How much do speakers contribute to the noise floor?

None. As the noise is fed to the speaker in the first place or passively re-radiated from incoming soundwaves (usually at the front). It's just waffle.

I’d say - nothing. The noise is generated by the electronics as you wrote.

Joerand, if the speaker is passive it can’t really add to the noise floor in a meaningful way. It can distort and cause artefacts, but that is not normally referred to as the ‘noise floor’.

A speaker however can offer great detail resolution and lack of phase distortion when coupled to the room it’s in and this can prevent the smearing or loss of fine detail and transients, giving the impression of greater realism.

Can a bad crossover add noise? Can a better crossover lower noisefloor?

Define what you think noise is ????    A bad passive crossover will make the integration between the drive units sound uneven, distorted etc.  but not noise or noise floor in the accepted terminology meaning.      An active crossover, is another story.

If it’s passive I can only see the noise level being added effectively below the resolution of the transducers, ie random white noise from wires and passive components that exist everywhere. A sub optimal crossover is more likely to rob resolution, provide more harmful phase distortion and not be aligned optimally for the transducers driving them to cause more distortion.

Some forms of distortion are interpreted by the brain as not being correlated to the signal.  The brain may interpret these as noise (even though they are actually non-harmonic distortions).  It is to this that they may be alluding.

Only computer simulated software can see a noise floor as such. Ears can certainly hear it - whatever it is.

Thanks for the replies. I use my Ryan speakers passively. If there was any merit to the claim, my thought was that it might be due to the speakers' 4-ohm rating, as opposed to a more conventional 8-ohm speaker. Perhaps less speaker hiss results from less resistance? Ryan do however make the same claim with their standmounts which are rated at 8-ohms.

Direct Acoustics manufactures a speaker model called the Silent Speaker II. Their claim is that "The concept of “Silence” is based on the premise that speakers can make sounds that are not on the recording. These sounds include all forms of distortion; irregularities in frequency response as well as resonances created within the speaker and acoustic distortions in the radiation of sound from the cabinet. When all of these sounds are eliminated, the presence of the speaker disappears."

So yes, seems to have to do with distortion rather than the noise floor.

The only thing I can think of that resembles noise floor from a speaker, would be the cabinet resonating. This would produce mechanical noise, versus electronic noise from the amplifier. 

Yes,  but that wouldn’t be ‘noise floor’ unless it resonated irrespective of whether any audio signals were passing into it or not. Noise floor is noise which occurs irrespective of any signal. If the noise is correlated to the signal then it’s distortion.

There may be something the speaker contributes as a result of back emf .... so if a speaker oscillates after an impulse it will generate a reverse current ... probably effecting the earth to a small degree.....I think it is quite a well known phenomena. .... but I am no engineer. Also any movement and vibration of crossovers in a magnetic field will also have an effect......

Very high efficiency speakers (i.e. horn  loaded) can exacerbate the hisssssss of quiescent noise in electronics whereby the  (fixed) noise becomes a significant percentage of the  very low signal required to produce ample listening levels. But in exchange you usually get wide dynamic contrasts. It is for this reason that Naim power amps that DO have low levels of hiss-quite possibly a consequence of fairly high gain-sound quite noisy with very high efficiency speakers, but very quiet with normal efficiency ones.