Signal to Noise Ratio - Sound Speeds

The noise floor of the room you're recording in will likely play a huge part in the decision to use production sound. Of course if you're recording a voice over or wild lines and the background noise is too high, you might as well record without the mic plugged in. In short, the noise level of the recording environment is directly related to the usability of that sound. In this video I'll explain why.
For more info on noise floor: en.wikipedia.o...
To learn about signal to noise ratio: en.wikipedia.o...
Make sure to subscribe for more content!
Have a question? Write it in the comments below or email [e-mail]
Join us on Facebook: SoundSpeeds
Follow us on Twitter: / soundspeedsyt
Interested in becoming a patron? www.patreon.com/SoundSpeeds
To make a donation: www.paypal.me/SoundSpeeds

[@JuliusSDeLaCruz]

You provide so much value in your videos. Thank you!

[@SoundSpeeds]

Thank you and thank you for watching!

[@hztan1263]

amazing video!

[@SoundSpeeds]

Thank you and thank for watching.

[@TomMeyer01]

Really great info man. Thank you !

[@SoundSpeeds]

Thank you and thank you for watching!

[@billgates3699]

Old movies are better for countless reasons and high noise is one of them. Noise makes for “deafening silence” so when the scene becomes completely silent with a high noise floor it builds tension. Notice old movies are never fully silent. Never. It really highlights the stillness and suspense going on.

[@SoundSpeeds]

It sounds more human to me too.

[@JustinPorterMedia]

If you are ok with it, I am making a "Audio bag" video and I speak shortly about Signal to Noise Ratio. This video is what helped me understand it in great detail and I want to link it/mention it in our video. I will of course credit you and add the link to this video in our description.

[@SoundSpeeds]

Sure thing. Thank you. I also did this video which might offer you some unique additional perspective as well: kzbin.info/www/bejne/gXjCZIl8eNWSprM

[@JustinPorterMedia]

@@SoundSpeeds you’re amazing. Truly appreciate your knowledge.

[@SoundSpeeds]

Thank you Justin. Feel free to email me if you're looking for some info and can't find it online. I may have a video about it.

[@JustinPorterMedia]

@@SoundSpeeds oh say no more. Truly appreciate it. You’re extremely knowledgeable. Thanks for making yourself available to one man bands like myself 😂

[@SoundSpeeds]

👍

[@anyuisbjoern]

In my last film ( Im not a professional filmmaker, just do it as passion ) we had one scene where a young woman sits at the table drinks some whine and talks silently self-talk. Then she stands up and puts a vinyl on quite loud and dances to it. I had a sennheiser mkh 416 plugged into a zoom F6 in 32bit float mode. So there was a bit of security. The music where captured with a boom pole. A second microphone were on the table which also had the purpose of having a backup of her voice.( A condensor-stereo mic connected on a H6 zoom). Later in editing I noticed how silent her voice were in comparison to the opening of the wine bottle with screwing metal lit or pouring in of whine or other high pitch sounds peaking. I later in post I pulled down all the high peaks manually but due that, the audio of the scene got slightly unnatural. Do you know any idea on how you would record this situation better?

[@SoundSpeeds]

Yes. First if all, 32 bit float on an F6 was the wrong call. It's not going to give you the ability to pull noiseless audio up to a normal level and furthermore sacrifices quality while running in 32 bit float. Check out this article: www.audiosciencereview.com/forum/index.php?threads/zoom-f6-portable-field-recorder-review.15668 What you need to do depends on the situation but I would have recorded 24 bit, gained up properly and asked the Director if the actress could give me some more volume. A whisper doesn't have to be a whisper, it can be a stage whisper for example. Talk to the Director and see if you can run wide shots separately from the tight shots so you can dig in close with the mic. The other stuff shouldn't be used all the time especially if they want usable dialogue. Communicate that and either you'll get your silence and proper dialogue recording or the Director doesn't care and he/she can sign off on it and you can write a note in the sound report then not care.

[@anyuisbjoern]

@@SoundSpeeds Thank you for your well thought answer! Thank you for posting the forum post about the quality of the F6 and 32bit float. Mmh yeah! That’s sounds logical, that we should have better planned the sound recording of the scene. The director, she wanted to have everything a bit like a performance and spontaneously. As a kinda beginner I just went for it and try to make the best with my limited knowledge. The actress she also asked me if she should speak louder and I denied cause my thoughts were, that this is more authentic. But for sure if you have the noises of glass or bottle lids popping in occasionally, the whispering voice isn’t as prominent anymore. We only had one long shot for this scene without break. Otherwise I could have adjusted the gain for each shot. It was the first time I used the MKH 416 and was suprised how sensitive it is to noises around. Probably the idea with the stage whispering would have worked great, but sadly we were so determined to have everything an authentic-homevideo-performance-character that we didn’t want to stage too much, as you would usually do on a film set.

