Freesound Forums

Some time ago i made a portable set for field recording. 2xOktava MK-012 + MixPre + EDIROL R-09HR. I always set R-09HR to volume 40 (default) and use MixPre Tape Out, but as so Oktava mics are rather quiet i need to keep MixPre potentiometers somewhere between 12 and 3 o'clock. But at this levels i begin to hear MixPre self noise. In this setup the MixPre 1kHz noise signal gives me -20 at the R-09HR meters. Despite the fact that this setup is recommended by SoundDevices its seems not perfect.
Also i experiment with MixPre XLR out to stereo mini jack connector. It outputs louder signal and i got -12 at the R-09HR meter. Its better, but this setup appears bulky in some run&gun; situations.
I also i try to connect a MixPre Tape Out to R-09HR Mic In and i got a -6 R-09HR meters from the MixPre 1kHz noise signal. Not too loud as for mic in.
Next i try lower the input volume to get my lovely -12 at the R-09HR meter and it sounds pretty nice.
I know that usually it is bad idea to use mic-in instead of line-in I decide make some tests and got pretty interesting results. Also i remember R-09HR uses Wolfson WM8786 A/D chip and it was some discussion about its specific line and mic inputs, but can't find it now.

Here is how i do the tests. first 5 seconds is MixPre 1kHz noise signal, then every 5 seconds i turn the MixPre potentiometer to the one mark more and more. Then I repeat this for various recording setups and volume settings. And at last i normalize all recorded files to level 0 to see things more clear. Additionally i made 2 records with Echo Audiofire 4 sound card. Spectrogram images generated in Izotope RX.
From this tests:
From this tests it looks that EDIROL line in and mic in are almost equal
EDIROL line in appears way noisy after normalization boost compare to same level of Echo Audiofire 4.
There is no any visible quality difference in MixPre XLR out and Tape Out
http://dl.dropbox.com/s/b25a6il53quktcb/123456-raw.jpg
http://dl.dropbox.com/s/kod7frhnnbekmv0/123456-norm.jpg

I'm not too technical sound geek but I got 2 fundamental questions here:
1. Is it safe in this situation to use mic-in with lower recorder volume or it can cause other kind of distortion or dynamic range loss which can't be seen in this test?
2. Is it safe to lower recorder input volume without quality and dynamic range loss?

here is also some specs from product manuals:

EDIROL Input Impedance Mic input:
30 k ohms
Line input: 15 k ohms

MixPre outputs:
XLR-Line: active-balanced, 120 ohm output impedance
Tape (3.5 mm) -10 dB Unbalanced, 2.1k ohm output impedance
http://dl.dropbox.com/s/ynzz1mztujtgnvt/mixpre-r-09hr.jpg

p.s. i done one more test: MixPre Tape Out to EDIROL R-09HR line in with volume boosted to 55 to got -12 levels meter. And after normalization there is way more noise than mic-in with same -12 levels meter! will post images later...

I think the main noise problem is simple.

The mics should be in the xlr input and powering from premix, but i guess you do already. The self-noise of your mics are like mine. So that is okey. The premix is a great mixer with low self-noise.

The the problem is the recorder i think. Try to find someone with a recorder with a balanced xlr inputs. So you can use the balances xlr output of the premix.

I use this gear chain: Rode NT4 > Sound Devices 302 >Tascam dr-100 Mk2.

In the beginning i did use a unballanced recorder (Sony pcm d50) and the tape out of the mixer. That was more noisy.

For results search , Tascam 302 klankbeeld field-recording
on freesound to listen to my Results.

Success

i tested MixPre XLR to balanced inputs in Echo Audiofire - yep, the spectre is better than unbalanced and it has less micro interference lines but those inputs in this sound card are too hot and can't be dimmed (i got about -6 or 0, depending of sound card input setting, in meters from Mixpre test signal)
few hours ago with same method i compared mics self noise vs Mixpre self noise vs edirol line/mic input self noise, and must say mics noise in quiet night room under the blankets and pillows about 2 times louder than Mixpre itself and the mixpre louder than recorder input itself
also looks like the mic-in self noise in edirol is a little less noisy than line in. but in real world i can't see the spectre difference between edirol line-in with mixpre potentiometer 3 o'clock vs edirol mic-in with mixpre potentiometer 12 o'clock. still need to know is it technically safe to lower the volume on edirol and use its mic-in.

Recorder specs say

Normal!
Mic in -33dBu
Line +2 dBu
I think using line is the best.

But
1 Pre set output mixer limit on +4 dB
2 sent 0dB testsignal from mixer to recorder
3 level mixer input on -10dB

Thats good and safe

So finally i tested R-09HR with RMAA (RightMark Audio Analyzer test). Tested with ECHO Audiofire 4 sound card and some tests with Oktava MK-012 mic. During playback test files Audiofire 4 was set to +4db output setting with adjusted output volume.
The results:
1. in general LINE IN is LESS noisier than MIC IN, but the difference is minimal.
2. Turning R-09HR input volume 15 clicks up and down from default mark 39 adds very tiny amount of self noise but has no any effect to spectral response.
3. Turning Mix Pre potentiometers even to 100% volume and connecting it to R-09HR LINE IN produces LESS noise than Turning Mix Pre potentiometers about 70% volume and connecting it to R-09HR MIC IN.
4. R-09HR has very nice headphones output spectral response, almost flat.
5. R-09HR Input self noise and dynamic range depend of source signal. Professional soundcards and mixers with strong +4db level signal can allow about 96db of dynamic range, and consumer soundcards with source -10db level signal allow about 93db of dynamic range. Its nice result for portable recorder, but still far from best professional desktop soundcards with 110db-130 of dynamic range.

The test charts are here:
https://www.dropbox.com/sh/nwn0y12np95dn1e/AACi-E7OHYN4BJaN_o39TIB7a?dl=0

and one more test measured with MixPre test tone signal:
One click of input volume button in R-09HR equal 0.5db
R-09HR vol. setting 39 corresponds to -20db input
R-09HR vol. setting 43 corresponds to -18db input