This is kind of the problem in gauging whether it's a hardware issue, I'd neglected the N900 a bit in terms of audio listening before I upgraded to PR1.3 and got more into using QSpot and gPodder - I'd written off Media Player as a viable player for me as it stands a while back, and gone back to the iPhone as my primary music phone. So I can only say that I believe I wasn't hearing anything as wrong before PR1.3.

However, your point about the headset got me thinking so I ran the N900 test once again with a jacked WH-205 connected, and as a comparison the iPhone 4G with a Shure MPA. Distortion was actually lower (although still way over most median MP3 players), and the rest of the figures weren't that bad - it still trails the iPhone 4G with the exception of dynamic range, in which the N900 was level pegging - and whatever you're doing to reduce distortion, it isn't working over a stock setup.

As for the Audioboost tonal improvements, I guess it's a matter of taste [audiosnob] or lack thereof [/audiosnob]...


(32o = 32 ohm load)

Cliff notes for anyone else reading this:

1) A bit of a shocker - using a headSET cabling configuration on the N900 in PR1.3 actually appears to give better audio quality than directly plugging in headPHONEs (regardless of whether you use Audioboost or not). Obviously however this is dependent on the head/earset you are plugging in - a headset with a relatively crappy earphone attached to it will sound crappy no matter what. I used a WH-205 converted with a jack socket (well - if I hadn't done this, it would have been impossible to RMAA the WH-205's effect in any case) vs the headphone socket for this particular test. However, the resulting audio is still worse than leading PMP/smartphone - particularly for distortion. If anything though, this points out how much us N900'ers need a WORKING jacked remote.

2) Audioboost has no effect on the volume (from a WH205 at least)


I ran all tests at a slightly higher then typical listening volume (as opposed to full volume, because this test is more realistic to what you'll actually hear) - so I'm not posting anything other than the frequency response, as the rest of the measurements would not have any relative relevance - enough people use RMAA to generate worthless data in any case, which are then quoted by people who have no idea how the measurements were carried out or how to interpret the graphs. Relative audio levels were matched using Digicheck.