Topic: Different L/R channel levels: is it a capsules problem?

[PaulCayard]

Hi, I'm a newbie and today I tried a stupid test on my new gears in order to improve the confidence with it.
Result of the test was unexpected, so I'm posting here hoping that some of you could help me to improve my knowledge :-)
I hope to use correct terms :-D

Gears:
- two couples of AT853 (4.7k mod) mics
- a pair of cardioid capsules
- a pair of sub-cardioid capsules
- Sony M10

(I've also a CA UBB - not used in this test - and a CA 9200 will join the party in the next weeks)

Test:
Thanks to http://www.audiocheck.net/audiofrequencysignalgenerator_sinetone.php I generated two sine tones that I play continuosly: a 10 sec 400Hz sine tone and a 10 sec 1000 Hz sine tone.
Then I put a couple of mic near to one of my cheap PC speaker, plug in to "mic" input of the Sony M10 (manual rec level to 7) and started a recording.
I tried a lot of configuration (mics/capsules) and then I've loaded all the recordings on Audacity.

Results:
1) When I record with cardioid capsules, a channel level is bigger of the other of 1 or 2 dB. Channel levels on audacity are between -21 dB and -15 dB. If I swap the capsules (left to right and viceversa), channel levels are swapped too. So I conclude that a cardioid capsule is more sensitive than the other of 1 or 2 dB.
2) When I record with sub-cardioid capsules, the result is the same but the sensitivity difference between the two subcards is 2 or 3 dB.



Question:
Is this test stupid? If no, should I worry about this sensitivity difference?

Thanks a lot for any contribution.

[beatkilla]

The capsules are not matched ,in the case of the sub cards they were only available thru AT parts center ,so they didnt hand match caps,you would have to buy many caps to get ones very close in output i think and since they are discontinued .....i think SP  sells matched card and omni caps though.

[PaulCayard]

Thanks for the response... "matched" is a new term that now I know (ok, it's not only about "sensitivity"...).
I've searched for "AT853 matched" and I've found some old discussions.

So I should boost a channel use Audacity effect "normalize" (with "normalize stereo channels independently" option checked) in post and I shouldn't be worried about it, isn't it?

[beatkilla]

I usually just raise the fader on the lower channel a bit in post production.

[bombdiggity]

Until you get to the realm of high end larger capsules the output (and overall response) likely varies significantly. 

It is luck of the draw as to how close any two are in the inexpensive unmatched realm. 

1-2 dB is pretty standard with the AT's.  If you get enough of them you can pick two that are close. 

[Gutbucket]

No need to worry about it too much.  Simply adjust the level of one channel to match the other afterwards.  If you know it's always, say 2dB higher on the left side or whatever, you can always adjust by that amount, but it's best to balance each recording by ear anyway.

The bigger problem is if they aren't matched closely enough in frequency response.  You may be able to EQ them to better match each other, but that's more difficult to do than simple level balancing.

[HealthCov Chris]

I am a bit confused by these responses.  Are you saying that you match the gain numbers on your recorder and fix level the channels later in post?  Or…level the channels on the recorder during recording using the gain options?  I have ck61's that are off a bit, but sometimes they are off by quite a bit and I am not sure what causes the change in variation.

[Gutbucket]

You can compensate during recording if your recorder provides individual control over the level of each channel, and/or you can compensate afterwards by adjusting levels on the computer in your editing software.  Either way, what really matters is how balanced it sounds.  If it doesn't sound balanced left/right, adjust it so that it does. 

Regardless of what the meters look like, an appropriate perceived left/right balance is all that really counts.  Of course you need to be careful and make sure you are only correcting the imbalance in perceived level on the recording and not compensating for an imbalance in your playback system or your own hearing.

[PaulCayard]

I usually just raise the fader on the lower channel a bit in post production.

It is luck of the draw as to how close any two are in the inexpensive unmatched realm. 
1-2 dB is pretty standard with the AT's.  If you get enough of them you can pick two that are close.

No need to worry about it too much.  Simply adjust the level of one channel to match the other afterwards.

Thank you guys for your answers! Now I know that it's a common problem "in the inexpensive unmatched realm" and how to fix it in post.

The bigger problem is if they aren't matched closely enough in frequency response.  You may be able to EQ them to better match each other, but that's more difficult to do than simple level balancing.

Just for curiosity, I tried another experiment (I'm a newbie so I imitated a test that another user described few days ago in another thread of this forum section) in order to obtain (maybe :-D ) the frequency response of 2 mics (I'm interested to compare the mics each other, not an absolute frequency response. So, no particular room... just a PC speaker...).
I generated a 10 seconds Frequency sweep (from 20Hz to 20k Hz), then I plotted (left and right channels spectrums and peaks hold) the result (that I've attached to this post).
Which conclusions can I draw from this graph?

Thanks for any contribution 

[Gutbucket]

Close enough to not be concerned about needing to compensate for frequency differences between the two.  Especially given the conditions of the test, which may be the primary contributor to the variances visible in the upper frequency range of that graph, rather than the response of the mics themselves.

Also try a perceptual test.  Listening to one at a time, switching back and forth with levels matched, do the two sound the same to you?

[PaulCayard]

I'm very busy in these days and I haven't done any other test... I'm very sorry, I hope to answer during easter holiday :-)

[PaulCayard]

Also try a perceptual test.  Listening to one at a time, switching back and forth with levels matched, do the two sound the same to you?

I've tried a lot of tests this afternoon... Perceptual test is negative: no difference between a "Normalize" track and a "Normalize stereo channels independently"!

I want to thank you guys for all the help :-)