Hi all, been reading through the forums and thought I would post my sound recording from a Blue Yeti Microphone to see if anyone can help me with the noise.
I record from a shelf in my wardrobe which has pyramid foam on all sides except the top. The opposite wall behind me has a blanket over it. For both samples below, the mic’s gain dial is set to 1/2 on the microphone and 0.2 recording level in Audacity.
The first sample below is recorded through my PC.
The second sample is recorded with my PC turned off, so im using a Windows tablet.
Can anyone help me with the background noise? I would like to Audition/record from home for voice over work / Audiobooks etc. Also, in these samples - the peaks don’t reach 0.5 or -0.5 which from what ive read is the correct way to keep the recording to allow for compression etc? Is this correct?
The good news is this isn’t the famous “Yeti Curse,” the unremovable whine encountered buy some Yeti users. I got your clips to pass ACX AudioBook compliance with relatively modest tools.
You should download and install ACX Check and Steve’s voice filter. They make this voice thing much easier. I don’t have the latest forum versions, but these are my local posts.
I’ll get much more detailed when it’s not 6AM and I can get to the grownup computer.
The goal is to have your voice peaks around -6 (yellow zone) on the Audacity sound meters. This is the same number as .5 (50%) on the blue waves. Too far under that and you run the risk of noise problems and too far over risks overload and clipping.
If you’re not in Audacity 2.1.1, you should be. The noise reduction tools are far superior to the older tools.
Actually I would characterize the noise level in both of those recordings as “not bad”. The acoustics of your set up sound fine as well.
Listening to the noise the bulk of what I hear I believe is electronic in origin. There is a little bit of USB whine and the rest sounds like thermal noise from the mic preamp.
There is a lot of noise below about 120Hz which seems to be a common failing with USB microphones. If the goal is voice recording then a low-cut filter such as the “L F rolloff for speech” will help with the numbers (but not much with what you hear).
You can probably turn the gain up just a bit more when recording. Your aim is to have the meters on the audacity panel bouncing occasionally into the yellow but never into the red. It is possible that which knob you turn (the one on the mic or the slider in the computer) will have more or less effect on the noise. But to determine this you’ll need to do some fairly careful tests to make sure you get the mic to the same overall sensitivity level.
Thanks for the replies. I’ve re-uploaded a sample with me reading a quick sentence of something laying about which shows peaks around -6db. To do this, ive left gain at 50% on the Blue Yeti, but ive now moved the recording from 0.20 to 0.50 in Audacity. Are you both sure about it not being the Frying Mosquito problem? It sounds awfully similar - its much more prevalent in the below sample now with the recording volume turned up.
Koz, I use Audacity 2.1.0 but im sure the noise reduction hasn’t changed with the last update? I’ll check. I do have the ACX check already, from reading your other posts while I was lurking about and I did try the voice filter but i couldn’t really tell the difference.
Flynwall, i live in an Apartment building and the AC voltage here is 50hz. Not sure if that helps. Dumb question but can you explain to me how the low-cut filter helps if i cant hear any change? Basically i just want to get to a point where i can record and use noise reduction so that background hum/hiss cant be heard.
What i don’t understand is, the ACX check passes fine on the above sample but the background hiss/noise is absolutely awful. Does that mean ACX would accept a whole audiobook that sounded like that? That doesn’t make any sense.
You appear to have cleared up any troubles in your presentation. Attached. Last sentence.
how the low-cut filter helps if i cant hear any change?
The ACX AudioBook system measures everything whether you can hear it or not. You can make a delightful presentation, perfect in every way, but still fail ACX Noise because of thunder or earthquake sounds. Many USB microphones made low pitched sounds as a natural thing because they’re electrically difficult to get rid of, and why not? Nobody can hear them anyway.
OK, so LF_Rolloff is the same no matter where you get it from. Install it in your Equalizer.
Adding Audacity Equalization Curves
– Select something on the timeline.