[@SoundSpeeds]

I hear that a lot from Directors. The question I ask is if they want the actors to try to recreate their amazing performance in an ADR session because if not, we can capture their audio performance with just a little tweak.

[@AntAskew]

I’ve never been able to find a solid answer for this, should you up the gain or the levels on an audio mixer. I know to a degree the two work hand in hand. But which one should me the main focus?

[@SoundSpeeds]

AntAskew Typically you set your fader to 0 db then adjust the Trim until the signal peaks at the level you like mixing to. Some people mix to -20 db but most mix to -10 db. You cannow start mixing so watch for your peaks to hit -10 db. You'll have extra headroom if you need it and on the fly you can tweak your levels. If you're riding the fader pretty high, dial up the trim a bit so you can back off on the faders a bit. The opposite is true as well.

[@hippo-potamus]

What does the S/N ratio mean in the context of a car audio amplifier? What would be considered the "noise?" Thanks

[@SoundSpeeds]

Buzz or hiss normally

[@hamedfannan4464]

if I want to choose a poor SNR, an average one and high SNR, what ratios do you suggest?

[@SoundSpeeds]

Regarding... what device? A mic - subtract the self noise from 94 to get the S/N. 20+ isn't great but 12 and below is.

[@underwoodvoice9077]

A while back, I had George Whittam (of Edge Studio) check a demo file I recorded for him. Turns out my noise floor, normalized to 0db, was -65 db. I was, as the Brits might say, dead chuffed.

[@Gregg0Palmer]

Ok I admit it....I had to google "dead chuffed"

[@SoundSpeeds]

That's awesome! I'm assuming you used an expander or gate.

[@SoundSpeeds]

Gregg Palmer I did too although I had an idea from his usage.

[@LoveToHateAggro]

I have bought a Zoom H1n to record interviews and stuff like that... Idk why but I thought that with a recorder and a lapel mic I would be able to 'isolate' background noise and just pick up the person's voice... turns out that the recorder is very sensitive, it picks up sound far away, any little noise in the background is picked up... Idk If a better/high end recorder would do a lot better... At the end of the day, I think there is no such thing as 'isolating' the background noise that much, right?

[@SoundSpeeds]

There are uses for lav mics and ways to make them more effective. Lavs pick up less background noise (despite having a higher self noise) than most mics but you still want a quiet background to give yourself the best chance to pull your dialogue out of the background. When you say a lav on the H1N is so sensitive it pulled out a lot of noise, that sounds to me like either incorrect gain staging, a less than ideal lav placement requiring more gain or just a cheap or malfunctioning lav. Without hearing the sound, I can't do more than speculate.

[@LoveToHateAggro]

Yeah I did some quick tests, I've set the lav with the clip-on that came with it... I intend to hide the lav under the shirt, sticking it to the skin or the undershirt, that would be ideal for me... the knob which sets the gain on the H1n was on 5 to 6 (I saw on a forum that gives about 20-25db of gain, have no clue), the lav is a PowerDeWise omnidirectional, it's a cheap one but very good, I think it's not the problem. Sorry for the noob question, 0 knowledge about recording sound, I'm getting into it right now, definitely need to learn a lot and your channel is a great resource... by the way, great video i got to understand what is signal to noise ratio that everybody talks about but no one explains what it is for us beginners. Thanks!!

[@SoundSpeeds]

Sure thing. Thanks for watching!

[@ebasoufeliz]

I love your videos. Lately I've been searching about recording equipment and techniques because I want to start recording music at home and your channels has been helping SO MUCH, really. Thank you. If it doesn't bother you, I'd like to ask a question. I have a budget dynamic mic and the noise floor I get is recorded at around -54db in audacity and, when recording, I get signals of around -30db consistently. Is this SNR THAT BAD? I feel like it is, as to me it sounds unbearably loud. I don't mean to eliminate the noise, but I kind of notice it just too much to be sure of the real quality. Even after post processing it bothers and makes me hate compressing or normalizing. I heard and read that preamps may help increasing gain and improving the SNR, but is it true? Also, I'm a starter in every aspect of recording and mixing, so I might be doing something wrong. I get the same noise floor on my headset's mic. I use an adapter for my dynamic mic, directly plugging it to my PC. Is it bad? Thank you, anyway!

[@SoundSpeeds]

Yes, that's bad! A dynamic into a computer is not the way to go. You need a USB audio interface and can get a Behringer UM2 inexpensively. Your dynamic mic is very low voltage out and your computer input get boost that enough to make it sound good.

