It's good to see you back! One thing that AI is already making big moves on is culling. So far, I've been playing with Filter Pixel and Narrative Select to to go through high volume sports photos. The tech isn't quite there yet for sports photography specifically, but I think that is one of the lower-level frontiers that will be done very well by AI very soon. Heck, it's actually already really good for portraits and still subjects. Some info on the AI I listed above: FilterPixel is also allowing photographers to "train" their own AI by feeding thousands of photos separated into the "good" and "culled" piles, and when that process gets really good, it could be a game-changer. Narrative Select has really good focus and eye-detection to notify you very quickly whether the photo is sharp or if someone is blinking. It's pretty impressive if you need that info.
Very interesting. Thanks for sharing. I may have to check out those programs. It doesn’t surprise me that checking the eyes is a key component of how one of them operates. I could see with some advancements and training it could work for selecting and culling sports photos.
Great to see you back! I have spent hours with all of your videos, to pick up years of invaluable knowledge. Knowledge on Position, understanding use of settings, and solutions for difficult lighting solutions, have helped my results immensely. It’s getting harder to cull the results. Too many keepers. Thank you.
Great video Jerry. I'm very interested (and somewhat frightened) to see where this leads our industry. I think it is fairly straightforward to see where this will go on the post-processing side of things: automation in ingesting, culling, cropping, toning, captioning, and uploading. In terms of how this affects shooters is much cloudier, especially with generative fill. One thing is for sure - I'm dusting off the DAM book. Even AI generated images need to be organized and stored!
Hey Luke! Generative fill will almost be certainly something that journalists like you or I will never use, but I can see it being used for marketing departments, ad agencies, etc. But it is a little scary how easy it is to fake an image so it will be interesting to see how things move going forward. Lots of thoughts on this running around in my brain that I'll try to put together in the future :-)
Great piece Jerry, I agree the backend will be the first area to be affected, just look at what Adobe has done. Machine learning and the golden rules are a solid base for humans to learn, AI will catch up.
Generally speaking, it will become increasingly difficult to discern genuine from fake images, news stories, social media posts, etc. Automated image capture, editing, and posting will increase, but as long as people enjoy doing these things themselves, none will go away...similar to film.
From what I have heard, I understand that camera manufacturers are developing ways so that images produced in-camera are 'digitally signed' for authenticity. So when an image is altered, it will be known (at least to those who know how to authenticate an image) that it was manipulated. I do agree with you - the fun of this field is the creation and knowing you framed and nailed all the elements naturally!
Jerry, I hope you're right on this. The deployment speed and exponential improvement in AI likely means that developments we think may take 5 years will probably happen in more like 2-3. Auto captioning (accurate) would be so welcome. I'm sure the guy who runs codereplacements won't be happy, though. That being said, having an ingest, process, and upload all in 2-3 minutes would be so much better than spending an hour or so in the photo room after a game.
Glad to see you back. Doing a LOT of post processing for my high school sports photography, I would welcome AI taking over much of that work. But there are many instances of me cropping a pic 2 or 3 times, to get different individuals in the pic or an individual and play shots. If I can think of that, AI can probably be programmed to do similar. I see a need to upgrade my computer again.
Absolutely, not only would AI programming handle your thousands of images in a mere fraction of the time that you or anybody else could, there's no reason to think they also couldn't chop up an image into multiple formats/crops. So while the focus of this video was more for mass-media, there is indeed application for those covering HS or pee wee sports.
Absolutely no doubt this will impact our industry. I think you did a good job here of laying out the different ways that we will see that impact and I agree with your thoughts. We will just have to embrace the changes that end up happening and modify our methods to incorporate those changes into our workflows. I don't do a lot of high-volume deadline work, but what little that I do, I can see that this automation will help me deliver faster and I have no problem with it. Everyone wants to get their images out fast!
While what I do touch on does indeed relate more to high-volume deadline work, I can see this workflow applying across all levels of sports photography. Shooting a running event, or a softball tournament? Just upload and let AI run its magic. Maybe the data isn't out there to caption every high school or pee wee player (although who knows with how social media collects data and its future impact), but it could at least apply team names or tournament names and dates. Parents out there would love to be able to see/share their images of kids in real time as well, not just mass media consumers.
