EQ'ing for video/film

Jon Fairhurst · August 24th, 2006, 01:32 AM

Hello Patomakarn,

You make a great point about using good noise reduction software. Here are some additional options:

* Arboretum's Ionizer
* Sony Noise Reduction plugin for Sound Forge
* Voxengo Denoiser
* 'cleaning audio lab" from Magix
* WavePurity (from Germany)

And, if you're cheap like me...
* Goldwave (which you can try for free - it has a really retro uiser interface thought.)

On a previous project where there was lots of noise, Goldwave worked really well. On my most recent project the noise was much lower. As things got close to the noise floor, the residual noise was very "rubbery" and processed sounding. No problem! Because the initial recordings were very clean, I only mixed in the noise reduced dialog when the dialog was too faint. The "rubbery" sound was so faint that it was completely masked by the unprocessed track.

Had I used higher quality NR, I probably could have gotten away with just using a processed dialog track.

Another tip: capture some clean anbient noise. You can use this as a "pad" under your noise-reduced dialog to keep things natural. Don't use anything with a hum or buzz. Find some ambient noise that sounds more neutral, but still reflects the sound one might expect to hear in your setting.

The most important thing is that your noise level shouldn't jump around from scene to scene, except as the viewer would expect (e.g. a transition from a plush high-rise office to a busy Manhattan sidewalk.)

Jeff Cottrone · August 30th, 2006, 11:40 AM

Jon,

Thanks so much for your post. The EQ thing is a huge help. I have a few questions about how to accomplish the other stuff. I'm new to Adobe Audition and hope you use that program and apologize for the newbie questions, but I haven't figured it out yet. I have Total Training, but it's more musical based, when I'm concerned with cleaning and sweetening the dialogue.

1. Mixing nominal dialogue to peak at -12. How is this done? If I'm in edit view, do I use the amplify/fade effect and set the initial and final amplification to -12? It seems like this is reducing the volume of even the quiet parts (will applying your noise red. trick bring those parts up?).

2. Using envelopes. In edit view, is this under Amplification/Envelope (process)? So do I just highlight the part that is too loud and apply an envelope? What should the form look like? Can I use a preset?

3. Noise reduction trick. How do I mix the second track (aggressive noise reduction) into the first track? Is there a level I ride so the second track is like muted except when I raise it in the quiet parts I want to increase? How do I do this? In multitrack view?

4. How do I ride the volume levels up and down in real-time? Or do I have to highlight certain areas and apply it separately?

5. Lastly, what order is your workflow. Certain scenes in my movie are louder than other scenes. And within a few scenes I have vocals that are too quiet, which I need to boost. There's also different degrees of background noise in each scene. My goal, of course, is to make the volume level consistent on both the global level and within each scene, and also to reduce the noise (without effecting the vocal quality). Does the workflow order matter?

Before I saw your post, I've been trying to level off the whole track with global compression. Then boosting everything to a hard limit level. Then doing noise reduction in each scene separately. Then trying to mess with the overall EQ. I know this is wrong. Is your post in your workflow order: EQ 1st, then fader to -12, then noise reduction?

Jon Fairhurst · August 30th, 2006, 12:41 PM

Hi Jeff,

I've never used Audition, so I can't help with the specifics. With Vegas, you right click the control area for your track and enable the volume envelope. You can then set an overall level with the fader, and mix phrases and syllables up or down locally with the envelope. You just add some points to the envelope line and drag them up and down to make changes anywhere in time.

Regarding my workflow, I do this:

1) Edit your video cuts.

2) Roughly adjust the volume for each segment.

3) Render the audio at 24 bits.

4) Open your noise reduction application. Listen to the various segments. Copy some "clean noise". Apply it to everything that has the same noise signature. Find a balance between noise reduction and distortion. Save to a new file.

5) Put the rendered audio on one track. Put the noise reduced audio on another. (Note: if the noise is really bad, you might make a heavily noise reduced track and a light noise reduced track.)

6) Set a nominal level on your normal track. Watch your meters. For normal spoken dialog, the meters should peak at 12 dB. Screams and whispers will vary.

7) Set a nominal level on your super-noise reduced track. Set its envelope to zero.

8) Send the output of both audio tracks to an Aux bus or group. Apply your EQ here.

Note: This assumes that your actors can all tolerate the same EQ. That's not always the case. You might cut up the dialog so each actor gets their own tracks (normal and noise reduced) and EQ.

9) Review the dialog scene by scene and word by word. Adjust the envelopes of the two tracks as needed. Allow an acceptable amount of noise from the normal track. Pull down things that are too loud (such as P-pops) with very local adjustments. Push up dialog that is too quiet with the super-noise reduced track. You can be somewhat sloppy here, as you won't hear much of anything between words.

10) Mix all of your music/foley/effects into an aux/group bus and apply your EQ there. Mix it around the dialog as needed. If things get too loud, try pulling down just the bass instruments on local hits. It's like a manual multi-band compressor.

11) When it's all done, review the whole thing and double check that your peaks are 1 to 2 dB below full scale. Adjust as needed.

12) Render out.

There are countless other tricks and effects that you can use, but this is a great starting place for on-screen dialog and music/foley/effects. Basically, you get your dialog right first, along with noise control, levels and EQ. Add the other elements next. Balance all of the elements, then review the overall effect and deal with any remaining peaks.

All the best...

