Getting the volume right without hiss

D

David Evans

Well-known member
Hello,

I've been using the MM1 as my main preamp with great results. I have some questions that always puzzled me though...

1 - What industry standard do you use to set the volume of your movie? Do you use dialogues to measure the desired loudness and set everything else accordingly?

2 - How to push the volume of each recorded sound without introducing hiss? Let's assume I'm talking about digital recordings and not analogue. Is the hiss a direct result of the clip being pushed further than it should? But what if it's still not loud enough?

Thanks!
 
Hmmm.,

After reading this a few times I'm going to make some assumptions.

I'm assuming this is a post audio question, though it starts out like it's a production audio question, so?

I'm also going to answer as if you were mixing a decent budget feature. You are not but you talked about industry standards so...

The first step is to record you production at a good level and get you mic close enough. Your actors should be able to touch the mic if they reach out.

Then you are into post. You should go to the stickies and read the parts on post, it's a lot more than I want to repeat.

Big films are mixed in calibrated rooms by ear. The "standard" is what you calibrate to, after that it's as loud or soft as you and the director want it to be.

TV/Cable etc is a different world and you need to get your mixes to comply with "standards" that vary by network/ station. It's a mess and you just have to find out what "standard" they want you to hit and alter your mix to match.

I would still mix like you were a big budget film and then do whatever it takes to meet individual places requirements.

I'm also going to assume that you are calling basically everything you don't want to hear "hiss". True his in the digital age is either because you have a very noisy mic, really crappy mic pre. mixer, recorder, very low levels or a combo of some of that. Most of the true "hiss" can be eliminated by having decent consumer gear or better and recording at a decent level.

The "other hiss" mostly comes down to getting the mic close enough and listening for noises you don't want like incoming planes.

In "real" films post sound adds a lot of sounds and the only thing you get from production is dialog. Even on a super low budget film done well 90% of what is kept from production is only dialog. So hiss in production recordings can be a big PITA, BUT it is also being buried in added sounds so unless it's really bad you probably don't hear it.

As a very general "rule" properly recorded production is always turned down in post. If you have to turn production up something went wrong. That is not 100%, and sometimes you just can't get close enough and S&^t happens, but if you are having to turn up production regularly something is wrong.

For the low rent filmmaker one of the best things you can do for your sound is to calibrate your system and then LEAVE it there. You can go high tech or you can just take something that is as loud as you want to get, say an explosion, and set you speakers so that it is just about too loud. Now do not change the amp/ speaker/ computer/ whatever/ levels. now as you go through your film adjust clip levels and volume automation and stay away fro the speaker/ amp controls.

This works just about as well as a "real" calibration and anybody can do it. If you have done that and your dialog is still too low you need to change how you record it because something is wrong.

I and just about anyone who mixes films will mix the dialog first. I love SFX but dialog is king, period. So mix your dialog so it sounds the right level, with nothing else playing. Now add you backgrounds and hard FX etc. keeping them at levels that allows the dialog understandable. In maybe one/ two BIG explosion everything is on fire everyone is shouting scenes maybe you cheat up the dialog so the SFX can get to where they need to be but then get it back down to your original levels. You can also duck the SFX under key bits of dialog, even when the world is blowing up (you just need to be slick getting in and out of that).

After the film plays well that way is when you should bring in the music. At this point you may want some scene to be more about music and you will take the SFX as a whole down a bit to let the music be the boss. BUT you always want some SFX under the music because that is what grounds you sound in the STORY'S WORLD rather than switching your film into a music video.

Unless your making a music video music should almost never be the lead player. Dialog is the lead, SFX (I'm lumping Hard FX, Foley, Backgrounds etc in that group) is a supporting player and music is mostly an extra.

Some sequences are usually all about the music (like dream sequences) but music dominated sound is inherently unnatural so IF you slather music on as so many mini- indi's do you pull the audience right out of your stories "World" and into a theatre watching a music video. At the right time and place music in a feature can be transcendent, but that moment was built on SFX and it's only a moment. 99.9% of the time music should be someplace between an extra and a supporting player, and really outside of music videos and experimental films it should never be a lead.

