Need help with input levels

Gesine · Post by **Gesine** » March 25th, 2011, 2:05 pm

I'm going through another phase of trying to optimise my recording set-up, so that I can get back to into recording (haven't done any for months, which is sad).

I just upgrade to the version 4 of Sound Studio (Mac only). I have a Samson Q1U USB mic.

My input levels look like this - note: the mic is just lying beside me during the screenshot, but same thing happens when I record: although there is some input on the R channel, it's about half or a third less at all times than the L channel.

My questions:
1. Why does this happen (I've linked L+R so they should be in synch)?
2. Does it matter?

Here's what the Sound Studio help file says about input levels:

The Input Levels window shows the audio levels on the audio inputs specified in the preferences.This window also has controls for the input gain, which adjusts the levels before the audio is digitized. The "M" slider is the master slider, which controls the gain for the entire device, and the "L" and "R" sliders control the gain for individual channels on the device. The "Link L+R" option causes the "L" and "R" sliders to move in sync. The "Play-through" option enables a preview of the audio inputs by playing it on the audio outputs; this option is the same as the play-through checkbox in the preferences.

earthcalling · Post by **earthcalling** » March 26th, 2011, 12:06 am

Hi Gesine,

No, that doesn't look right, and it's not what my Sound Studio 4 does. What happens with the mic unplugged, i.e. using your internal microphone as the source?

Oh, and I suppose you can work out whether it matters by doing a quick recording...

David

bish · Post by **bish** » March 26th, 2011, 9:57 am

Hi Gesine,

I use Sound Studio 4 as well... from what I see (assuming the mic's picking up ambient noise etc.), the left channel looks as I'd expect, but the right seems somewhat inhibited. I agree that your settings look good on SS4, so the problem may be further back down the chain. I have no experience with the Samson Q1U, but I do have a CO1U (earlier model I think) that I just use for Skype now. This came with Samson's own driver software (Samson SoftPre) and this used to cause me no end of problems (I ended up dumping it). I don't know if you have that on your system, but it would be my first place to look. Failing that, is it the same after a re-boot, or on a different USB port? Also check you system Audio Midi set-up. All SS4 does when you link L&R is to "gang" the controls, not anything with the signal path itself... so if it's low on one side coming into SS4, it won't compensate.

Cheers
Peter

Gesine · Post by **Gesine** » March 26th, 2011, 2:34 pm

Thank you, David and bish!

I did a bit more testing.

First finding: Yes, with just the internal mic the L+R channels are beautifully in synch.
Second finding: I listened to the recording I made yesterday with the unsynched channels. Sounds okay, I think.

Samson has discontinued SoftPre, which is odd because when I first got the mic a couple of years ago, I couldn't even get a decent input level without it (tiny tiny waveform). When I upgraded my OSX to Leopard, SoftPre stopped working. Input volume is no longer an issue even without it, so I wasn’t worried. However, I tried to get it now and found that Samson has discontinued it, but I found a working version and downloaded it.

Once I'd installed (but not yet opened!) SoftPre, something must have changed in my audio settings somewhere because two things happened:

1. my L+R levels are in synch even with my Samson mic, so that's fixed
2. my input volume is again not high enough, to the extent that there's no apparent waveform when recording with the mic.

So I opened SoftPre and did what I used to do before: cranked up the volume to 42db. Now I have a decent wave, but still not loud enough, i.e. I'd have to amplify. Or I could crank up the volume to 46-48, where I get a loud enough recording, but the background noise is even louder.

Here are the three waveforms (recorded with Samson Q1U in Sound Studio 4):

Recording yesterday, channels not in synch, no PreSoft on the system:

Recording today, channels in synch, PreSoft set to 40db:

Recording today, channels in synch, PreSoft set to 46db:

Here are the three files - completely untreated, just recorded and saved as .wav:

No SoftPre, unsynched 1.72MB
SoftPre at 40db 1.46MB
SoftPre at 46db 1.02MB

I'm trying to understand what my best option is. All three recordings need noise-cleaning and some dynamic compression, which is not ideal. Should I ditch PreSoft again? Is it better to have a good input volume and more noise?

David, how does my background noise compare to yours, and also the input vol?

I'll do some more testing with noise-cleaning, but if any of you have suggestions on how to sound-edit these files to make them sound good, please don't hold back!

earthcalling · Post by **earthcalling** » March 26th, 2011, 3:19 pm

This is all sounding horribly familiar...

... in theme, if not in detail.

The hiss you're getting on all 3 recordings is from the microphone. I used to get that with my C01U. Possibly mine wasn't quite that bad, or perhaps I'm not used to hearing it now that I've changed to a Blue Yeti, but I'm pretty sure it's the same thing. For a long time I ignored it, and didn't noise-clean but just faded in and out so it didn't jar at the start and end of the recording. Having heard some of the wonderful recordings in iambik.com, I got to thinking I needed to do something. First step was to get an SE Project Studio Reflexion Filter, thinking it was ambient noise but also having a slight suspicion it was the mic itself. That did help with the quality of the vocal sound, but didn't eliminate the noise. Next step was to replace the mic with a Blue Yeti.

Now, before taking either of those steps, I was in the same position as you, having taken a break from recording and then finding problems setting up again. With me, also newly in Snow Leopard, the Mac wouldn't recognise the C01U at all. Nichts. Nada. Rien. After much hair-pulling, research in forums etc., the only solution that worked was to remove Samson SoftPre and reinstall the system software. Then I had no problem at all recording straight into Sound Studio. The noise issue wasn't new, or made any worse through removing SoftPre. The only thing that changed there was my tolerance of it.

