Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Recording audio with smartphone mic input. White noise

Status
Not open for further replies.

frilance

Member level 1
Member level 1
Joined
Aug 16, 2012
Messages
37
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Activity points
1,703
Hi!

I decided to solder a 3.5mm 4-pol phone connector to a 6.35 mm one in order to record electrical instruments (guitar, keyboard... maybe mixing desk too) directly in my smartphone.

You can see the first version of this "adapter" in the picture attached. DSC_0016.JPG

I used it together with two apps (MyVoiceFree and WavePad free) and recorded the signal that came through the phone's mic input... but the result was not completely satisfactory. There is plenty of white noise whose origin is still unknown to me.

You can hear an example here:

**broken link removed**

I used a 32 bits conversion with a 44 KHz Fs... so I don't it comes from the phone's ADC... could it be related to the quality of the cables used? any other thoughts?


Thanks a lot in advance.

Best regards
 

Your adapter has a direct short circuit between the left and right channels on the 6.35mm plug then unshielded wires (they pickup mains hum and other interference) to a 3.5mm plug on the mic input tip.
The direct short circuit can destroy your signal source. A mixer circuit should be used which can be two resistors. Shielded audio cables are always used for audio.

I hear your recording at a low level. When I turn up my volume control then I hear background hiss. It will have less hiss if you increase the recording level.
 

Phone microphone is designed for voice- with a bandwidth of 4000Hz - not really music quality. They also have horrible response curves. 32 bit conversion will not help if you use a poor quality mike. Studio quality mikes are expensive. Some say damn expensive.

What I heard (your example sound) is good. But try to record lows and highs so that you will get the idea of the mike limitations.
 

Phone microphone is designed for voice- with a bandwidth of 4000Hz - not really music quality. They also have horrible response curves.
Yes, for a 50 years old and older carbon microphone in a wired phone. But cell phones use an electret mic that has a frequency response as flat as a ruler to 15kHz or more.

When I designed and built my Mod4 FM transmitter I used the tiny electret mic from my daughter's old cell phone and its frequency response is as good as an FM radio station (15kHz).
Here is the frequency response of a typical electret mic:
 

Attachments

  • Electret microphone.PNG
    Electret microphone.PNG
    16.8 KB · Views: 201

But cell phones use an electret mic that has a frequency response as flat as a ruler to 15kHz or more.

The graph shows "flat as a ruler" only upto 5000 Hz. Why they plot the graph from -50db to +20db when the graph stays within +/- 10db?

A 3db change is equal to a response change of signal output by a factor of 2. Most of these graphs use signal power and I guess a 6db change will mean a signal output change by a factor of 2.

The wriggles after 4000 Hz are due to resonance. Larger units have better low frequency response where as smaller ones have better high frequency response.
 

Most microphones have a frequency response that goes up and down a lot at different frequencies. Look here:
 

Attachments

  • microphone comparison.png
    microphone comparison.png
    67.4 KB · Views: 143

I agree. The mike with the curvaceous response curve is certainly more expensive. Don't ask me why.
 

I agree. The mike with the curvaceous response curve is certainly more expensive. Don't ask me why.
The expensive Shure SM58 dynamic microphone was designed for singing and speaking vocals:
1) It rolls off low frequencies because it is directional and up close it produces the "proximity effect' which boosts low frequencies and makes a voice sound boomy.
2) It has a peak in the upper midrange to give "presence" so it sounds like the singer or talker is very close to you.
 

Thanks a lot guys for the nice input about microphones' response.

But I'm a bit puzzled. From my naive point of view I think I should not be concerned about that.

I mean, I'm trying to record electric signals directly from guitar/guitar amp / digital keyboard headphones output to my smartphone via the minijack mic input, so no transducer (with its intrinsic response) involved.

That said, maybe what some of you meant and I did not get is that the electrical freq response of the phone amplifier right after the minijack input is optimized for voice frequencies.

Well, in order to discard some potential causes of this white noise I improved the cable by: adding the stereo-to-mono passive adding circuit (10K R, 10K L and 100 K to ground), used a shielded cable and added a voltage divider with a potentiometer to attenuate the output signal from the device in the cable itself and not by reducing the volume of the guitar/keyboard.

Nothing really helped. Results were similar as in the audio attached.

Then I decided to remove more components of my system, and basically record the audio when having only a short at the input (I shorted the 4-pol mini jack GND to MIC)... and I still had the noise... so I'm afraid that the smartphone itself is limiting me... either the front-end electronics or the ADCs... does this conclusion sound right to you? Maybe I'm going into so much trouble when I could get way better results with a raspberry Pi and the additional ADC... and have way more flexibility.
 

Electret mikes need a voltage source and the special connector for the smartphone has a power source. You need to check the connections properly. For the mike has two pins - one marked + and the other marked (obviously) -, the audio travels on the same cable marked + via a small cap. See the connection diagram on https://en.wikipedia.org/wiki/Electret_microphone and do a connection accordingly. Usually the resistor is 1K-10K and the capacitor is 1uF.
 

As I said in my first reply, your recording level is too low so of course when I turn up my volume so it is loud enough then the hiss in your recording is also turned up.
I do not know why the recording level is too low. Is the headphones output level from the amplifier loud enough for headphones or is it also too low? Then turn up the volume on the amplifier.
Is your smart phone producing the hiss? If you record "silence" then is there hiss on the recording?
 

Is your smart phone producing the hiss? If you record "silence" then is there hiss on the recording?

Exactly, that's what I meant Audioguru. I tried to check if the smartphone was the responsible, recording the "silence" by shorting the 3.5 minijack input (between MIC and Ground channels of the CTIA/AHJ standard) and the hiss was still there...
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top