that you can use as a reference for the noise reduction plugin. This only needs to be a few milliseconds with nothing but the room noise. I check that print, and if the level is above -70dB, I run one pass of the Noise Reduction plugin only removing 12dB of noise. Next, I run some light compression. This may seem aggressive to some people, but I compress at 1.6:1 with the threshold set at the RMS level of the file. After that step, I check my noise print again. If it is above -70dB, I do another 12dB removal pass. Next, I do my final normalization. I go to -18dB RMS, but you might want to go to -20dB to stay well within the guidelines you were given. Now I do one last check of my noise print and run a final -12dB pass if it is still above -70dB. You can obviously substitute your own acceptable noise floor level for my figure of -70dB.

That's just my way and there are probably millions of variations. Short version is, record at a good level. EQ to remove some of the unwanted frequencies. Lightly compress. Normalize to your final level. Noise reduce in short steps after each major processing step.

Rob.

Subject:RE: Help

Reply by: rraud

Date:11/13/2013 2:59:01 PM

If you have SF Pro, you have Wave Hammer. It's a DX plug-in available in the Effects menu or can be added to the 'FX favorites' or in a plug-in chain.
Set-up the first comp. stage as 'roblesinge' suggested.. On the second vol. max. limiter stage, set the output level to your desired absolute peak, (dBFS) and it will not exceed that setting. Gradually pull down the threshold adj. to bring up the volume with minimum 'gain reduction showing on the far right 'gain reduction' meter... BEWARE, excessive gain reduction will induce distortion and there's no one pre-set that fits all, so you will have to tweak the parameters in both stages.
I would advise not to use a hard gate (noise gate) at all. To lower the breaths, it's best to do it manually with a 'Graphic Fade' tool with a smile type envelope.. or Waves makes a plug in designed for it. never tried it though.
Additionally 'roblesinge's 120Hz HP filter may be a little to high and remove the 'balls' from your voice. Again you will want to tweak this as well.
Another standalone 'normalizer' type app you may want to try is the free 'Levelator', easy to operate but not really 'tweak-able'... a good option for folks w/o compressor/limiter skills.
Edit:
http://web.archive.org/web/20130729204551id_/http://www.conversationsnetwork.org/levelator/

Subject:RE: Help

Reply by: HeyJude

Date:11/15/2013 9:50:28 AM

Thanks for the suggestions. I'll work thru them and see if I can make this work. I am figuring out that I have a somewhat more "discounted" version of Sound Forge Audio Studio 10.0 and it doesn't have all the bells and whistles that other versions have. For example in Normalize I don't have the option to normalize to "Average RMS Level". It's a conundrum

My microphone isn't probably the best either. I have a logtech noise reduction headset microphone and I've also tried a plantronics one.

The first company I've narrated for hasn't seemed to have any difficulty with my product. This new company must have a tighter standard and I'm not totally sure I have the ability to meet. I'll give myself a 50/50 ignorance quotient as a handicap.

fuss and bother.

Thank you

Jude

Subject:RE: Help

Reply by: HeyJude

Date:11/15/2013 9:51:36 AM

Thanks. That all sounds very helpful. Unfortunately my version is Sound Forge Audio Studio 10.0...not pro sadly.

Subject:RE: Help

Reply by: rraud

Date:11/15/2013 10:49:29 AM

"I have a Logtech noise reduction headset microphone"
- That is your biggest issue. That is a consumer item designed of phone calls and such.
You really need a good quality narration mic and quality mic preamp. Your computer's sound card is likely a consumer product as well. Typically, a large diaphragm condenser mic is the first choice for VOs. However radio host mics are popular as well. like the EV RE20, 27 or 320; Shure SM7 and Sennheiser 421 to name a few. There are some semi-pro audio USB mics as well which would work, just stay away from gear designed for consumers. If your on a budget, you can get a decent sounding set-up for around $100-200 Even less would be a large step in quality over the POS Logtech noise reduction headset mic.
Aside from that, SF pro has more tools for sure, but the recoding sound quality (or lack thereof) would be the same whatever version of SF or software you have.

FYI: You should also record and edit in a non-data compressed PCM file format, (WAVE or AIFF) with a 44.1kHz or higher sampling rate (48kHz if video is involved) and a 16 or 24 bit depth setting.

Addendum: Ty Ford (a trusted source) states the $50(USD) MCA SP-1 Cardioid Condenser mic sounds very good despite it's extremely low price. You will still want a pro audio preamp/soundcard though.
http://tyfordaudiovideo.blogspot.com/2012/03/mca-sp-1-cardioid-condenser-mic.html

Additionally, the topic VOs is discussed frequently over at http://www.dvinfo.net/forum/all-things-audio

Subject:RE: Help

Reply by: roblesinge

Date:11/15/2013 11:31:47 AM

I agree with rraud. The mic is your first weak point. There are some VO applications where a little quality can be sacrificed. However, better-paying jobs are going to require better quality. Honestly, with a decently quiet and non-reflective space, you can get good results with something as simple as a Blue Yeti USB mic. I've done some VO with mine and have gotten some good results. If you're looking for better results, upgrade your mic first thing.

Second, SF Audio Studio isn't going to have everything, but you can still do most of what we described. If you have to normalize by peak, do so. I would norm to -6dB. If your recording levels are good and you do the compression steps, that should get you to roughly -20dB RMS. I'll have to do some digging, but I think even Audio Studio has basic EQ and Compression (might be called Dynamics).

Rob.

Subject:RE: Help

Reply by: Benny Bear

Date:11/18/2013 5:18:21 PM

I don't do voice over work, but I do use a reflection filter to cut down the room sound when recording vocals. There is less work post recording to get a decent sound. They've come down in price a lot and work well.