Glad to see you out with a new video. I can see this happening, but I also wonder if the quality will be the same. Will AI be able to capture or write the best stuff? I've had photos where a facial expression of someone on the sideline or someone raising their arms for a touchdown told the story of the receiver running after a catch, or the fact that this was that 1000th strikeout - whatever, it would take a lot to program the AI for all those stats and background awareness to ensure it's actually used. I'm sure we'll get to the point where it is usable before it's optimal. And I don't look forward to those times.
Great to see another video! I think most of your suppositions make sense, and agree with most of it. I do, however, disagree that if AI saves money in the post processing portion, that money will be reallocated. I can see this money being 'saved' for sure, but then kept as profit, or just less overhead.
This is indeed a fear, and admittedly there will be some money-saving spreadsheet-crunching organizations that choose to work this way. But I can also say with high level of certainty that there is more content out that many companies wish they can cover but cannot due to existing budget. I guess we'll see what prevails!
Interesting video. Not a fan of AI, I think it’s going to widen the gap between rich and poor as administrative jobs tend to pay more than in the field jobs. Especially with me as a Systems Admin, I’m positive AI could do much of my “preventative maintenance” type work. I guess the positive is just like any other computer program, there will need to be patches and updates to AI machines.
@JerryTLai, as a sports photo editor for one of the major agencies, I'm curious on your opinion as to the acceptable use of AI tools today for editorial photos. In the past, cropping and basic toning and color correction were generally the limits. However, AI based tools have significantly increased the capabilities to improve an image. I think we can safely assume that adding/removing content is still unacceptable for editorial photos, so no generative fill. However, there are AI tools to improve photo quality such as removing noise or blur (e.g. Lightroom, Topaz, etc). Is the use of those tools acceptable for editorial photos? Are your photographers allowed to use these AI based capabilities to improve their images? I appreciate you insights.
Haha, you're a step ahead of me. These are issues that rack my brain these days, and will probably be discussion points in my next video. The short answer is currently these tools are not acceptable for editorial photos. While I don't doubt some of them do a very good job and can fool editors, some of them do an extremely rough and poor job and are indeed flagged and spiked. It's only going to continue... anyways, thanks for watching & stay tuned!
A question on a different subject if it's OK. On some of the shots you showed of shooters with big glass at games a lot of them don't use lens shades. I see that a lot and no one could clue me in as to why.
In tight packed-in spots, shooting without telephoto lens hood gives a reduction of about 8 inches in total lens length (e.g., for a Canon EF 400 mm 2.8 V2 + EF/RF adapter for mirrorless camera). In the same situation, greater rotation visibility is offered to the adjacent photographers without the lens hood when side-by-side. Dealing with light flares without the hood will be an issue when locked in a fixed position. Otherwise find an area with space to not interfere with the view of other photographers or spectators.
@par4dcourse45 is right on. Taking off the lens hood is a courtesy to other photographers if shooting in packed locations. It gives everybody a chance rotate and pivot the way they need to based on what plays they want to follow. If I have plenty of room I will have the hood on but if it's tight I will take it off to play nice with other people near me.
The AI industry will not stop at anything. Photographers and photography completely changed and for the most part obsolete as it stands now. Resolution will be increased to where you could place 2-3 cameras around the field with an entire view, then start them shooting before the game begins.Then have AI pick and and edit keepers from the total footage, send them out, etc.. I hope I'm full of hooey but this doesn't seem to be out of the question to me.
In theory this would be the quickest way to achieve automation of the actual photography, but shooting everything wider just doesn't have the same effect of photographing things with a long lens due to depth of field effects. I think more advanced actual robots will have to come into existence before sideline photographers are obsolete. Fortunately, I think we are at least still a ways off from this scenario from coming to play.
I'm a bit sad, because of our country, photographers usually are not paid well... so you have to do some other jobs with it.... you are mostly outta free time... is that the same in other countries as well?
i'm a semi retired fitter / machinist and amateur photgrapher. Over my 45year working life I've seen CNC machines and computerised manufacturing systems take over many manual processes. I feel professional sports photography is now beginning to be on a similar path.. 🐯
It's good to see you back!
One thing that AI is already making big moves on is culling. So far, I've been playing with Filter Pixel and Narrative Select to to go through high volume sports photos. The tech isn't quite there yet for sports photography specifically, but I think that is one of the lower-level frontiers that will be done very well by AI very soon. Heck, it's actually already really good for portraits and still subjects.
