Zoom H6, Deity S-Mic 2 , Sennheiser MKH 416

Any preamp will generate noise, especially when it is cranked up. That's one reason that dynamic mikes like the SM7B and SM57/58 don't always work well. They put out very low signal levels.

You haven't mentioned anything about the software that you are using. There is noise reduction available in most DAW packages, some better than others. I processed this in Reaper using the REAFir plug-in. Does this sound quieter to you? Its your TR2 Wave file from above.
Thanks, TalismanRich! I record to SD card in Zoom H6 and then use Adobe Audition CC to process the file. I usually use the DeNoise option, and with 10-20% I will get the noise removed. If I use more than 20%, the voice get slightly distorted.
In your sample, I cannot hear any noise, but the voice also sounds differently, less natural.
 

TalismanRich

Well-known member
A FETHEAD will boost the signal, but if it is microphone self noise, then you will simply boost the noise level of the mic along with the signal. You really need determine if the source is from the preamp or microphone (probably a combination).

In your tests, you had the microphone plugged in. Try this: Plug in the mic, set the gain at 7 (that seemed to be a noisy level) and record about 15 seconds of sound. Stop, unplug the mic, and record 15 more seconds. Compare the two. This should tell you how much each component is contributing.

As for posting some of your mediations, it won't matter if they're in English, Russian or Martian! We just want to compare your voice to the signal level. Unprocessed would be best. My concern is that if we can distinctly hear your breathing, you're probably speaking too softly.
 

TalismanRich

Well-known member
Anton,
Definitely the mic is contributing most of the noise. I ran a quick analysis of the two and the mic is adding almost 10dB of noise to the signal.

Anton Noise Analysis.jpg
I tried a couple of things to reduce the noise, but, like you said, it affects the overall tone. Using a noise gate doesn't work well because you use breathing examples. Set the gate too high and the breathing gets cut out. Set too low, and it doesn't cut anything. The noise reduction algorithms are better, but the breathing has lots of high frequencies and they start to "warble" in and out as it goes through the noise reduction threshold. I tried using a low pass filter from about 4-5 K to see if that helped, it did a bit, but there are still artifacts.

Which leaves us with two paths... 1) get a mic and/or preamp system with much lower noise (and remove as much ambient sound as possible from the room) or 2) adjust your technique. #1 is the easier way, but more expensive. Watching an evaluation of the Deity S on Booth Junkie's Youtube channel, the Deity seems to be very similar to the Senn MK416, but he is definitely talking at a high volume. He also has a booth that seems to be VERY quiet.

As for adjusting technique, I assume you have your headphones on, and are monitoring things as you speak. Maybe try something like removing your headphones, and speak as if you are working with a group of 4-6 people in a live situation. I bet your voice will be louder. That might give you enough boost to overcome the noise.
 

keith.rogers

Bobby'); DROP TABLE USER
Voice recordings shouldn't be normalized to -0.1dB, but to an appropriate loudness (dB LUFS) measurement. I've heard that for a single mono track, you might want as low as -23 dB LUFS, which will convert to about -20dB LUFS if mixed to dual mono. If you are applying any spatial/time effects (e.g., reverb) to create a stereo mix, then you'd mix master that to about -20dB LUFS, or no more than -17dB LUFS, I'd guess. The peak level is irrelevant, unless getting to the LUFS level would create clipping, and so then you'd need to investigate some compression and/or limiting.

I did hundreds of recordings with H6, but in live venues with people that were singing, and not mumbling into a microphone, typically something like an SM58, so the preamps were never an issue. I think you should work on your presentation a bit, though; but, honestly, a Zoom H6 would not be my choice for doing voiceover or audible-book type recordings at all. It's really a live, audio-for-video kind of device. If you want gain with low noise, buy a Zoom F6 or used F4 if you can find one, or get an actual audio interface with lots of clean gain, and ditch the H6.
 
Voice recordings shouldn't be normalized to -0.1dB, but to an appropriate loudness (dB LUFS) measurement. I've heard that for a single mono track, you might want as low as -23 dB LUFS, which will convert to about -20dB LUFS if mixed to dual mono. If you are applying any spatial/time effects (e.g., reverb) to create a stereo mix, then you'd mix master that to about -20dB LUFS, or no more than -17dB LUFS, I'd guess. The peak level is irrelevant, unless getting to the LUFS level would create clipping, and so then you'd need to investigate some compression and/or limiting.

I did hundreds of recordings with H6, but in live venues with people that were singing, and not mumbling into a microphone, typically something like an SM58, so the preamps were never an issue. I think you should work on your presentation a bit, though; but, honestly, a Zoom H6 would not be my choice for doing voiceover or audible-book type recordings at all. It's really a live, audio-for-video kind of device. If you want gain with low noise, buy a Zoom F6 or used F4 if you can find one, or get an actual audio interface with lots of clean gain, and ditch the H6.
Thanks for your comments! I use normalized -0,1bd to work with the recording (so I can hear everything), then I use "match loudness" and set it for -19db LUFS. This is what I heard you need for youtube and other places in the Internet.
What about Rode AI-1 or Focusrite Scarlett Solo?
I only have a tablet with Windows 8.1, which, I guess, could be used for recording in this case (it is fanless).
 
TH
Anton,
Definitely the mic is contributing most of the noise. I ran a quick analysis of the two and the mic is adding almost 10dB of noise to the signal.

View attachment 108068
I tried a couple of things to reduce the noise, but, like you said, it affects the overall tone. Using a noise gate doesn't work well because you use breathing examples. Set the gate too high and the breathing gets cut out. Set too low, and it doesn't cut anything. The noise reduction algorithms are better, but the breathing has lots of high frequencies and they start to "warble" in and out as it goes through the noise reduction threshold. I tried using a low pass filter from about 4-5 K to see if that helped, it did a bit, but there are still artifacts.

