SSIM is a quality assessment method that compares the structural similarity between the original image and the compressed image. <\/p>\n\n\n\n
It provides a score that ranges from:<\/p>\n\n\n\n
- \n
- 1<\/strong> – representing identical pixel structure – to –<\/li>\n\n\n\n
- 0<\/strong> – representing low similarity and poor image quality.<\/li>\n<\/ul>\n\n\n\n
\n ![\"Original](\"https:\/\/shortpixel.com/blog\/wp-content\/uploads\/2023\/02\/picture_original-1024x683.jpg\" "\\"\\"")
<\/a>Original picture<\/figcaption><\/figure>\n<\/figure>\n\n\n\n \n ![\"\"](\"https:\/\/shortpixel.com/blog\/wp-content\/uploads\/2023\/02\/picture_quality_low-1024x683.jpg\" "\\"\\"")
<\/a>Highly compressed picture – low picture quality<\/figcaption><\/figure>\n<\/figure>\n\n\n\n Example 1: High compression level = low visual quality<\/strong><\/p>\n\n\n\n
Above, you’ll find the original image, while the same image is highly compressed underneath it.<\/p>\n\n\n\n
The file size has been reduced by 81.31%<\/strong> from 273 KB<\/strong> to 51 KB<\/strong>, but the visual quality has been degraded. <\/p>\n\n\n\n
The SSIM value of 0.8924<\/strong> confirms that the quality is not great.<\/p>\n\n\n\n
\n ![\"\"](\"https:\/\/shortpixel.com/blog\/wp-content\/uploads\/2023\/02\/picture_quality_82-1024x683.jpg\" "\\"\\"")
<\/a>Slightly compressed picture – high picture quality<\/figcaption><\/figure>\n<\/figure>\n\n\n\n Example 2: Low compression level = high visual quality<\/strong><\/p>\n\n\n\n
When compressing images by lowering the quality factor to a certain value – a technique used by most image optimization services and plugins – it can be difficult to detect any changes in our test image. <\/p>\n\n\n\n
However, the file size was reduced by only 10.21%<\/strong> to 245 KB<\/strong>. <\/p>\n\n\n\n
The SSIM value for this compression level is 0.9907<\/strong>, which is very good, but we can get even better results in terms of file size.<\/p>\n\n\n\n
\n ![\"Optimally](\"https:\/\/shortpixel.com/blog\/wp-content\/uploads\/2023\/02\/picture_quality_glossy-1024x683.jpg\" "\\"\\"")
<\/a>Optimally compressed image<\/figcaption><\/figure>\n<\/figure>\n\n\n\n Example 3: SmartCompress = optimal visual quality<\/strong><\/p>\n\n\n\n
Using ShortPixel’s SmartCompress with the Glossy optimization level<\/a> results in a better-compressed version of the picture, indistinguishable from the original to the naked eye, and reduces the file size by 26.73%<\/strong> to 200 KB<\/strong>. <\/p>\n\n\n\n
The SSIM value of 0.9909<\/strong> means that the quality of the image is almost identical to the original image, and the file size is smaller compared to the example above.<\/p>\n\n\n\n
The picture optimized by ShortPixel isn’t only much smaller<\/strong> than the picture compressed with a quality factor of 82, but also has a better SSIM factor, which means that it also looks better<\/strong>!<\/p>\n\n\n\n
One of the key advantages of SSIM is that it compares images at multiple weighted resolutions and performs scaling in linear-light RGB. This means that it’s sensitive to distortions of different sizes and blends colors correctly to detect errors such as chroma subsampling.<\/p>\n\n\n\n
In addition, the algorithm uses the Lab* color space for SSIM calculation, allowing better measurement of brightness and color compared to metrics based on the average of RGB channels. Furthermore, supports alpha channels and images with color profiles, and also uses multi-core CPUs to ensure fast processing.<\/p>\n\n\n\n
How does SmartCompress work?<\/h2>\n\n\n\n
As the images above illustrate, the file size of an image has a direct impact on its quality. To achieve the best results, it is important to find the perfect balance between file size and image quality<\/strong>.<\/p>\n\n\n\n
By using SmartCompress, each image is automatically optimized multiple times<\/strong> until the perfect, smallest possible image is achieved. That means you don’t have to worry about under- or over-optimized images. You can rest assured that your images will look great and load quickly while keeping file size to a minimum.<\/p>\n\n\n\n
Here you can see how ShortPixel’s SmartCompress compares to the optimization methods of other image compression tools and software:<\/p>\n\n\n\n
\n\n\n![\"\"](\"https:\/\/shortpixel.com/blog\/wp-content\/uploads\/2024\/03\/pexels-photo-417273-1024x683.jpeg\" "\\"\\"")
<\/a>Original file<\/a> – 978px<\/figcaption><\/figure>\n<\/div>\n\n\n\n \n![\"\"](\"https:\/\/shortpixel.com/blog\/wp-content\/uploads\/2023\/02\/pexels-pixabay-417273-shortpixel-1024x683.jpg\" "\\"\\"")
<\/a>ShortPixel – Glossy compression – 213KB – Reduced by 78% <\/strong><\/figcaption><\/figure>\n<\/div>\n<\/div>\n<\/div>\n\n\n\n \n\n\n![\"\"](\"https:\/\/shortpixel.com/blog\/wp-content\/uploads\/2023\/02\/pexels-pixabay-417273-adobe-photoshop-65-1024x683.jpg\" "\\"\\"")
<\/a>Adobe Photoshop (Save for web at 65% quality) – 309KB – Reduced by 68%<\/figcaption><\/figure>\n<\/div>\n\n\n\n \n![\"\"](\"https:\/\/shortpixel.com/blog\/wp-content\/uploads\/2023\/02\/pexels-pixabay-417273-tinypng-1024x683.jpg\" "\\"\\"")
<\/a>TinyPNG – 353KB – Reduced by 63%<\/figcaption><\/figure>\n<\/div>\n<\/div>\n<\/div>\n\n\n\n \n\n![\"\"](\"https:\/\/shortpixel.com/blog\/wp-content\/uploads\/2023\/02\/pexels-pixabay-417273-resmushit-1024x683.jpg\" "\\"\\"")
<\/a>reSmush.it – 327KB – Reduced by 66%<\/figcaption><\/figure>\n<\/div>\n\n\n\n \n![\"\"](\"https:\/\/shortpixel.com/blog\/wp-content\/uploads\/2023\/02\/pexels-pixabay-417273-kraken-io-1024x683.jpg\" "\\"\\"")
<\/a>Kraken.io – 327KB – Reduced by 66%<\/figcaption><\/figure>\n<\/div>\n<\/div>\n\n\n\n As demonstrated, ShortPixel’s SmartCompress is the clear winner by significantly reducing the file size by 78%<\/strong>, from 978KB to 213KB without even using Lossy compression<\/a> <\/strong>(the most aggressive level), maintaining high quality at a much smaller file size.<\/p>\n\n\n\n
\nTry SmartCompress!<\/h3>\n\n\n\n
Want to improve the performance and visual appeal of your website through advanced image compression and optimization? Give SmartCompress a try for free and see the results for yourself.<\/p>\n\n\n\n
- 0<\/strong> – representing low similarity and poor image quality.<\/li>\n<\/ul>\n\n\n\n