If you have done the above hiss is not a worry.
 
What sort of mic are you feeding the preamp with?
What distance is the mic from the sound source?
What level are the people speaking at?
 
Thanks for your response! Well, to be more specific:

1 - I'm currently using a Audix SCX1 HC for most of my indoor stuff.
2 - I'm feeding it into a MM1 preamp.
3 - MM1 goes to camera XLR input.
4 - I always get the mic as close as possible to my subject.

Now, I've read in a lot of places that digital recording doesn't have to be "hot", like analogue. Which means that we should record dialogues leaving a lot of headroom for post. Is this even true? From Noiz2 response, I'm guessing no.

My question is:

At what volume should my dialogues be, before distributing the movie, and how to avoid unnecessary hiss if it starts to show up when pushing certain clips to that desired volume...

Thanks
 
As I mentioned there is no standard level for what you mixed dialog should be at. Some places have a specification you are required to meet but that varies a lot and is based on the whole sound not the dialog alone.

Digital does not pick up tape hiss which is an issue with analog tape. But there is still electronic noise.

But as I also mentioned most people who talk about "noise" are not actually talking about noise, they are referring to room tone and other sound they don't want.

Anything you record will get louder if you crank the gain, including any noise, room tone, planes over head, etc.

But if you are recording so your dialog peaks are somewhere in the -6 to -3 dB range it is very rare you will need to turn anything up.

Recording low in digital effectively lowers your bit depth. If I remember correctly it's around one bit per three dB. So if most of your dialog is is at -24 dB then your 16 bit recording is a lot closer to a 8 bit recording. I probably have the math wrong but the point is that recording low is not a good idea.

What kind of films are you making? And how are you distributing them?
 
While you have lots of dynamic range in the recorders, the signal to noise is set by the noisiest piece of kit you use. So a noisy receiver or something else can be recorded quite hot - leaving the noisy stuff quite high. The key for me is recording in a way that adds as little extra noise. Some people like to keep more headroom, others less - but what is bad is when you have lots of mixed level tracks in the edit. One fader pulled back because the track is very hot and another pushed because it's low. You have less control on the fader that has little travel, and maybe need a bit more on the one that is full already. Whatever you do, consistency is the vital thing.
 
David Evans said:
Now, I've read in a lot of places that digital recording doesn't have to be "hot", like analogue. Which means that we should record dialogues leaving a lot of headroom for post. Is this even true?
Click to expand...
Recording (or "tracking") the original sources (i.e. dialog) should ALWAYS be made at the most advantageous SNR (Signal to Noise Ratio). You are trying to capture the signal as far out of the "mud" of noise floor as possible without clipping the peaks.

There should be NO CONSIDERATION AT ALL for how the track will be mixed in post-production. It is the job of the recordist to capture the cleanest signal for the production mixer to use down the line. How the signal is mixed in post is of no concern to the production recording on location.


At what volume should my dialogues be, before distributing the movie,
Click to expand...
As others have already observed, that is a post-production mixing decision made in concert with all the other factors, SFX, Foley, music, mood, etc. etc. It has nothing to do with the recording levels during production shooting.

and how to avoid unnecessary hiss if it starts to show up when pushing certain clips to that desired volume...
Click to expand...
The production recordist is responsible for capturing the cleanest dialog during shooting. That goes all the way back to location selection to minimize ambient noise. Unless you have a really low-quality operation, lousy location selection, crappy equipment, inexperienced sound crew, low-priority given to sound by the director, etc, you should not be seeing "hiss starting to show up".
 
The right microphone makes a difference. Mics reputed to be better than long shotguns are noticeably better at interior dialog capture. I have used the AT 4053b and have noticed a better sound. Some directors cannot hear a dufference, it seems, and don't want to change mics.

