The picture quality has been improved! LR enhancement details_ Lightroom turns on AI photo detail enhancement: picture clarity increases by 30%

Use super resolution , The image quality has been improved , Giving pictures for nothing can also save , It can take 100 million pixels , Do you ？

​​Adobe In the latest version of Lightroom CC, as well as Lightroom Classic CC The function of super-resolution image is updated in , To pass through AI Technology to double the resolution .

Simply speaking , The image quality has been improved .

Chapter 1 How to use ？

---

First , You need to make sure you use Lightroom It has been updated to the latest version .

As we enter Lightroom Classic In the gallery , Right click the photo you want to enhance , choice ” enhance “ after , What pops up is this Lr Updated super-resolution function interface .

Right-click menu , choice “ enhance ”

 

Enhanced interface

You will notice that the super-resolution function is used , The software will automatically save another resolution double The new DNG（RAW A format of a file ） Images , At the same time, the file size will also soar to the size of the original image 4 times . For example, a 82MB The size of RAW file , Generating 100 million pixels DNG after , The amount of data is outrageous 523MB, At the same time, the loading time of images will also be doubled . So before using this function , Please ensure that there is enough hard disk space to support you to complete the whole process .

The subsequent process is consistent with the usual map revision process . in addition , You can also use noise reduction 、 Sharpen and other tools to further optimize your picture . that , How about the final imaging effect ？

Chapter 2 Has the picture quality improved ？

---

In a word ,yes and no.

In order to make the result more obvious , We used a 2018 Nikon D700（1200 Megapixel full frame camera ） Hainan Airlines B787-8 To simulate large-scale cutting .

We decided to start with the resolution , Confirm the difference before and after using this function .

Export two photos with the highest image quality after using the same parameters to repair the picture . The left side is the exported image of the original resolution , On the right is the exported image after using the super-resolution function .

 ​

The number of pixels of the original image after cutting is about 800 Mega pixels . After using the super-resolution function , The number of pixels is amazing 3846 Mega pixels . As can be seen from this picture , The enhanced image font is not as empty as the original image , At the same time, the details of the fuselage have also been improved .

On the other hand , Upload pictures to JetPhotos When waiting for the platform , It is often necessary to compress pictures , The sharpness of the exported image also determines that your image is Accept Or rejected soft. We set the long side 1920 Export two images , And enlarge the exported image to the same level .

 

 

The enhanced picture is in “BOEING 787-8” The words of are more clearly displayed . Because the resolution is consistent , Other details are not much different , But it is also enough to show the sharpness advantage brought by higher pixel compression . Simply speaking ,“ Young man , The picture you took is very high-definition ”.

Chapter 3 There is something else I want to say

---

actually , Use AI Computing to improve image resolution is not Adobe Pioneering ,2015 year IEEE Scientific researchers of have introduced a high pixel computing method , A few years ago, some developers launched based on this technology “Waifu2x” Algorithm , To compensate for the pixel loss of the animation image after compression, so as to achieve the effect similar to the original image （ The finished product is quite suitable for wallpaper ）. But only animation pictures with this algorithm have the best effect , The effect of camera photos is not satisfactory .

Of course, there are also hardware to improve the definition . Sony camera has a common name “ Shake “ Multi pixel conversion shooting , Read four photos through the movement of the anti shake component , Then use professional software on the computer to synthesize , This synthesis method is direct color stacking . But because the four shooting angles need to be exactly the same , The object needs to remain completely stationary , The anti shake component participates in shaking but does not participate in anti shake , Basically ruled out the possibility of holding a piece .

So can you save all kinds of free maps ？ Here is the "Yes and no" Medium no 了 . Maybe a little noise is ok , There is no problem with large-scale cutting of pictures . however , Your original image itself must be focused , No matter how large the image of virtual focus is, it will not achieve the effect of counting the passengers on the plane .

What about 100 million pixels ？ We made an attempt , take 4200 The original image of ten thousand pixels uses super-resolution , After cutting and coloring , Finally got a 9838 Mega pixels “ Hassu boutique ”. Interested partners can get all the original pictures used in this article through the link below , And this one " Hassu boutique ".

---