– Effect > Equalization > Save/Manage Curves > Import
– Select LF_rolloff_for_speech.xml > OK. (it won’t open the ZIP. You have to decompress it)
– LF rolloff for speech now appears in the equalization preset curve list.
LF_Rolloff is used with the “length” slider all the way up.
In general that’s first step no matter what you have wrong.
It shouldn’t be. There is a cross-check. Right-click on the playback meters > Preferences > “RMS.” The meters will drop the colour scheme and turn into a dual meter system. The light green reads RMS or loudness. As you play your room tone (silent) segment, it should bounce to the left of -60. This is all ACX cares about. Or I should say that’s all the ACX robot cares about. If you pass that, then you go on to the human tester.
A common mistake is to crank your speaker volume all the way up while listening to your clip and exclaim how dreadful the noise is. The real test (assuming you pass the instruments) is set your voice volume for a pleasant level and, without touching anything, go back and listen to the Room Tone.
It’s also possible you have an unrealistic speaker or headphone system. Do you like EDM or Modern Disco? Tailored sound systems are rough because you will always be hearing different from us.
If it still bothers you, you can apply gentle noise reduction.
Noise Reduction
– Drag-select Room Tone, silence or the flat area between spoken phrases.
– Effect > Noise Reduction: Profile
– Select the whole clip or show by clicking just above MUTE.
– Effect > Noise Reduction: Settings 6, 6, 6 > OK
For the most part, the clip will not have changed except the noise will have dipped a little bit. Better? If you want more, UNDO that one and apply slightly stiffer Settings of 12, 6, 6.
You are warned against creating a performance against the Blackness of Space silence. ACX has a failure they call “Excessive Processing.” That’s also the reason we stopped calling it Noise Removal. Too many people wanted it to Remove Noise…to zero.
I think you have achieved the natural limit of what you can do with your microphone and environment. As FlynWIll points out, you have rain-in-the-trees hiss from the microphone amplifiers and a little Yeti-Curse digital whine waaay in the background. Only about a third of the total noise is room/environment and probably not worth the effort to get rid of.
It shouldn’t be. There is a cross-check. Right-click on the playback meters > Preferences > “RMS.” The meters will drop the colour scheme and turn into a dual meter system. The light green reads RMS or loudness. As you play your room tone (silent) segment, it should bounce to the left of -60. This is all ACX cares about. Or I should say that’s all the ACX robot cares about. If you pass that, then you go on to the human tester.
Yeah I have this but, for example, during the room tone parts the meters bounce to the right of -60. I guess I should keep
A common mistake is to crank your speaker volume all the way up while listening to your clip and exclaim how dreadful the noise is. The real test (assuming you pass the instruments) is set your voice volume for a pleasant level and, > without touching anything> , go back and listen to the Room Tone.
Yes! This is exactly what I have been doing. I have studio headphones and I have the volume at maximum - I thought this is what we are supposed to be doing? If I can hear background noise in an audition, then so can the person on the other end. I assumed this was not professional and that if a potential client can hear background noise then they wont hire me? I mean you never hear background noise in a demo, right.
It’s also possible you have an unrealistic speaker or headphone system. Do you like EDM or Modern Disco? Tailored sound systems are rough because you will always be hearing different from us.
I use Sony MDR-7506 headphones for all the editing.
You are warned against creating a performance against the Blackness of Space silence.
Ok, so for Audiobooks, background noise is perfectly fine as long as it’s at -60 or lower. It seems I will still need to upgrade my setup then because I don’t think it’s acceptable for auditioning for voice over work.
And yes, your voice works very well.
Thank you for that, it’s very kind of you to say so.
See what happens if you up the Noise Reduction correction to 18, 6, 6? That may do it for you.
The next step is to cast off the Yeti. There are limits to what can be done with an inexpensive USB microphone.
You have my permission to crank up the volume during Room Tone. Hear that mosquito back there? EEeeeeeeeEEEEeeeee. That’s the Yeti digital sound damage. It’s pretty much permanent and wedded to the microphone/USB Cable/computer combination.
