Saving a WebP file with lossy compression requires you to set a quality value, and the quality value determines the file size and image quality of the saved WebP file.Lowering the value reduces the file size but the image quality also gets poorer.
You need to choose the best quality value considering both the file size and image quality.It depends on the content of the image, but a value between 50-80 is moderate in most cases.
The reason why a fixed value cannot be suggested is that human eyes have different sensitivities to different types of objects.For example, humans can detect minor changes in a human face or skin color, and therefore the quality value for an image containing such things needs to be quite high, at least 70-90.On the contrary, if an image contains computer-generated graphics or rapidly moving objects, the quality value of about 50-60 may be good enough.
Click the following link to view WebP images containing much or little movement, and compare them to each other.
The following samples are WebP images saved with several different quality values.
Lossy compression of WebP
Q 90 (148KB)
Q 80 (96KB)
Q 70 (77KB)
Q 50 (60KB)
Q 30 (42KB)
Q 10 (24KB)
Hover a mouse cursor over a sample to enlarge it; the samples with a quality value of more than 70 have no noticeable deterioration in the image quality.The sample image with Q 50 shows the objects' edges slightly rough.The details of the objects in the sample image with Q 10 are severely damaged.
Animated WebP with little movement
Q 90 (670KB)
Q 80 (429KB)
Q 70 (340KB)
Q 50 (280KB)
Q 30 (217KB)
Q 10 (154KB)
In this example, it is difficult to notice a difference up to quality Q80, a slight noise phenomenon starts to occur from the case of Q70, and deterioration of image quality begins to be noticeable in the case of Q50.For Q30, the deterioration in image quality begins to be noticeable.
It is hard to sense the quality deterioration in a WebP image with a low quality value if the image contains much movement as the samples above do.The quality deterioration becomes slightly noticeable only when the quality value is 50 or less.
As shown by the samples above, you may consider the content of an image and adjust the quality value to achieve both the small file size and high image quality at the same time.