Help I Use Adobe Adudtion! Noise floor problems

[brittanyw86]

I've noticed that a lot of you use Audacity, but if someone who uses Adobe Audition could help I really could use some solid advice. I've never had a problem with ACX passing my work, but here I've been told my noise floor is high and when I use checker (app recommended by y'all) I always get this error message ""The volume falls outside of the target range (86–92 dB)". I don't even know where to check my noise floor in audition and have looked everywhere. If you can assist with letting me know where that is too I'd be much obliged lol. So I ran denoise, but it seemed to muddy the audio and I disliked how it sounded. Then I found that if I used the average rms amplitude after running amplitude statistics before normalizing to -3db I could then set a noise gate to that number specifically it practically eliminates the noise floor. It sounds great, however I'm concerned that it gives off that complete silence vibe that is jarring to the ear that is deemed a no no of audiobooks.

[knotyouraveragejo]

A noise gate will do that if it is set too low. Can you have a link to one of your files? Both with and without the noise gate would be best. You can upload to the tests folder if you need to upload a file for this purpose.

[TriciaG]

Also, note that ACX is much more picky than we are at LV.

I just downloaded your "house of love" file, and it has almost no noise. Unless you want to improve the vocals on that, the noise level is WAAAY acceptable.

P.S. Checker is flagging the volume because it's too high. House of Love is 93.4 dB, and we aim for about 87-92. It doesn't have to do with your background noise.

[brittanyw86]

A noise gate will do that if it is set too low. Can you have a link to one of your files? Both with and without the noise gate would be best. You can upload to the tests folder if you need to upload a file for this purpose.

Thanks for telling me about the test folder !

https://librivox.org/uploads/tests/houseoflove_10_cheney_128kb.mp3
Here's a section I did. I got the noise floor down with running denoise, but I did feel like it muddied up the dialogue a little. They said this file's noise floor is ok now. I did this after feedback that the noise floor was too loud. I initially just edited it using the presets I usually run for ACX submissions that I run that pass acx's checker, but after installing the Checker app here it did not pass

Channel 1
Peak Amplitude: -3.01 dB
True Peak Amplitude: -3.00 dBTP
Maximum Sample Value: 23163.74
Minimum Sample Value: -22660.11
Possibly Clipped Samples: 0
Total RMS Amplitude: -18.33 dB
Maximum RMS Amplitude: -7.97 dB
Minimum RMS Amplitude: -87.13 dB
Average RMS Amplitude: -27.49 dB
DC Offset: 0.01 %
Measured Bit Depth: 32
Dynamic Range: 79.16 dB
Dynamic Range Used: 67.20 dB
Loudness (Legacy): -15.65 dB
Perceived Loudness (Legacy): -14.21 dB
ITU-R BS.1770-3 Loudness: -18.33 LUFS

0dB = FS Square Wave
Using RMS Window of 50.00 ms
Account for DC = true

https://librivox.org/uploads/tests/houseoflove_10_cheney_128kbnoisegate.mp3

Here's where instead of denoise I set a noise gate in the manner that I described above. I didn't have much luck setting the noise gate lower because it too seemed to clip the audio. I like the way this sounds better, but it does seem to have that void of sound I mentioned . After some more reading I found someone who used this method and then just added using multitrack a looped audio of good room tone. Not sure I'd want to go that route though if I can just find the right way to do this lol.

Channel 1
Peak Amplitude: -3.04 dB
True Peak Amplitude: -3.02 dBTP
Maximum Sample Value: 23095.73
Minimum Sample Value: -22456.64
Possibly Clipped Samples: 0
Total RMS Amplitude: -18.43 dB
Maximum RMS Amplitude: -8.05 dB
Minimum RMS Amplitude: -84.57 dB
Average RMS Amplitude: -30.15 dB
DC Offset: 0.02 %
Measured Bit Depth: 32
Dynamic Range: 76.53 dB
Dynamic Range Used: 63.70 dB
Loudness (Legacy): -16.01 dB
Perceived Loudness (Legacy): -14.09 dB
ITU-R BS.1770-3 Loudness: -18.42 LUFS

0dB = FS Square Wave
Using RMS Window of 50.00 ms
Account for DC = true

These stats that I am also listing are what I see when I run amplitude statics. Another thing I'm really trying to figure out is where do I find the information that PL are seeing when they check in Audacity and let me know about so that I can check for it in Audition that way I can check before submiting/fix any issues. Thanks for your assistance!!

[brittanyw86]

Also, note that ACX is much more picky than we are at LV.

I just downloaded your "house of love" file, and it has almost no noise. Unless you want to improve the vocals on that, the noise level is WAAAY acceptable.

P.S. Checker is flagging the volume because it's too high. House of Love is 93.4 dB, and we aim for about 87-92. It doesn't have to do with your background noise.

Sorry, I should have been more clear in that their Audiolab (https://www.acx.com/audiolab) will pass using the same presets (granted they want it at 192kb) that I ran for Librivox that aren't passing in Checker. I'm trying to see what I need to tweak preset wise for it to work for Librivox and find out where in Audition I can find out what they're seeing when they check the stats in Audacity.

Thanks for looking into the House of Love file. That file did get cleared for noise floor. It was just a general question so I can learn to do things correctly in the future.

This is what I see when I run a scan to check the file via amplitute statistics and it's not showing me that -93.4 db which is why I'm at a loss as to how to solve it/check it when editing. The ones I put asterisks next too are what ACX is looking for and the noise floor needs to not exceed -60db which I look at by just manually testing the room tone by playing it and seeing where it is in Levels.

Peak Amplitude: -3.01 dB
*True Peak Amplitude: -3.00 dBTP
Maximum Sample Value: 23163.74
Minimum Sample Value: -22660.11
Possibly Clipped Samples: 0
*Total RMS Amplitude: -18.33 dB
Maximum RMS Amplitude: -7.97 dB
Minimum RMS Amplitude: -87.13 dB
Average RMS Amplitude: -27.49 dB
DC Offset: 0.01 %
Measured Bit Depth: 32
Dynamic Range: 79.16 dB
Dynamic Range Used: 67.20 dB
Loudness (Legacy): -15.65 dB
Perceived Loudness (Legacy): -14.21 dB
ITU-R BS.1770-3 Loudness: -18.33 LUFS

0dB = FS Square Wave
Using RMS Window of 50.00 ms
Account for DC = true

The file you listened to I used denoise as sparingly as possible to lower the noise floor, but I didn't like how it muddied the dialogue. Any suggestions to improve the vocal aspect too are greatly appreciated . Thanks again for your assistance!

[lurcherlover]

I found the noise floor in both those example perfectly good, and maybe a bit too quiet? I only listened for a few seconds and found gaps so short that the noise floor was not noticeable at all. Maybe you are a bit too close miked? For the recording gain level you could back off a few inches perhaps without any need to increase the gain. But it's a matter of opinion and you might well disagree.

[brittanyw86]

I found the noise floor in both those example perfectly good, and maybe a bit too quiet? I only listened for a few seconds and found gaps so short that the noise floor was not noticeable at all. Maybe you are a bit too close miked? For the recording gain level you could back off a few inches perhaps without any need to increase the gain. But it's a matter of opinion and you might well disagree.

Thanks! Yeah I think it’s a bit too quiet too. Trying to find a better way that won’t make it not so quiet lol. Thanks I will definitely try experimenting with the gain too !