Hey, it's not my film; I have to tell myself that if I disagree, and made my opinions known.
 
Just as an aside, I picked up an Evidence Lyric HG XLR cable (short) for my setup between my pre-amp and my Roland R26 recorder.
I believe that purchasing this high quality XLR cable cleaned up the "hiss" by about 50% by itself.
 
spinkscapes said:
I believe that purchasing this high quality XLR cable cleaned up the "hiss" by about 50% by itself.
Click to expand...

In terms of ambient audio noise, I've never noticed any difference between a cheap and premium XLR cable. Yet I still buy and use Canare Star Quad cables because I don't want to pick-up RF noise and I want a cable that will remain reliable even after years of use.
 
Noiz2 said:
As I mentioned there is no standard level for what you mixed dialog should be at. Some places have a specification you are required to meet but that varies a lot and is based on the whole sound not the dialog alone.

Digital does not pick up tape hiss which is an issue with analog tape. But there is still electronic noise.

But as I also mentioned most people who talk about "noise" are not actually talking about noise, they are referring to room tone and other sound they don't want.

Anything you record will get louder if you crank the gain, including any noise, room tone, planes over head, etc.

But if you are recording so your dialog peaks are somewhere in the -6 to -3 dB range it is very rare you will need to turn anything up.

Recording low in digital effectively lowers your bit depth. If I remember correctly it's around one bit per three dB. So if most of your dialog is is at -24 dB then your 16 bit recording is a lot closer to a 8 bit recording. I probably have the math wrong but the point is that recording low is not a good idea.

What kind of films are you making? And how are you distributing them?
Click to expand...


Tell me what you think this means then especially the adjusting up part. This is quote from Robert at Juicedlink talking about the dynamic range of the -16db safety track of his recorders/preamps. If I'm recording -6 to-12 and the safety tack is -16db lower then I'm at -22 to -28. Is seems he's saying -16 keeps the DR and adjusting it up decreases the DR . Is he talking about just a one off loud scream as the function of the safety track making the non scream part too low?

"The safety track is not adjustable, and that's how I wanted it. If I let people adjust it further down (-18dB, -20dB), then you start to get into a dangerous area, and I didn't want people who didn't know what they were doing to mess themselves up. Giving people the option to adjust it up (-12dB, -8dB) merely decreases the dynamic range you get from the safety track, for no benefit. This has a juicedLink low-noise preamp in it, so the SNR is going to be limited by the SNR of the lav in any of these adjustment ranges, so there is no SNR benefit to doing this."
 
The safety track would only be used for the audio segment (a scream for instance) that clipped the 'main' track... of course if a radio mic transmitter or preamp clips, the safety track will be the same. just at a lower level. That's why limiters (and competent sound folk) and equally important. An occasional digital clip is easily repairable, not so with preamps an such.
 
People toss terms like dynamic range around all the time with out really knowing what they are talking about.

At it's core DR is the difference between the loudest and softest signal. Generally with recorders they are talking about the usable range so you will set the bottom at where the inherent noise swamps the signal.

The lower you set your level the lower your top is and the bottom is fixed based on your equipment and it's base noise level. So inherently lowering your peak level will lower your possible dynamic range. Practically though clipping is a lot worse than a somewhat squashed dynamic range. So in real world terms you top is wherever you have to set it so that you don't clip.

Now the lower your peak is the lower your effective bit depth is. That isn't dynamic range it's resolution. If the bit depth gets too low you end up with kind of "grainy" sounding tracks.

And as Rick mentioned this is only the recorder/ preamp section of things. Mics, both wired and wireless, can clip and that can't be fixed by anything down the line.