The rain in the trees fffffffffffff is the noise that the microphone preamplifier makes. You’re stuck with that in any microphone, but it tends to go down as the cost of the microphone system goes up.
The rest is probably computer fan noise?? I can’t tell. You might be able to get rid of some of that by isolating the microphone from the desk.
noise is perfectly fine as long as it’s at -60 or lower.
Assuming it’s gentle rain on a spring afternoon ffffff noise. That fades into the background if not puts everybody to sleep.
Not so Frying Mosquitoes noise generated by The Yeti Curse. That has a tonal characteristic of baby screaming on a jet. Almost impossible to get rid of without damaging your voice and it doesn’t matter what the loudness value is.
“It’s at -70 and I can still hear it. Make it stooooooop!”
This is all about the ACX system. They adopted roughly the same sound standards as broadcast with the idea that no kid with a Mister Microphone is ever going to pass them. Then they made performers submit to The Robot which automatically checks for the three major sound values, background (noise), peak (overload), and overall loudness (RMS). No Mister Microphone ever passes all three. A number of Yetis won’t pass all three.
Several posters report after making it through Audacity processing and testing, they have been accepted by ACX for publication. ACX-Check seems to work the same way that the ACX Robot does. Darn good guess because I don’t think we ever got a straight answer how their robot works.
~~
If you have another application, like voice-overs (or voices-over, see, the noun is “voice” and takes the plural, not “over” which is a preposition/modifier), then you need to find out what their specifications are, and I wouldn’t guess at it. Passing ACX and sounding good is a darn good place to start. If you think the noise is still too high, see what you sound like when you suppress it further with higher numbers. Remember to check for clarity and quality of voice as you do that. Noise Reduction works on everything, hopefully the noise more than the voice. The old Noise Removal would destroy voices given any chance at all. It wasn’t particularly valuable.
Another note. Many people wouldn’t dream of submitting a professional voice recording without DeEssing and Sibilant Processing. Our DeEsser needs significant hand-holding to make work. The instructions and interface are in development. I’ve never used it.
Thanks for these samples. The only noise I hear (with HD201 cans), is LF USB power noise. And from that I can tell your computers have good power supplies. So, no frying mosquitos.
The noise should be controllable in Audacity with a highpass filter that sets in at 125 Hz and a not to steep rolloff. -12 dB or less. If needed, a bit of noise removal could follow.
Also, if you’re a tinkerer, get a USB extension cord and a good 5V, 500 mA power supply. Cut the power wires (red and black) in the extension cord, route them to the power supply. That should get rid of the Yeti curse. Don’t forget to isolate the unused power wires coming from the computer.
Soundwise, the Yeti seems one of the better USB mics. Worth rescuing, imho.
FlynWill and I don’t agree. We both hear that annoying 1KHz data management whine waaay in the background. It’s not at all obvious, so it’s easy to miss—or suppress with energetic Noise Reduction.
You shouldn’t have any trouble meeting noise with any of those Noise Reduction values. Have you tried any of them?
If you’re going to go all purist on us, then you can stop using the Yeti. I was able to make an ACX compliant voice sound clip with very little processing in my nearly soundproofed third bedroom but I did it with a rock-band analog microphone, a good, tiny, but well-behaved analog sound mixer and a Mac.
That can be done, but probably not with a USB microphone.
There was a blog posting recently of someone who accidentally locked herself in a hotel closet while she was recording an overdue AudioBook. She wasn’t using a USB microphone, either. She was using a stand-alone recorder and had stuffed the closet with blankets and quilts from the bed.
All went well until she tried to leave…
Was the hotel trying to prevent the guest’s Wellingtons from breaking out? I don’t know that we ever got the explanation of why the closet doors lock, but she apparently does this all the time.
The noise should be controllable in Audacity with a highpass filter that sets in at 125 Hz and a not to steep rolloff. -12 dB or less. If needed, a bit of noise removal could follow.
I gave this a go at and around 125hz using the last sample I uploaded to this thread but it made no difference.
