Natural Hiss Vs Digital Silence in Narration track

[PositiveFuture]

In my narration track I have three distinct sounds:

1. The narrator speaking
2. The analogue hiss of the track when he's pausing for a few seconds (sounds pleasant enough)
3. Total digital silence where the analogue hiss drops off because I inserted a gap.

The Problem

The difference between the analogue hiss and digital silence is BIG, and jars when you listen to it. You really notice the drop off into digital silence.

Two choices:

1) Change the noise floor so there's less hiss, and the difference between the analogue hiss and digital silence is not so great. (Maybe I'll lose some of the sound character like this.)

2) Insert some analogue silence in the digital gaps, so there's a smoother more continuous hiss background. (I can do this but it's fiddly and'll take a couple of days)

What do the pros do????

Thx if you answer this.

 

[Gonzo_Entertainment]

You'd lay a "room tone" under the digital silence. In this case you could either cut out and duplicate the analog hiss from a spot where he stops talking, or just record silence with the same recorder you used to record him, and lay that under it.

 

[PositiveFuture]

Cool. So that's what I figured out this evening, I just didn't know the lingo. So that's a standard practice, and you figure that's better than losing quality by over-effecting the signal??? Got it.

Looks like I've got a heavy day of cutting and splicing tomorrow... Darn.

Thanks for the speedy response.

 

[Gonzo_Entertainment]

The practice by audio guys is ANYWHERE you are recording sound, even voice overs, you record a few minutes of "silence" that captures the natural sound of the room you can then lay under any ADR, foley, etc... that you record in post. It's called a "room tone".

 

[PositiveFuture]

Yeah read about room tone a few times. I didn't realise you'd need it for standard narration but now I see why. Lucky there's a 30 second pause for me to get what I need.

Some room tone links for anyone interested:

http://www.videomaker.com/learn/production/audio/sound-design/room-tone/

http://en.wikipedia.org/wiki/Presence_(sound_recording)

Room Tone
(other terms are Presence and Atmosphere)

A location´s "aural fingerprint" - nonspecific sounds on the upper end (somewhere between 2 000 and 8 000 Hz)

Each room has a distinct presence of subtle sounds created by the movement of air particles in a particular volume. A microphone placed in two different empty rooms will produce different room tone for each.


* Room tone is recorded during 'production sound recording'

* Room tone is used to match the production sound track so that it may be intercut with the track and provide a continuous-sounding background.

* Room tone may smooth out edit points and give a feeling of life in a sound-deadened studio. The soundtrack "going dead" would be perceived by the audience not as silence but as a failure of the sound system.

 

[Alcove Audio]

Good one, Gonzo!

When you edit your VO cut right at the beginnings and ends of the lines. Use the various fades to make it smooth. Then use a low-pass filter to reduce the hiss (however, don't kill the high end so much that you "dull" the voice); there is a possibility that you may not need a room tone. If you do need a room tone create an additional continuous track. Also keep in mind that if there is music or other "noise" underneath you may not need the room tone.

Hiss is an "unnatural" occurrence; it is created by the electronics used to capture sound. All audio capture devices - microphones, pre-amps, mixers, recorders, A/D-D/A converters and the like - have some degree of self-noise. As you get up in quality/price the self-noise becomes less apparent.

A $2k microphone plugged into a $4k mic pre plugged into a $3k A/D-D/A converter recorded in a sonically isolated, acoustically treated room is remarkably quiet. Add in the fact that experienced talent knows how to work the mic, and the engineer knows proper gain-staging and recording technique, and afterwards the VO is properly edited by an experienced editor and mixed by an experienced rerecording engineer. At that level you don't have to worry about room tones; there is no hiss that would be noticed by the average listener.

 

[IndieBudget]

Ok here's a Q.. with basic dialog post work, with decent beginning sound, is the order of cleanup / sweetening
1.Low pass
2 Compressor
3 EQ
4 Normalize

THEN add room tone , foley etc ??










.

 

[2001 Productions]

Then use a low-pass filter to reduce the hiss (however, don't kill the high end so much that you "dull" the voice)

In editing dialog for my feature I've been filtering everything below 200Hz and above 7900Hz. In experimenting with various frequencies, that seemed to get rid of 90% of noise picked up by my Senny M66 without too much voice dulling. Does that range sound about right to you, Alcove?

 

[gpforet001]

I don't compress dialog, nor do I normalize. I EQ when necessary, but it's typically automation based so I only apply it where needed to get a phrase or part of a phrase to cut thru.