As to Juice Link... That safety level seems really low. But they are a consumer company building stuff for people who have no clue about sound. So they are kind of the "iPreamp". It just works, till it doesn't. They are also marketed to basically home movie folks. That "bar" is fairly low. They seemed to have stepped up their game a bit since they started, they started pretty low on the quality scale, and I'm not sure where they are on the consumer trash to professional splendor scale they are now. They don't give you a lot of control so they are definitely not in the "professional" end but their electronics might be decent now? I only played with an early model and it was noisier than , a place I don't think I can mention. More recent reports imply that they have made some big improvements in that respect.

Either way I would not spend a lot of time on their audio wisdom since it's likely to be the "... for idiots" version and less than really useful.
 
Hey guys,

Sorry to bring this thread back, but I recorded a short audio ilustrating what type of hiss I was talking about. Now, maybe this is acceptable, but I really wish I could record a bit cleaner...

I'm recording in a very small room, at about 1/2 meter (20") from the mic, using:

- Audix SCX1 Hyper --> SD MM1 --> Canon XH A1 XLR line input

Maybe it's the XLR cables that are introducing the noise, as I'm sure the gear is good enough to be silent. What do you guys think?

https://www.dropbox.com/s/jf8x2puk49xpv13/hiss test.wav?dl=0


EDIT:

Just realize the link is not producing any sound... not sure if it's a problem with my pc... Can anyone hear anything when playing the file?

EDIT 2:

I've put it on Vimeo. It's not that big of a hiss, but the gain is not that cranked as well... and I'm recording pretty close to the mic.

https://vimeo.com/181298546
Pass: hiss

Thanks
 
Last edited:
David Evans said:
Hey guys,

Sorry to bring this thread back, but I recorded a short audio ilustrating what type of hiss I was talking about. Now, maybe this is acceptable, but I really wish I could record a bit cleaner...

I'm recording in a very small room, at about 1/2 meter (20") from the mic, using:

- Audix SCX1 Hyper --> SD MM1 --> Canon XH A1 XLR line input

Maybe it's the XLR cables that are introducing the noise, as I'm sure the gear is good enough to be silent. What do you guys think?

https://www.dropbox.com/s/jf8x2puk49xpv13/hiss test.wav?dl=0


EDIT:

Just realize the link is not producing any sound... not sure if it's a problem with my pc... Can anyone hear anything when playing the file?

EDIT 2:

I've put it on Vimeo. It's not that big of a hiss, but the gain is not that cranked as well... and I'm recording pretty close to the mic.

https://vimeo.com/181298546
Pass: hiss

Thanks
Click to expand...

I think people tend to record at lower than necessary levels with setups like this. You have a mic that puts out a relatively strong signal at medium working distances, 3-4 feet. We use Audix hypercardioid mics in ceiling mount setups in small video conference rooms where working distances are typically 4-6 feet.
You have a nice preamp with an effective peak limiter (i'm assuming you mean a Sound Design MM-1).
If your XF305 audio meters are like my old GL2, they show a dBfs digital reference scale with "0" level at hard clipping at the right end of the scale. Minus 12dB FS is a reasonable "0 VU" equivalent recording level. The goal is to keep your recording levels for normal speech range centered around -12dBfs within +/-6dB leaving an extra 6dB of clipping headroom.
Do an experimental setup with your mic set at typical working distance. Set your Canon to manual levels and set the gain controls to the middle of the adjustment range. This is the 0 gain point and should give you close to maximum S/N for the camera.
Start with the MM1 gain at 42dB, limiter off, low cut filter at 80HZ.
Talk in a normal conversational tone of voice and see where the levels fall on your XF meters. You want them to average around -12 and peak a little bit higher, say -9. Adjust gain on your MM-1 as needed to get there.
Yell really loud to see if you can hit 0dBfs at that gain level. If yes, turn on the limiter. Yell again to see if the limiter keeps you below 0dBfs. Is it working? You want to trim gain balance between camera and preamp so the limiter keeps you from exceeding -3dBfs with a loud input. Turn down the gain on camera first to see if the limiter is actually operating. If it isn't you can use more gain on the preamp. You want to find the balance where the limiter is not cutting in at -6dBfs and is limiting anything that exceeds -3dBfs.
This should be pretty close to optimum dialogue recording setup for best S/N. Adjust gain only on the MM-1 to adapt to scene gain requirements. Loud conversations a bit less gain, quiet conversations a bit more.
If the XF lets you mono sum the input to both camera channels you can set channel 2 say 6dB lower than 1 for protection. But with this setup neither channel should ever hard clip. Or you can calibrate limiting for the lower gain 2nd channel so you get maximum unlimited range on channel 1, but potentially it can clip while channel 2 is fully protected.
 
