Farcaster

hey @farcaster is there anyway for you to make the image preview centered when posting? 

for example here’s the full photo i posted yesterday compared to the preview. 

instead of centering the preview is cropping a ton of the bottom. 

i’ve literally cropped photos just to post on farcaster because i know it’s messed up. but it’s been this way for like a year haha 

@ted i know you see what i mean lol

A sufficiently decentralized social network. farcaster.xyz

building farcaster | surfing @ venice, ca | eating @ gjusta / gjelina | “lucky me, lucky mud” 🤠

personality hire 🌈🤙 fullstack & ux, building /uno

building @dynamic co-founder & prev cto @bello /bello 🍄 @a16zcrypto csx alum prev eng: lunarcrush, nbc, comcast, lockheed martin

i know exactly what you mean, but no not right now :(

we're currently revamping embeds in advance of Pro though so will take this into account

i think the issue is that the preview in the cast vs. the preview in the preview of the cast vs. the actual cast are all different previews. @erica has also noticed this lol

haha yess they don't all appear the same across the client rn  

what's interesting is twitter actually wrote a ML-driven algo for image cropping so the most interesting aspect of the image was always in the centre of the crop

however the early stages of the algo had a bias towards prioritizing certain skin tones...eep

oh I meant the farcaster app/client itself doesn’t show images in the same way everywhere !

on /uno they aren’t cropped at all bc the true center isn’t always the “most interesting” part. smart cropping takes a lot more effort than just not cropping 😅

yeah super interesting! i feel like if the image has a height of 500px you could just take 20px from the top and 20 from the bottom. that way the focal part of the image doesn’t get cut off. 

i only use the farcaster client so not sure how the other ones look tbh

haha yess they don't all appear the same across the client rn  

what's interesting is twitter actually wrote a ML-driven algo for image cropping so the most interesting aspect of the image was always in the centre of the crop

however the early stages of the algo had a bias towards prioritizing certain skin tones...eep
https://blog.x.com/engineering/en_us/topics/insights/2021/sharing-learnings-about-our-image-cropping-algorithm