sorry to start a new thread but the old one was locked:
http://talk.maemo.org/showthread.php?t=21387

and it appears the n810 had a similar problem
http://talk.maemo.org/showthread.php?t=26272

but I don't think the answer was satisfactorily laid to rest, so I want to post an update on some experiments I've carried out.

Firstly, the phenomenon. when connected to certain *active* devices such as noise cancelling headphones, active loudspeakers or FM transmitters there's a significant background hiss when media player runs; if media player pauses after short while the hiss stops. different people experience different levels of the problem. The noise is generally not affected by the volume control other than totally muting the sound.


in my personal experience,
* With noise cancelling phones (Sony MDR-NC7) I only get the noise when NC is turned on. The noise I hear in my isn't plain noise, it sounds like quantisation noise mixed with white noise, a distorted quiet version of the main sound track, and changing the N800's volume control shows it's a steady noise (other than when muting sound).
* With an FM transmitter, it's more a plain "white" noise phenomenon in the car radio
* I don't get it with some simple Logitech active speakers
* I don't get it with some 32 ohm Sennheiser portable/walkman headphones
* It's not affected by backlight
* It's not affected by fiddling with the FM radio, e.g. turning it on, changing channel, changing FM radio's own volume control

I wondered if it was simply that active devices have higher input impedance preventing electrical noise being suppressed, so as an experiment I've connected various different resistor values across the audio ouput. I found that higher value resistors over 100 ohms had little effect, but 32 ohm resistors made a noticeable difference to the higher frequency noise, and had a slightly lesser effect on the sound levels of the media being played. I was disappointed it didn't help as much as I wanted, otherwise I would make a short flying lead with this simple attenuator in so as to allow me to use my NC phones with it.

The fact that I can hear a heavily distorted version with white noise implies to me that there's a flaw in the analogue circuitry. Most digital-analogue converters have a low-pass filter with analogue amplifiers afterwards - there will also be a mixer to combine the FM radio output in the N800. I am wondering whether the low pass filter and mixer/amp leaks the unfiltered DAC output and its clock sources into the final amp stages in such a way it bypasses the volume control. Normal headphones and speakers don't reproduce it as they are naturally low pass filters themselves, but active devices are significantly affected by these high frequency sources.

Is there someone with an N800 and an oscilloscope who can look at the audio output and confirm my theory? For bonus points, construct a simple low pass filter with a corner frequency of about 15 kHz and see if the affect can be significantly mitigated.

If my theory is correct, I will make myself an inline passive low-pass filter. Without the noise the N800's playback is pretty good, so I would be able to make better use of it as a media player!

BTW, I don't have many sound effects enabled; I think that comments about have screen tap clicks turned on are a red herring, it's merely because it wakes up the audio output, which I image is turned off to save power.


sorry if people have done this to death before, but I never found a satisfactory explanation and particularly I never came across a solution.