I work all my dialog first, checkerboarding for each character, cutting out dead air, then look for anything that needs surgery (i.e. pops, clicks, snaps, cars, airplanes, floor creaks, clothing noise, etc.). Once each persons dialog is clean and on an individual track, I run the track and automate levels and eq for a fairly steady and consistant dialog for that character, I print that track and that becomes my baseline dialog track for that character. Once each character's dialog track is done I move to foley and again foley for each character making the foley track(s) for each character. Next comes practical fx (doors, windows, car engines, phone rings, knocks, bumps, basically everything that wasn't covered by foley). For things like gunshots, explosions, car sounds, and other featured sounds, I'll do these separate with several layers. I then work on room tone layers for each scene, which typically consists of a base room tone with other ambient sounds to set the location (birds outside, muted lawnmover in background when inside suburban home, muted street sounds for indoor scenes in the city, office machine noises, beeps, whirls, restaurant chatter, silverwear, ice clinking in glasses) you get the idea. Lastly, I'll add synthetic tones, drones, and other cinematic sound components but this is usually done in collaboration with the person scoring, so we don't step on each other's toes.

Things like compressing and sweetening I tend to do last and apply it to the the master mix.

Now, I'm sure the big boys have a far more elaborate and logical approach then this, but this is the process I'm currently using on a 12 minute short and it seems to be working well.

Since I primarily focus on films, my post work is never just basic dialog but all of the supporting elements which make up the sonic landscape of the scene.

Ok here's a Q.. with basic dialog post work, with decent beginning sound, is the order of cleanup / sweetening
1.Low pass
2 Compressor
3 EQ
4 Normalize

THEN add room tone , foley etc ??










.

 

[Alcove Audio]

Every project is going to be different, and it will depend upon the tools available to you.

After the production sound/dialog editing is complete I do Noise Reduction. NR is, in some ways, low/high pass and notch filters, but the better NR programs actually run a frequency analysis that prorates the noise; nowhere near perfect but better than simple EQ. Some also have dedicated hiss and rumble filters that work somewhat better than basic EQs. It is not unusual to have several different NR and EQ programs running simultaneously. It is also a case of "less is more" where, if you are printing the reduced noise clips, you do just a little with each pass; you get much less aliasing and other artifacts while getting a bit more NR without overly affecting the dialog. As I work in a Native environment I print the individual dialog tracks after completing the NR process.

EQ is next in line, first on individual channels to get all of the disparate clips matching each other - which can be a major PITA - and then a final EQ on the DX stem.

I tend to not use any compression, or, if I do, a very light touch. I will put a limiter on the DX stem, but it gets hit hard only three or four times during an entire feature.

I usually do not normalize at all, unless I have unusually quiet production sound clips, and the gain of those clips are increased during the editing and NR process.


The dialog processing chain was the subject of a quite lengthy thread on a post production audio forum. Some of the participants have Oscars, Emmys and/or other awards on their shelves. Each one had her/his own individual DX chain and personal favorite plug-ins. There is no "right" answer except "what sounds good." Just keep in mind that the speakers and the room (not to mention expertise and experience) also play a large part of the process; what sounds good on a laptop or computer speakers doesn't sound as good on $150k worth of speakers at a Dolby or THX certified facility.

 

[2001 Productions]

Just keep in mind that the speakers and the room (not to mention expertise and experience) also play a large part of the process; what sounds good on a laptop or computer speakers doesn't sound as good on $150k worth of speakers at a Dolby or THX certified facility.

I'm switching between a halfway decent pair of Senny headphones and the built-in speakers on my monitor. Nice if I could afford a good studio (if there was even one available in my area) for the mix, but I'm hoping that between the ideal (headphones) and worst case (TV speakers) I'll find a happy medium.

 

[Alcove Audio]

GP and I work in pretty much the same way. DX comes first, BGs/ambiences, Foley, sound FX, score/music spotting and then mix. I do the BGs second to see how much noise I can hide in it; that reduces the amount of NR I have to do. I'm not lazy, it's one more step that degrades the quality of the dialog that can be avoided. Other than I'm the same as GP and as all the other one-man-band sound designers.

When you have a budget the work gets divvied up amongst specialists - dialog editors, ADR recordists & editors, Foley walker/recordist teams, Foley editors, field sound FX recordists, sound FX editors and the rerecording team. There is also the usual tech support staff, archivists and interns, as well as the all important Supervising Sound Editor and the Sound Designer, who may not be the same person, and who may or may not be a part of the rerecording team.

 

[gpforet]

Gee, Alcove, maybe the reason we work the same is that I have taken to heart nearly everything you've written here and elsewhere. You and a few others here have and continue to be my mentors. Oh, and I've got a new short I just started and I'm going to try the BGs second to reduce N/R. Thanks for affirming I'm at least on the right page.

GP and I work in pretty much the same way. DX comes first, BGs/ambiences, Foley, sound FX, score/music spotting and then mix. I do the BGs second to see how much noise I can hide in it; that reduces the amount of NR I have to do. I'm not lazy, it's one more step that degrades the quality of the dialog that can be avoided. Other than I'm the same as GP and as all the other one-man-band sound designers.

When you have a budget the work gets divvied up amongst specialists - dialog editors, ADR recordists & editors, Foley walker/recordist teams, Foley editors, field sound FX recordists, sound FX editors and the rerecording team. There is also the usual tech support staff, archivists and interns, as well as the all important Supervising Sound Editor and the Sound Designer, who may not be the same person, and who may or may not be a part of the rerecording team.