[@ebasoufeliz]

@@SoundSpeeds I will definitely check it out. Thank you!

[@SoundSpeeds]

Sure thing.

[@VoyVod242]

Hey there, I'm starting out in sound after graduating from film school and I'm trying to get better at making informed decisions about noise floors on set. I already know what is easy to get rid of based on what part of the spectrum it occupies, and I have some experience in determining what kind of S/N-ratio is acceptable, but I'd like to hear your take on it, especially considering noise reduction in post, if you have insight on that as well. Thus far I've found that a signal to noise ratio of ~20dB is just about the minimum that's going to sound decent and I always go for something better, obviously. What would you consider the minimum S/N-ratio for something useable?

[@SoundSpeeds]

Ilari Sivil That entirely depends on post. I've been on shows that ADR everything by policy then compare ADRed dialogue to production audio and they don't care about S/N until they make the decision as to which one to use. Crazy, right? There's not really a number I'm aware of because it depends on what kind of noise, the frequency range, if it's a constant or oscillating, etc. I don't let noises go on set if they can be shut down because I believe you should always strive for the lowest noise. One exception... If there's no dialogue I'll let some go if it's comfort related like heaters at video village or the actor's heating tent during cold months. The mixer may demand me shut everything down but I try to pick my battles. Some mixers that use a Schoeps SuperCMIT 2U mic engage the noise cancelling feature and use that as reference for what you can get away with on set. Others may engage CEDAR noise cancellation (if they have the box) and route a second mix track thru it as a courtesy for post to use in the dailies mix. It's also a good reference if you have $4000+ to spend on one. Does this help?

[@VoyVod242]

Sure, this helps. I've used the CMIT5U on a couple of occasions and I'd love to try the SuperCMIT. I already kind of knew about some mixers using CEDAR, but neither option is financially available to me at this point, and I don't even know of anyone renting either out in Finland, where I operate for now. The reason I asked is because I've been working in a lot of uncontrolled environments lately, and thinking about what can be reasonably removed has been keeping my sanity intact. Corporate stuff, the occasional interview, documentary-type conditions where there's not too much I even get to do. What makes matters a bit more difficult, is that I've mostly had new clients that I don't have a long working relationship with, so being really strict feels like a social risk. I strive for the lowest possible noise like everybody with an ounce of common audio sense, but I also have to pick my battles. I think having an idea on what can be reasonably removed will help. On narrative stuff, everything that can be removed on set gets removed on set, but I don't want to be the guy who is constantly slowing down production. If someone asks me whether take 7 in a bad location was OK since it was the quietest, but still kind of gnarly, I'd love to be more certain of the call I have to make. I'm working on finding the balance, but I'm already pretty conscious about limiting my requests so that I can be sure that I'll get what I really need to not have to be ashamed of my tracks.

[@SoundSpeeds]

Oh yeah, I hear you! You don't get more money from clients if you do use a CEDAR or SuperCMIT so it's difficult to justify the cost. Do your best and when you have, you gotta just make notes in your sound report or you can push and push until that last little thing is resolved.

[@sebek0311]

Hello, Thank you for the fantastic video! I have one question if I may. Does noise floor of a phono stage or an amplifier affect how much hiss it will be produced by those devices?

[@SoundSpeeds]

Thank you. If an amplifier's noise floor is below that of the signal put thru it, you shouldn't unless the noise floor frequencies are different and the background noise level is in a different range,

[@sebek0311]

@@SoundSpeeds Many thanks for the replay. It helped a lot!

[@SoundSpeeds]

Glad to help and thank you for watching!

[@AlbertMajjor]

Thanks for the video great information but being an engineer myself, background noise will always exist, and this is why post production, creating of Foley and sound effects, music and automated dialogue replacement (ADR) exist

[@SoundSpeeds]

Very true. You don't need a sterile audio track anyway.

[@edenlim5542]

Hi all gurus, i need some clarification. soundblaster k3+ has a -93db SNR while scarlett 2i2 has a -128db, which is better?

[@SoundSpeeds]

Sorry but there's not enough info to calculate one from the other. I looked in the user manual too and there's only the -93db you mentioned.

[@jack_lion]

RX by izotope has some amazing tools for this issue

[@SoundSpeeds]

Yeah they do! Top notch at that

[@SunriseWaterMedia]

That is sound advice.

[@SoundSpeeds]

Hey! That's my line.

[@cuisined]

3:45 kill the dog lol (kidding!)

