Anderton
Maybe the speed at which the compression was performed makes a difference? (You know, the "faster/slower compression" option a lot of encoders have). It does with the LAME encoder - if you export a file with "quality = better" or "quality = faster," they won't null even though the kbps is the same. Also maybe a high or lowpass filter is enabled behind the scenes.
 
Encoders may use the same algorithms, but that doesn't mean the implementation is identical.

I suspect you're right. An on-the-fly encoder might well employ a lower quality setting for the sake of efficiency. But wouldn't that kind of defeat the "hear what we're taking out" feature? (Not that I place much stock in it to begin with; after all, it's only playing us the stuff we're not supposed to hear anyway.)

I suspect the Cake limiter is amplifying the MP3 artifacts in order to make them easier to hear.  That would not necessarily be a bad thing, but if that is happening, it should be explained so that people will understand that the fidelity loss is exaggerated.

Its not exaggerated. As Jon explained above its showing the encoding behavior at the default MP3 encoding settings.

Noel Borthwick [Cakewalk]
Its not exaggerated. As Jon explained above its showing the encoding behavior at the default MP3 encoding settings.

I see.  It does seem a bit misleading.  I'm not saying that the Cakewalk implementation is misleading by design.  I'm just saying the effect is a bit stronger than what people would normally experience with MP3 compression.  I don't suppose you would consider adding the faster/slower option.  (Maybe that would do more harm than good, if it eats too much CPU.)

cparmerlee
...the effect is a bit stronger than what people would normally experience with MP3 compression.

You're right about that. It's not realistic, but it wouldn't necessarily be more realistic even with higher quality settings.
 
Not unless the option was given to exactly match your own settings - assuming that a) you use the same settings all the time, that b) you actually know what those settings are, and c) that you use the LAME encoder yourself.
 
Most of LAME's commandline options are unique to LAME. I don't use LAME. Consequently, even if CW added configuration options I'd not be able to match them to the encoder that I use.
 
The lesson: don't put much stock in anyone's artifact-revealer tool. Although it has some educational value, it yields very little in the way of actionable information, regardless of implementation.

bitflipper
The lesson: don't put much stock in anyone's artifact-revealer tool. Although it has some educational value, it yields very little in the way of actionable information, regardless of implementation.

 
Some material handles data omission better than others; listening to the artifacts can help determine what you can get away with if minimum memory/fastest streaming is a priority.