Today I had a problem posting my edit to the playraw category as I usually reduce the image size to 1080 for export when being used as a playraw post of my edit. However, all the detail was softened by doing this. So then I posted a full size JPG and the file size was huge. So I then tried a full size JPG but quality set to 60% and this seemed a reasonable compromise retaining detail and giving reduced file size. I am just wondering what suggestions others would have for posting images to playraw so that file size is reasonable and detail is maintained. Thanks in advance.
Someone somewhere told me that jpg quality visibly suffers below 93%, so that is what I use. I have no idea whether that was right, but I was studying a resource that I thought was reliable and trustworthy.
Yes I normally use the default from the program, but the file size was too big for the playraw category. It seemed lowering the quality to 60% had minimal impact for internet screen viewing, but I would never use 60% for other uses.
I usually use 3000px * on the long side, at 60% jpeg quality.
The jpeg compression is noticeable if you zoom in, but to me the overall quality/clarity is better, at about the same file size as a 1080px at 90% jpeg.
I’ve always been slightly puzzled that many use low jpeg compression at the expense of pixels when aiming for a certain file size. But there may well be aspects I’m missing.
- Panoramic images are an exception - I more often go 3000 or 2500px on the short side.
1 Like
Interesting, I think I’m going for 1900x1200 at 85%. It works well, but sometimes I would like a bit more resolution. So, I’m gonna keep your tradeoff in mind and try it, when needed 
1 Like
It really depends on the program/JPEG compression library. There’s no standard for what a given quality setting means (and, indeed, many programs use different scales). ImageMagick, for example, only does chroma subsampling below 90, but that may not be true for other programs.
That said, a broad consensus seems to be that 85 is a good compromise between quality and filesize most of the time, if it’s for final display on the web. If you pixel peep you can see some artefacts, but at normal viewing sizes it’s unlikely anyone will notice.
2 Likes
Have some of us tested webp ? Would it be a viable alternative to jpg ? or even AVIF ?
personal tests
Webp produce smaller file on my 1file test and AVIF is way too much cpu intensive to be an alternative for me…
even if it’s smaller than PNG in lossless mode and I’m still waiting for the result in lossy mode …
How much you can push compression without it being noticeable also depends a lot on the content of the image. On large areas with smooth gradients artefacts become visible much faster than on higher frequency textures. So I think it’s not really possible to give a general number.
2 Likes
WebP has never been a good format for photos. In lossy mode, in particular, you will always lose color information.
You can see some comparisons in the article I posted in the JPEG XL thread:
2 Likes
I export to 4k and then have a resize action that adds a little sharpening optimized for 100% view on a screen. For playraws I use 1920x size, compression is 85% if I remember right.
For general web galleries, it is often better to have a larger pixelcount and compress harder - that is the one thing where webp actually works. But I consider the format a general fail due to it’s opaqueness to what it does to an image.
So 3000x at 60% will almost always look better than 1500x at 90% and the filesizes won’t differ much.
1 Like