Razz16mm said:
I think people tend to record at lower than necessary levels with setups like this. You have a mic that puts out a relatively strong signal at medium working distances, 3-4 feet. We use Audix hypercardioid mics in ceiling mount setups in small video conference rooms where working distances are typically 4-6 feet.
You have a nice preamp with an effective peak limiter (i'm assuming you mean a Sound Design MM-1).
If your XF305 audio meters are like my old GL2, they show a dBfs digital reference scale with "0" level at hard clipping at the right end of the scale. Minus 12dB FS is a reasonable "0 VU" equivalent recording level. The goal is to keep your recording levels for normal speech range centered around -12dBfs within +/-6dB leaving an extra 6dB of clipping headroom.
Do an experimental setup with your mic set at typical working distance. Set your Canon to manual levels and set the gain controls to the middle of the adjustment range. This is the 0 gain point and should give you close to maximum S/N for the camera.
Start with the MM1 gain at 42dB, limiter off, low cut filter at 80HZ.
Talk in a normal conversational tone of voice and see where the levels fall on your XF meters. You want them to average around -12 and peak a little bit higher, say -9. Adjust gain on your MM-1 as needed to get there.
Yell really loud to see if you can hit 0dBfs at that gain level. If yes, turn on the limiter. Yell again to see if the limiter keeps you below 0dBfs. Is it working? You want to trim gain balance between camera and preamp so the limiter keeps you from exceeding -3dBfs with a loud input. Turn down the gain on camera first to see if the limiter is actually operating. If it isn't you can use more gain on the preamp. You want to find the balance where the limiter is not cutting in at -6dBfs and is limiting anything that exceeds -3dBfs.
This should be pretty close to optimum dialogue recording setup for best S/N. Adjust gain only on the MM-1 to adapt to scene gain requirements. Loud conversations a bit less gain, quiet conversations a bit more.
If the XF lets you mono sum the input to both camera channels you can set channel 2 say 6dB lower than 1 for protection. But with this setup neither channel should ever hard clip. Or you can calibrate limiting for the lower gain 2nd channel so you get maximum unlimited range on channel 1, but potentially it can clip while channel 2 is fully protected.
Click to expand...

Thank you so much for your help! I'm not sure I understood everyting, but it actually helped me solve the problem. I had the Line in level of my camera at maximum, so I dialed it down to half way and pushed the gain in the Sound Devices MM-1. And that was it! I just tried it and sound was clean! MM1 rules, I love that piece of gear.

Thanks for your help!
 
Kind of a rule of thumb is get your gain as early in the chain as you can. Start by setting everything to unity gain (0 db), then turn your preamp up until you get the level you want, or it starts to get noisy. If at that point your levels are not high enough raise the gain of the next stage. In my experience, if two gain stages do not do it, something is wrong with your equipment or technique.

The whole idea here is do not amplify noise, only signal. Any gain stage past the first, or second one, should be considered a fader to reduce a too-loud signal. I am being precise here "signal" is sound you want, "noise" is sound you do not want!

On pro equipment unity gain is usually marked in some form. If not it is usually going to be 9 o'clock on a gain or trim control, or 3 o'clock for a fader control. Some digital stuff uses the center of the scale for both. Please note, that means you can actually turn the sound down some on the gain stage and up some on the fader stages, it is not a good idea to do either if you can avoid it.
 
You must log in or register to reply here.
Back
Top