This was recorded on the C01U (after removing SoftPre):-
http://ia600403.us.archive.org/32/items/quiettalks_prayer_1103_librivox/quiettalksonprayer_01_gordon.mp3

This was with the C01U and the Reflexion Filter:-
http://ia600403.us.archive.org/32/items/quiettalks_prayer_1103_librivox/quiettalksonprayer_04_gordon.mp3

This is on the Blue Yeti and the Reflexion filter:-
http://ia600403.us.archive.org/32/items/quiettalks_prayer_1103_librivox/quiettalksonprayer_06_gordon.mp3

All files are processed in the same way, i.e. no noise cleaning, just normalisation.

David

Gesine · Post by **Gesine** » March 27th, 2011, 8:22 am

David, that's very interesting. The sound is very good! Thanks for posting. Have you tried the Blue Yeti without the Reflexion Filter?

Why did you buy the Blue Yeti, i.e. what other mics did you consider?

Have you thought of getting a mixer/pre-amp?

Gesine · Post by **Gesine** » March 27th, 2011, 10:16 am

Also, may I ask what settings you use in SS to normalise? Do you ever use dynamic compression, or any of the other sound gizmos?

I just got a friend's mic stand, and also his Samson CO1U, to test some more.

Already two interesting things:
1. It works just with Sound Studio, without having to open SoftPre. Once SoftPre is opened, I have to crank the volume up to 40db to get the same input volume as without SoftPre. There's no noticeable sound difference between having SoftPre on or not, I think (but I can only find crappy headphones, and it's currently very noisy outside, so not very accurate testing). Interestingly, with the Q1U I have to set vol in SoftPre to at least 46 to get a good enough sound wave (but I read condenser mics are louder). Also, it won't work at all without SoftPre.
2. Using a mic stand seems to reduce the noise considerably on both mics, but I'll have to test some more.

Will report back later.

bish · Post by **bish** » March 27th, 2011, 12:28 pm

I am reluctant to normalize at 0db (as is sometimes recommended here if I remember rightly). For other uses, I normalize anywhere between -6dB and -2dB (depending on who it's for), but for LibriVox, I've standardized at -2dB. I try to avoid normalization as a rule (as it can drag the noise floor up), and try to get my peak recording level as close to -2dB as possible... it allows for a little over-run on some peaks (without distortion), and then I can just treat those small sections to bring the highs down a tad. If I've done a section, and I'm peaking at -6dB, then I've not been diligent enough in setting my input level!

I avoid compression on any dramatic parts. Once the editor has put it all together, they may want to compress the whole thing slightly, and if I've done my bits once already, they will get over-compressed and sound odd in the whole soundscape. The editor has a hard enough job evening everything out, and I like to present my stuff as dry as possible so I don't screw them up.

Now, for full chapters or stories, I may play with the smallest hint of dynamic compression. The rule is, if you can hear that it's been done, then it's been done too much

Over-compressed audio can actually be painful to listen to for extended periods (well, I think so). You have to be careful because regular mp3 encoding compression can add an element of dynamic compression to the signal. If I've set everything to sound fine in the native aiff file, then the mp3 may well sound over-compressed.

As for the other effects, I tend not to use anything for LibriVox, although for other stuff I'll play with all the tools to get what I want. I recently had to be "an android customs officer on a space station" ... I really went to town with the effects for that!

Samson SoftPre... I did some research on this, and found that it's supposedly redundant now and discontinued by Samson. The core functionality of the Mac negates it's use. However, it seems that it's such a badly written little driver that you have to use a special removal tool and a re-boot to purge it from your system... otherwise it lurks and gets in the way. So, without having it open, SoftPre can dribble all over your signal path! You need to nuke it completely! David alluded to this in an earlier message... but I think he didn't have the benefit of the removal tool, and had to use the tried and trusted "Nuke and pave" method of fixing software problems

Here's a link to the Samson removal tool... http://www.samsontech.com/main/misc.cfm?pageID=45

Good news about the stand...
Cheers
Peter

Gesine · Post by **Gesine** » March 27th, 2011, 12:46 pm

Thanks for that, Peter! Once I have a cleaner recording to present (just setting up in the b/r now - in fact sitting on bed - it's the only place in the house with soft surfaces in the house and the smallest room, so I'm guessing it's a little less echo-y) maybe you and David can help me with normalisation/compression settings.

It's great to get so much knowledgeable help!

earthcalling · Post by **earthcalling** » March 27th, 2011, 1:49 pm

Gesine,

My normalisation is set to -1.5dB, peak level. I look out for particularly high peaks, and manually adjust them down, just enough so that the overall level is boosted by normalisation. The peaks are usually just a few sharp esses, if I've recorded fairly carefully. That's it. No compression or anything.

To be honest, I didn't do much research. The Blue Yeti was well reviewed, not expensive, and has the 4 different patterns to play with, which I thought might come in handy for group recording, interviews and the like.

I have tried it without the reflexion filter, and the difference wasn't huge. But my room isn't brilliant acoustically, and can sound a but echoey. The filter helps reduce that, and focuses the sound a little. I also find it helps me keep a consistent position relative to the mic. Mainly, though, I've bought it, so I use it, even if the benefit is marginal.

David