An undesirable background interference or disturbance that affects image quality.
The noise is commonly characterized by the standard deviation of signal intensity in the image of a uniform object (phantom) in the absence of artifacts. The measured noise may depend on the particular phantom used due to variable effects. Noisy images appear when the signal to noise ratio is too low. There are various noise sources in any electronic system, including Johnson noise, shot noise, thermal noise.

See also Interference Artifact.

(SNR) Signal to Noise Ratio is the ratio of the amplitude of a signal to undesired noise. The larger the signal noise ratio, the easier is the detection and measurement of a signal. The sensitivity of any device is limited by the SNR. The SNR for sound is often expressed in decibel.

From ALOKA Co., Ltd.;
'A Platform for Pure Harmonic Detection
Harmonic Echo™ technology constructs images using second harmonic components, which contain far less artifacts and noise than fundamental-frequency components.
Pure Harmonic Detection is a technology to transmit distortion-free, fundamental-frequency ultrasound beams. And when the fundamental-frequency ultrasound beam is a pure sinusoidal wave, both the fundamental-frequency image and the Harmonic Echo™ image are much clearer with less noise. This technology is especially effective for obese patients and in a variety technically difficult scanning conditions.'

Increasing the frequency of the transmitted power improves the image quality of ultrasound, but the improvement in resolution results in a decreased signal to noise ratio (SNR). Higher acoustic power levels can prevent the loss in SNR, but among other reasons, ultrasound regulations limit this to avoid heating or cavitation.
Coded excitation increase the signal to noise ratio without the loss of resolution by using coded waveforms. Coded excitation allows transmitting a long wide-band pulse with more acoustic power and high penetration of the sound beam.

Speckle noise affects the quality of the ultrasound images and can mask pathology. This artifact can be reduced by using a phase insensitive technique, or by canceling the undesirable linear-phase representation. More system samplings (needs more channels and memory) or a fuzzy logic algorithm can filter out the speckle noise.