Subject:Routine steps for processing audio

Posted by: claudebou

Date:8/18/2004 6:57:08 AM

I'm a new comer with this soft.
The main use for Sound Forge will be for recording voice over for personnal video using my NLE system.
Could anyone provide some advise on what can be the routine steps for processing a recorded voice over and witch Fx should be consider to provide clean, warm and good amplitude for a V.O.
As an exemple after recording is your first step is "Normalize" the track or you do something else.
I have very little knowlodge in sound editing so any advice will be welcome.

PS: I use a Sound Blaster Audigy 2ZS and Audio Tecnica microphone.

Claude bou

Subject:RE: Routine steps for processing audio

Reply by: Geoff_Wood

Date:8/18/2004 3:41:34 PM

Hi Claude ,

Sound Forge is not the right tool for recording voice-overs. It is a stereo editor. It does not record a new file while playing back another (audio or video) at all, let alone in synchronisation.

Vegas would be more appropriate. But you already have an NLE system - can this not record voice-overs ?

As for recording steps, you have the existing material on the timeline, record in the voice-over, and tweak to perfect the desired sound. This may include EQ , pop + sibilance treatment (better addressed at recording time) and compression.

Normalisation should NOT be the first process, if necessary at all. Any processing that gives any sort of boost may cause clipping to an already normalised file.

SoundBlasters , even Audigy, are not very good quality, even considering price. Also which AT mic are you using ?

geoff

Subject:RE: Routine steps for processing audio

Reply by: claudebou

Date:8/19/2004 6:19:25 AM

Geoff,
Thank you very much for all the advices.
My editing system is Premiere Pro, wich provide a voice over feature.
I must have express myself incorrectly(My english writing is not very good)
I usualy write my text and record after than place it in my NLE.
My main goals is to have the best recording voice possible for the NLE timeline.
You wrote "This may include EQ , pop + sibilance treatment (better addressed at recording time) and compression."
These are the kind of steps that I'm looking for. I did'nt know were to start from.
Can you provide some good sound card name to buy?
As for the mic it's the ATM31.

Claude

Subject:RE: Routine steps for processing audio

Reply by: Chienworks

Date:8/19/2004 7:11:05 AM

The biggest problem with your request is that we don't know what your recordings sound like, so we don't know what they need to have done to them. The best we can do is say that sometimes you will need to use EQ to correct the frequency balance, but we don't know how badly your recordings are colorized, so we can't tell you which EQ settings to use. There are often some pops and silibance, but fixing them depends on what they are like in your recording. Compression is often good; it is also often bad if applied too heavily or when it is not needed. We don't know how much or what type of compression is appropriate for your recordings. We would like to be helpful, but we can only tell you some very general things. Any specific help depends on exactly what your recordings sound like, and would probably vary from one recording to the next.

Probably one of the best things you can do is to experiment with the various filters and settings and see what they sound like to you. Don't be afraid to fiddle with the knobs. You can always Undo and start over again. When you get something that sounds good then you've found what needs to be done.

Subject:RE: Routine steps for processing audio

Reply by: claudebou

Date:8/19/2004 7:25:13 AM

Your absolutely wright Chienworks.
I'm looking for generals step to follow and I must experiment with the fine tuning.
Being advise not to normalise the audio track after recording as a first step or not use that feature at all is very interesting to know.
Thanks for your input.
Claude

Subject:RE: Routine steps for processing audio

Reply by: Chienworks

Date:8/19/2004 7:41:59 AM

I always use normalize as the last step if necessary. If you want to know my order of doing steps, generally i will do noise/pop/hiss reduction first, then EQ, then compression. Some of these steps may or may not be necessary for each individual case. Usually if compression is needed then that takes care of normalization automatically.

The best rule to follow is that less processing is better! Each time you add a process or an effect the sound is degraded slightly. The less you do to it, the less quality you lose.

Subject:RE: Routine steps for processing audio

Reply by: Rednroll

Date:8/19/2004 1:35:08 PM

If you're doing Voice over work, you should try to do the minimum amount of processing as mentioned. This includes normalization. You should look into getting a Mic pre channel strip, something like a DBX 286a is a good choice. This has the eccential tools to record a voice over. Those being a good clean mic-pre, a compressor, a desser, and EQ. The compressor you would start with something like a 4:1 compression ratio with an attack of 5mS and a release around 300mS. Raise the level of the Mic Pre until your gain reduction meter on your compressor is reading around -3dB to -6dB. If the voice over has loud annoying "S" sounds, then you would add your Deesser into the signal path. Ajust the frequency setting of the deesser so that it only triggers on the sibilance parts. This is usually around 5Khz-8Khz for male VO, and 6Khz to 9Khz for female VO. Once you find the frequency you need to adjust the threshold on the deesser so it only triggers on the "S" sounds of the VO, therefore nothing else gets turned down besides the sibilance. Next adjust the EQ, so that the VO sounds natural. I actually recommend not applying any EQ during the recording stage, because it is best to do your EQing when everything is assembled and therefore only adding EQ if it needs to be adjusted to blend in with the rest of the audio. Adding EQ during the record is a bad idea, because what may sound good when recording, may not sound so good when playing it with other audio elements that they are being mixed with. Therefore, you may have to make double EQ adjustments on the VO, and like previously mentioned you want to keep the processing to a minimum and only use it IF NEEDED. Not adding EQ during the record also gives you a better chance to match up previously recorded material with a record you may have to do at a later date. It's not always easy to remember exactly what EQ settings you set for different VO's, where these are easily recallable in software. The final adjustment of the channel strip is your output adjustment. Set this so that the level on your meters in Sound Forge are peaking around -10dB to -6dB, therefore leaving yourself with at least 6dB of headroom before the signal will get distorted. This means the audio can go up to twice as loud before you hit the digital distortion ceiling, yet giving you maximum bit resolution.

When it comes to mix down time, add your EQ adjustments if needed. You may also find a little reverb may help the VO to sound fuller. Try adding a small Plate reverb or a small room.

Subject:RE: Routine steps for processing audio

Reply by: mhbstevens

Date:8/20/2004 12:03:19 AM

In a Vegas thread SoundForge was said to be a good tool formastering. So I ask where would you mix? Can all documentary mixing be done satisfactorilt in Vegas?

Subject:RE: Routine steps for processing audio

Reply by: Rednroll

Date:8/20/2004 6:28:02 AM

Vegas is the best tool for doing Voice over work, mixing and editing the audio. Vegas is a true multitrack program and has the necessary editing features for voice over editing work. Sound Forge is a better mastering tool. Sound Forge can perform more specialized disecting and analyzing processes to an audio file than Vegas.

Subject:RE: Routine steps for processing audio

Reply by: Geoff_Wood

Date:8/20/2004 4:23:09 PM

More than that, It is pretty much impossible to mix in SF with any degree of accuracy wrt timing, unless you are a masochist and like jotting down times and cross-referencing them.