Which leaves us with two paths... 1) get a mic and/or preamp system with much lower noise (and remove as much ambient sound as possible from the room) or 2) adjust your technique. #1 is the easier way, but more expensive. Watching an evaluation of the Deity S on Booth Junkie's Youtube channel, the Deity seems to be very similar to the Senn MK416, but he is definitely talking at a high volume. He also has a booth that seems to be VERY quiet.

As for adjusting technique, I assume you have your headphones on, and are monitoring things as you speak. Maybe try something like removing your headphones, and speak as if you are working with a group of 4-6 people in a live situation. I bet your voice will be louder. That might give you enough boost to overcome the noise.
Thanks a lot! I will try! And yes, it is similar to Mk416 according to Booth Junkie, that is why I bought it.
 

keith.rogers

Bobby'); DROP TABLE USER
Thanks for your comments! I use normalized -0,1bd to work with the recording (so I can hear everything), then I use "match loudness" and set it for -19db LUFS. This is what I heard you need for youtube and other places in the Internet.
What about Rode AI-1 or Focusrite Scarlett Solo?
I only have a tablet with Windows 8.1, which, I guess, could be used for recording in this case (it is fanless).
Around here, the NT1 probably gets more votes than the NT1A - I don't have either, but the (obviously smoothed) curves show the 1A has a brighter top end. It may add "air" to some voices, but also likely enhances sibilance in others. Both are very quiet, at least per the specs. Make sure to get one of the kits with a pop filiter, or an accessory one, if you get a mic like this. Here's a comparison: https://www.dawsons.co.uk/blog/rode-nt1-vs-nt1a-what-are-the-differences

The Solo is fine for a mic like the RODE, but is a may be "gain challenged" for something like a Shure SM7b. There's a guy on the internet (Julian Krause) that has done a lot of reviews of entry to pro-sumer (maybe) interfaces and he does a good job of measuring and identifying plusses and minuses. Here's a review of his:

 
Around here, the NT1 probably gets more votes than the NT1A - I don't have either, but the (obviously smoothed) curves show the 1A has a brighter top end. It may add "air" to some voices, but also likely enhances sibilance in others. Both are very quiet, at least per the specs. Make sure to get one of the kits with a pop filiter, or an accessory one, if you get a mic like this. Here's a comparison: https://www.dawsons.co.uk/blog/rode-nt1-vs-nt1a-what-are-the-differences

The Solo is fine for a mic like the RODE, but is a may be "gain challenged" for something like a Shure SM7b. There's a guy on the internet (Julian Krause) that has done a lot of reviews of entry to pro-sumer (maybe) interfaces and he does a good job of measuring and identifying plusses and minuses. Here's a review of his:

Yes, I know him. Just yesterday I saw his video with Rode Ai-1 interface and Zoom H5 comparison.
I liked this bundle:
 

TalismanRich

Well-known member
The Rode A1-1 looks nice enough for a single microphone setup like you are using. I haven't listened to either interface, so I can't say if one is better than the other.

Something you need to research for both devices is if there are drivers available for Win 8.1, and if your tablet has enough USB power to drive the interface. Both the A1-1 and the Scarlett Solo are bus powered, there is no option for external power. Some older USB ports may not provide enough power, in which case you might need a powered hub to run it.

Are there any stores where you live that might allow you to test the various interfaces.
 
The Rode A1-1 looks nice enough for a single microphone setup like you are using. I haven't listened to either interface, so I can't say if one is better than the other.

Something you need to research for both devices is if there are drivers available for Win 8.1, and if your tablet has enough USB power to drive the interface. Both the A1-1 and the Scarlett Solo are bus powered, there is no option for external power. Some older USB ports may not provide enough power, in which case you might need a powered hub to run it.

Are there any stores where you live that might allow you to test the various interfaces.
I cannot test anything here, no mics, no interfaces, that is why everything is so bad. I have to bring it, test it, and then try to sell it if it is not suitable. I live in Ecuador. But I will figure it out! Thank you, guys, a lot!!!
 

TalismanRich

Well-known member
Here is a review/test of the AI-1. The Rode certainly sounds quieter in the noise test than a Scarlett 2i2 at 3:27.
 

TalismanRich

Well-known member
The F4 records at the same 24 bits as the H6, which is a measly -144dB noise floor. The preamps are rated at -127dBu, which are slightly better than the 120dBu H6. I don't see what 32bit float will give you. You already determined that the mic is giving you the majority of the noise in your system. That should be the first point to address.

What do you think having a -192dB theoretical noise floor is going to give you? Its like buying racing tires that can go 200mph without blowing apart, and a car that tops out at 100mph.
 

ecc83

Well-known member
Slight qualification? There are mic pre amps such as those in the Sound Devices recorders that use "dual converter" technology (no, I don't know how it works either!) that give very low noise and near infinite headroom. This is useful for portable recorders where sound levels can range massively.
But yes, for any reasonable 'studio' purpose, 24 bit operation is easily good enough. The main problem is that people seem terrified to record at what they perceive as a low level. My son sends me classical guitar clips PEAKING to -25dBfs never mind averaging 20dB! No matter. He has a quiet LDC and interface and is lucky to have a very quiet flat most of the time. I can boost to a 20dB average, peaks to around -8dB fs and still have a very decent recording.

Dave.
 
Thank you very much! As I understood, 32 bit float allows me to record at very low gain level and then boost it without boosting noise. Maybe, I understood something wrong.
 
Top