 

[PositiveFuture]

Thanks everyone for your input. It's good to know that my basic questions are helping other people out, and a few people are benefitting from the discussion. First: Anyone know any good 'warming' VST's released recently. Yep they are all fake, and nothing like the real thing but you'd be surprised.

This morning I thought through the process logically, made a 15 second room tone track, and stuck it in the digital gaps using audacity. Faded it in and out. Basic, micro budget, but listenable. It didn't take as long as I thought it would.

When you edit your VO cut right at the beginnings and ends of the lines.

Makes sense. I have been sloppy on this.

In editing dialog for my feature I've been filtering everything below 200Hz and above 7900Hz.

Thanks, I didn't know what an acceptable spectrum was. Do you cut it RIGHT out or roll it off smoothly?

The room tone sounds fine to me. Adds an atmosphere to the track.

I work all my dialog first, checkerboarding for each character, cutting out dead air, then look for anything that needs surgery (i.e. pops, clicks, snaps, cars, airplanes, floor creaks, clothing noise, etc.). Once each persons dialog is clean and on an individual track, I run the track and automate levels and eq for a fairly steady and consistant dialog for that character, I print that track and that becomes my baseline dialog track for that character. Once each character's dialog track is done I move to foley and again foley for each character making the foley track(s) for each character. Next comes practical fx (doors, windows, car engines, phone rings, knocks, bumps, basically everything that wasn't covered by foley). For things like gunshots, explosions, car sounds, and other featured sounds, I'll do these separate with several layers. I then work on room tone layers for each scene, which typically consists of a base room tone with other ambient sounds to set the location (birds outside, muted lawnmover in background when inside suburban home, muted street sounds for indoor scenes in the city, office machine noises, beeps, whirls, restaurant chatter, silverwear, ice clinking in glasses) you get the idea. Lastly, I'll add synthetic tones, drones, and other cinematic sound components but this is usually done in collaboration with the person scoring, so we don't step on each other's toes.

And here was me complaining about having to tidy up a single narration track. Respect... Takes days, right?

 

[Alcove Audio]

Anyone know any good 'warming' VST's released recently. Yep they are all fake, and nothing like the real thing but you'd be surprised.

The really big dogs mix on analog/digital hybrid consoles (you know, the $250K+ Euphonix and similar boards). Right below them the mix is run through extremely high end compressor/limiter/pre-amps (Manley, Drawmer, Avalon, etc.) on close to zero settings just to add a little tube warmth. When it comes to plug-ins that mimic tube and tape warmth you get endless differing opinions, allegiances and hates; you pays yo money and you takes ur choice.

A far as high and low pass, as I have said many times, start at zero affect (low pass - 20kHz; Hi pass - 80Hz) and slowly increase/decrease the frequency level (as the case may be). Hit the bypass to compare the original to the effected version. It's quite often a compromise between reduced noise and dulling or thinning the dialog too much.

And here was me complaining about having to tidy up a single narration track. Respect... Takes days, right?

To do proper audio post takes between one and ten hours per linear minute per each audio post category - dialog, Foley, sound FX, score/music spotting, mix. That's five to fifty hours per linear minute; so a 100 minute feature could take 500 to 5,000 man-hours of work - sometimes more. I spent a little over 600 hours on the audio post of my most recent 110 minute feature project. I would have spent more time but that was all the budget there was. It was a basic drama, so not a lot in the way of "exciting" sound FX - one "mean" car, a single gunshot and a big fire were about it. The dialog was pretty good for the most part, so the bulk of my time was spent on Foley. I spent almost three weeks mixing.

 

[PositiveFuture]

endless differing opinions, allegiances and hates;

I'm sure there are massive debates. If you know any off the top of your head please say... Better hear it here than spend hours researching as I am on a tight but extendable deadline.

Thanks for telling me how long it takes to do all the background sound work. If it takes a pro that long I don't feel so bad.

I understand now about starting way out on the EQ then bringing it in. Makes sense.

 

[Alcove Audio]

I'm sure there are massive debates. If you know any off the top of your head please say...

I never use them myself. Most of them are Pro Tools TDM/HD plug-ins, meaning that you must use very expensive Pro Tools hardware. Of the ones left I don't know of any that are VST compliant; they all seem to be RTAS or AU. Here are a couple of possibilities:

Softube CL1B
URS Classic Console Strip Pro Native

 

[2001 Productions]

I didn't know what an acceptable spectrum was. Do you cut it RIGHT out or roll it off smoothly?

I hope that's an acceptable spectrum. It sounds good to me and, as I said, really quiets the track.

I'm using SoundTrack Pro in Final Cut Studio, so my capabilities are nowhere near as sophisticated as what Alcove uses. It has high-cut and low-cut filter tools. They default to 7000Hz and 390Hz, respectively, but those numbers seemed a tad extreme, so I created user presets at 7900Hz and 200Hz, which still does the job without significantly damaging my tracks. I'm guessing the filters just cut those frequencies off, but I'm not sure.