[@SoundSpeeds]

Poor dog

[@cuisined]

Hey..Left you a question on your podcast KZbin channel.😎

[@SoundSpeeds]

Ok, I'll check it

[@Jakartans]

what is DB? what -12DB means in GH5 for example, and why noise level of mic have several ways of measurements... some says 12DB, 10 DB some says -69Db... -80DB... bit confuse the difference.. thanks

[@ChrisPFuchs]

Jakartans dB stands for decibels, the unit of loudness. dB-SPL measures the sound pressure level of an acoustical waveform. The perceived volume of sound is basically proportional to the logarithm of energy, meaning it takes more and more energy to increase the perceived level of volume. Humans also hear different frequencies at different levels, check out the fletchur-munson curves that demonstrate this. So multiple types (if you see A-weighting vs another weighting) is taking this into account. Analogue audio has to have its own way of measuring dB since you can't measure the Sound Pressure Level of electricity in a current, and digital audio has its own way of measuring since you can't measure the Sound Pressure Level or electrical current of digital 1's and 0's. It starts to get complicated and kind of depends on the bit-depth to determine the possible dynamic range, but for digital, it's measured in a way that 0=clipping (no more bits to define the amplitude of a specific sample), so all audio will be read as in the negatives. Hope that makes sense and I didn't mix something up. I'm sure someone will correct me if so.

[@Gregg0Palmer]

One bel = is a power of ten difference 10^1. One decibel (db) is one tenth the power of ten 10^.1. So a +20db gain is 10^2= 100, a -20db loss is 10^-2=1/100.

[@Gregg0Palmer]

A dB ratio is almost always a logrithimic comparison to a reference value, (e.g "dBFS" = max digital value or full scale, dBV = 1 volt, dBu= .775V volts, dB SPL = sound pressure level relative to quietest thing humans can hear)

[@Jakartans]

Can you explain it in human language..:)? Does it means +20dB gain means 100times of the current vol? I heard that if I want to reduce the volume by half... I need to turned it down by -6dB some say -10dB? I can turn down the gain level of my GH5 to max -12dB, what is it actually means..?

[@Gregg0Palmer]

dB +20 dB is 100 times increase of power or energy, however we do not perceive audio nor light in a linear fashion of power instead we perceive these much more in a logarithm fashion, meaning we have a huge range of perception, but the our perception sensitivity steadily decreases with increasing energy. The net effect is it takes more and more energy for us to perceive a doubling of the sound. So 6 dB or 10^.6 or 3.98 increase in the energy is required for us to perceive a doubling of the sound....I don't have a GH5 so I'm not sure about it's settings, but for general guidance: .... if you have audio meters that peak at 0 dB (0 dBFS), that means at 0dB the camera can read no higher volume. Any volume above 0dB and will in fact be recorded as 0db by the camera, this is clipping (a very very bad thing). A clipped signal will have a very distorted sound when played, as all the signal peaks will flattened to 0db, making it sound very unnatural and usually unusable. When recording audio you want to leave plenty of head room above the average peak signals you are seeing on the meters for unexpected increases of volume, so worst case nothing will ever go over 0dB. Most people record audio with a peaks target of -12dBFS or -20dBFS. That way there is plenty of room for actors to suddenly change the volume of their delivery (yelling, or laughing) or for sudden instantaneous sounds (slamming a book down, clink of dish, slamming of a door). I would adjust the camera audio sensitivity such that the audio meter peaks are between -12 and -20 dB. If there is any yelling, laughing, or slamming things go for -20 dB, otherwise -12dB usually works well. You never want to see the signal reach 0dB, ever, because you don't know if the signal was clipped or not....They key is you need to adjust the camera audio volume by seeing what the meters are reading and then make adjusts to bring the peaks into a range where there is plenty of room for unexpected sounds. ....hope this helps....

[@piotrmasztalski423]

Please don't use these flashing white screen transitions. Very annoying especially while watching in the dark.

[@SoundSpeeds]

I see. Didn't think about that. I'll try fast dissolves and see how that works. Old habits die hard. Jump cuts drive me nuts and look incredibly sloppy to me. Thanks for mentioning this.

[@sebek0311]

It is sad that all you had to say after watching this great and informative video was that. No appreciation at all.

[@SoundSpeeds]

It is the way... in 2020

[@TrickGas]

@@sebek0311 Agreed with J Alex, all I thought while watching the video was how informative it was. Not something as minute as the transitions lol this is a video about Audio.. Maybe our friend Piotr was having a bad day, cheers!

[@SoundSpeeds]

I did change that transition shortly after this video. It's no biggie to me either way and if it helps someone concentrate on the content of my video, I'm fine with the change. Thank you so.much for your support and for watching!