August 24th, 2006, 01:32 AM	#16
Jon Fairhurst Inner Circle Join Date: May 2006 Location: Camas, WA, USA Posts: 5,513	Hello Patomakarn, You make a great point about using good noise reduction software. Here are some additional options: * Arboretum's Ionizer * Sony Noise Reduction plugin for Sound Forge * Voxengo Denoiser * 'cleaning audio lab" from Magix * WavePurity (from Germany) And, if you're cheap like me... * Goldwave (which you can try for free - it has a really retro uiser interface thought.) On a previous project where there was lots of noise, Goldwave worked really well. On my most recent project the noise was much lower. As things got close to the noise floor, the residual noise was very "rubbery" and processed sounding. No problem! Because the initial recordings were very clean, I only mixed in the noise reduced dialog when the dialog was too faint. The "rubbery" sound was so faint that it was completely masked by the unprocessed track. Had I used higher quality NR, I probably could have gotten away with just using a processed dialog track. Another tip: capture some clean anbient noise. You can use this as a "pad" under your noise-reduced dialog to keep things natural. Don't use anything with a hum or buzz. Find some ambient noise that sounds more neutral, but still reflects the sound one might expect to hear in your setting. The most important thing is that your noise level shouldn't jump around from scene to scene, except as the viewer would expect (e.g. a transition from a plush high-rise office to a busy Manhattan sidewalk.) __________________ Jon Fairhurst

August 30th, 2006, 12:41 PM	#18
Jon Fairhurst Inner Circle Join Date: May 2006 Location: Camas, WA, USA Posts: 5,513	Hi Jeff, I've never used Audition, so I can't help with the specifics. With Vegas, you right click the control area for your track and enable the volume envelope. You can then set an overall level with the fader, and mix phrases and syllables up or down locally with the envelope. You just add some points to the envelope line and drag them up and down to make changes anywhere in time. Regarding my workflow, I do this: 1) Edit your video cuts. 2) Roughly adjust the volume for each segment. 3) Render the audio at 24 bits. 4) Open your noise reduction application. Listen to the various segments. Copy some "clean noise". Apply it to everything that has the same noise signature. Find a balance between noise reduction and distortion. Save to a new file. 5) Put the rendered audio on one track. Put the noise reduced audio on another. (Note: if the noise is really bad, you might make a heavily noise reduced track and a light noise reduced track.) 6) Set a nominal level on your normal track. Watch your meters. For normal spoken dialog, the meters should peak at 12 dB. Screams and whispers will vary. 7) Set a nominal level on your super-noise reduced track. Set its envelope to zero. 8) Send the output of both audio tracks to an Aux bus or group. Apply your EQ here. Note: This assumes that your actors can all tolerate the same EQ. That's not always the case. You might cut up the dialog so each actor gets their own tracks (normal and noise reduced) and EQ. 9) Review the dialog scene by scene and word by word. Adjust the envelopes of the two tracks as needed. Allow an acceptable amount of noise from the normal track. Pull down things that are too loud (such as P-pops) with very local adjustments. Push up dialog that is too quiet with the super-noise reduced track. You can be somewhat sloppy here, as you won't hear much of anything between words. 10) Mix all of your music/foley/effects into an aux/group bus and apply your EQ there. Mix it around the dialog as needed. If things get too loud, try pulling down just the bass instruments on local hits. It's like a manual multi-band compressor. 11) When it's all done, review the whole thing and double check that your peaks are 1 to 2 dB below full scale. Adjust as needed. 12) Render out. There are countless other tricks and effects that you can use, but this is a great starting place for on-screen dialog and music/foley/effects. Basically, you get your dialog right first, along with noise control, levels and EQ. Add the other elements next. Balance all of the elements, then review the overall effect and deal with any remaining peaks. All the best... __________________ Jon Fairhurst

August 30th, 2006, 11:40 AM	#17
Jeff Cottrone Regular Crew Join Date: Sep 2005 Location: Rochester, NY Posts: 100	Jon, Thanks so much for your post. The EQ thing is a huge help. I have a few questions about how to accomplish the other stuff. I'm new to Adobe Audition and hope you use that program and apologize for the newbie questions, but I haven't figured it out yet. I have Total Training, but it's more musical based, when I'm concerned with cleaning and sweetening the dialogue. 1. Mixing nominal dialogue to peak at -12. How is this done? If I'm in edit view, do I use the amplify/fade effect and set the initial and final amplification to -12? It seems like this is reducing the volume of even the quiet parts (will applying your noise red. trick bring those parts up?). 2. Using envelopes. In edit view, is this under Amplification/Envelope (process)? So do I just highlight the part that is too loud and apply an envelope? What should the form look like? Can I use a preset? 3. Noise reduction trick. How do I mix the second track (aggressive noise reduction) into the first track? Is there a level I ride so the second track is like muted except when I raise it in the quiet parts I want to increase? How do I do this? In multitrack view? 4. How do I ride the volume levels up and down in real-time? Or do I have to highlight certain areas and apply it separately? 5. Lastly, what order is your workflow. Certain scenes in my movie are louder than other scenes. And within a few scenes I have vocals that are too quiet, which I need to boost. There's also different degrees of background noise in each scene. My goal, of course, is to make the volume level consistent on both the global level and within each scene, and also to reduce the noise (without effecting the vocal quality). Does the workflow order matter? Before I saw your post, I've been trying to level off the whole track with global compression. Then boosting everything to a hard limit level. Then doing noise reduction in each scene separately. Then trying to mess with the overall EQ. I know this is wrong. Is your post in your workflow order: EQ 1st, then fader to -12, then noise reduction?