Also, if you’re a tinkerer, get a USB extension cord and a good 5V, 500 mA power supply. Cut the power wires (red and black) in the extension cord, route them to the power supply. That should get rid of the Yeti curse. Don’t forget to isolate the unused power wires coming from the computer.
I’d like to try this, but I just need more in depth explanation - like routing the wires to the power supply. Also, even if I route this blue yeti usb directly to a power supply - then how does the sound get recorded? I’m missing something obvious here.
This is all about the ACX system.
Unbeknownst to me, ACX is only for US/UK people wishing to narrate audiobooks - so that door is now firmly shut but I will still keep the requirements in mind if I ever try audiobooks in another avenue.
The sample below is the best I could do. I used the LF Rolloff, followed by Noise Reduction of 22/10/0. I could never reduce the sound without making the breath sound distorted.
Cut it. Reconnect the green and white (or yellow and blue) wires. Alternatively, peel off the outer jacket and don’t cut those. Cut only red and black.
Attach the red and black wires to a good 5V power supply on the side of the cable that is going to the mic.
Done.
You could also use a USB Y-cable and cut off one end to connect to the red and black wires. These are made to power portable harddrives from 2 USB ports. Maybe some only have power wires, but I’ve never seen those.
If you have a DMM, check the output of the power supply. It’s not rare for USB chargers to put out much more if there’s no load. A battery charger circuit doesn’t mind, but digital electronics might just blow when fed more than 5.25 V.
Now that I can really listen to it, it’s mostly digital noise. And Koz and Flynwill are right, it’s digital noise. You can almost see the transmission bits if you zoom in enough. There’s another digital noise superimposed on it. And that’s impossible? to filter out.
And that points to insufficient filtering in the Blue Yeti. Or even bad design, though I would expect better from Blue Microphones.
Thanks for the detailed reply, I fully understand now. Makes sense. I wont be able to get my hands on a DMM here though, but the 5v power supply should be ok. I’m sure there is some sort of powered device that I could buy just to take the p/supply from it.
And if I do this, you think that it should completely remove the digital noise?
Flynwill was able to remove the very tiny high-pitched digital whine from his Behringer UCA-202 by this external battery or power supply trick. He did it inside the cabinet of the device rather than breaking into the cable. It won’t do a thing for the other noises and it doesn’t affect the low frequency rumble.
The rumble isn’t a problem. Steve’s LF-Rolloff is a known good performer. It’s possible this filter will be included in future Audacity as a native filter.
I used the LF Rolloff, followed by Noise Reduction of 22/10/0.
I’m curious why you chose those Noise Reduction values, and I’m still curious why you chose that particular headphone.
We use ACX and their standards because they’re rational, functional, and they’re nearly the same as broadcast. So it’s a much wider net than just the prickly requirements of a single US company. It’s as close as we can get to a guarantee of a marketable product.
We do insist this has nothing to do with theatrical content and ACX humans do actually listen now and again, particularly if you’re a new reader. If people cover their ears and flee when you speak, no filter is going to help you.
I’m curious why you chose those Noise Reduction values, and I’m still curious why you chose that particular headphone.
These particular headphones an industry standard, and I used them when choosing the reduction values because its what sounded best to me in the short time I played with it.
I’m curious why you chose those Noise Reduction values, and I’m still curious why you chose that particular headphone.
Please don’t cherry-pick the questions. Answer all of them. It saves a lot of time.
Why did you pick those Noise Reduction values? I posted multiple times about the reduction values many people have found valuable for clear, desirable voice recordings. Starting with 6, 6, 6, (the noise reduction of the beast), 12, 6, 6 and 18, 6, 6 if necessary. Other reduction value combinations may be needed for high noise music presentations, but they can cause voice problems.
I asked about the headphone because they work well, are very popular in Hollywood, but tend to be a “hidden secret.” They’re not something a beginner would naturally know about. No audio operator would be caught dead without at least one pair. Do you know people in the movie industry?
