I came across this: http://www.dapreview.net/content.php?article.133 while looking for solutions to some problems I have with my new ipod 5g i.e. "normalization", and ipods HORRIBLE EQ.
The article seemed to have a combination of factors that seemed a little different than anything else I've read. Any thoughts? I want to use replaygain/mp3gain, or something to deal with volume consistancy, but NOT soundcheck on itunes. Has anyone else come across similar eq "tweaks" for ipods?

Everybody bashes SoundCheck, but I'm not certain why. Aero has said that in his testing of foo_pod, the values produced by SoundCheck were not usually very much different than the values produced by foobar's ReplayGain. In my own testing, SoundCheck has worked well enough for me to not notice any extraordinary volume shifts, or at least, not any more so than using RG values.

Using the volume adjustment like he recommends in the article above has caused excessive clipping in my own experience. It's not really all that safe to screw around with those volume controls too much.

As far as EQ, well, there's really no fixing that. The iPod has a 5 band EQ and that's all there is to it, really.

so there's no validity to his claims that mp3gain makes the ipod equlaizer more usable? I find the distortion it adds horrible, but he seems to have eliminated it in his testing.

I've read the whole article and it seems reasonable to me (except for the last part about "boosting" the volume of the MP3gained files on which I cannot comment as I own no iPod; also, he misrepresents how MP3Gain alters the MP3 data, but that doesn't really matter as it is reversible as he writes). Reducing the volume to 89db should give enough headroom to prevent clipping, provided this is really the problem with the EQ on the iPod, which I don't know.
In another thread, someone (Otto42 IIRC?) wrote that SoundCheck does not prevent clipping as it applies the volume changes after decoding the MP3, so it works well for adjusting the volume but not for clipping prevention. For this you'd indeed need to use MP3Gain. But maybe it's better to use a higher value like e.g. 93db to avoid having to use the "volume boost" hack.

I can confirm that this works. After mp3gaining all my mp3s and AACs there is no audible distortion with neither my 2nd gen mini nor my 5th gen ipod.

However, with highly dynamic music there might still be clipping. But I didn't notice any.

Of course Apple should rather fix their equalizer. I have read the wolfson datasheets and the solution would be simple: use the wolfson digital attenuation function (which takes effect directly before the eq for a reason) to attenuate the signal by the highest boost in any band. Then equalise and finally offset for the attenuation via the analog amplifier..

It's kinda lame that Apple still doesn't get such a super-simple thing right on a multi-hundred dollar product after five generations...

SoundCheck doesn't prevent clipping, but then neither does MP3Gain. MP3Gain will just tell you when it will be clipping, and you can lower the volume to not cause it to clip. If you use MP3Gain wrong, you can clip the hell out of a file.

SoundCheck, though, can't cause clipping either, at least on the iPod. It's just adjusting the power level in a more analog manner, I thought. That's why the SoundCheck info is stored in the iPod's database as microwatts, it's not a digital volume boost on the iPod. The per-song volume level thingy *is* digital, and causes clipping very easily, in my experience.

This is all on my 3G iPod, other versions may be different.

It seems like the problem might be that when the db value going into the equalizer is too hot for a given track, it will cause clipping and problems as it comes out of the EQ due to the EQ's inability to deal with "hotter" tracks. I've noticed that with some files the EQ isn't that bad, but with others it distorts quite noticably. The reason Sound Check doesn't help is because it comes after the EQ in the chain of command, and the EQ has already done its damage.

I'll try using MP3gain to knock down a problem sample to 89db and see if it works for me. AS long as you aren't boosting the MP3's digitally, there shouldn't be any clipping anyway. In this case, we want to decrease the level.

I can see why you would want to boost the player's output power to compensate for low volume mp3's but it doesn't seem like it would be warrented (or a good idea to actually do).

Am I making sense?