Some info on the AI I listed above:
FilterPixel is also allowing photographers to "train" their own AI by feeding thousands of photos separated into the "good" and "culled" piles, and when that process gets really good, it could be a game-changer.
Narrative Select has really good focus and eye-detection to notify you very quickly whether the photo is sharp or if someone is blinking. It's pretty impressive if you need that info.
Very interesting. Thanks for sharing. I may have to check out those programs. It doesn’t surprise me that checking the eyes is a key component of how one of them operates. I could see with some advancements and training it could work for selecting and culling sports photos.
Great to see you back! I have spent hours with all of your videos, to pick up years of invaluable knowledge. Knowledge on Position, understanding use of settings, and solutions for difficult lighting solutions, have helped my results immensely. It’s getting harder to cull the results. Too many keepers. Thank you.
Happy you're back! Can't wait for more sports photography videos! 😀
Good info Jerry. Great to see you back making videos.
Great video Jerry. I'm very interested (and somewhat frightened) to see where this leads our industry. I think it is fairly straightforward to see where this will go on the post-processing side of things: automation in ingesting, culling, cropping, toning, captioning, and uploading. In terms of how this affects shooters is much cloudier, especially with generative fill. One thing is for sure - I'm dusting off the DAM book. Even AI generated images need to be organized and stored!
Hey Luke! Generative fill will almost be certainly something that journalists like you or I will never use, but I can see it being used for marketing departments, ad agencies, etc. But it is a little scary how easy it is to fake an image so it will be interesting to see how things move going forward. Lots of thoughts on this running around in my brain that I'll try to put together in the future :-)
Good insight! Glad you're back!
Great piece Jerry, I agree the backend will be the first area to be affected, just look at what Adobe has done.
Machine learning and the golden rules are a solid base for humans to learn, AI will catch up.
Generally speaking, it will become increasingly difficult to discern genuine from fake images, news stories, social media posts, etc. Automated image capture, editing, and posting will increase, but as long as people enjoy doing these things themselves, none will go away...similar to film.
From what I have heard, I understand that camera manufacturers are developing ways so that images produced in-camera are 'digitally signed' for authenticity. So when an image is altered, it will be known (at least to those who know how to authenticate an image) that it was manipulated. I do agree with you - the fun of this field is the creation and knowing you framed and nailed all the elements naturally!
Jerry,
I hope you're right on this. The deployment speed and exponential improvement in AI likely means that developments we think may take 5 years will probably happen in more like 2-3. Auto captioning (accurate) would be so welcome. I'm sure the guy who runs codereplacements won't be happy, though. That being said, having an ingest, process, and upload all in 2-3 minutes would be so much better than spending an hour or so in the photo room after a game.
Welcome back!
Glad to see you back.
Doing a LOT of post processing for my high school sports photography, I would welcome AI taking over much of that work. But there are many instances of me cropping a pic 2 or 3 times, to get different individuals in the pic or an individual and play shots. If I can think of that, AI can probably be programmed to do similar. I see a need to upgrade my computer again.
Absolutely, not only would AI programming handle your thousands of images in a mere fraction of the time that you or anybody else could, there's no reason to think they also couldn't chop up an image into multiple formats/crops. So while the focus of this video was more for mass-media, there is indeed application for those covering HS or pee wee sports.
Jerry, very thought provoking!
Great to see you back on RUclips! Always enjoy your videos. Looking forward to more great sports photography tips.
Absolutely no doubt this will impact our industry. I think you did a good job here of laying out the different ways that we will see that impact and I agree with your thoughts. We will just have to embrace the changes that end up happening and modify our methods to incorporate those changes into our workflows. I don't do a lot of high-volume deadline work, but what little that I do, I can see that this automation will help me deliver faster and I have no problem with it. Everyone wants to get their images out fast!
While what I do touch on does indeed relate more to high-volume deadline work, I can see this workflow applying across all levels of sports photography. Shooting a running event, or a softball tournament? Just upload and let AI run its magic. Maybe the data isn't out there to caption every high school or pee wee player (although who knows with how social media collects data and its future impact), but it could at least apply team names or tournament names and dates. Parents out there would love to be able to see/share their images of kids in real time as well, not just mass media consumers.
Thanks for sharing your interesting thoughts 😊
Very interesting and also thought provoking
Glad to see you out with a new video. I can see this happening, but I also wonder if the quality will be the same. Will AI be able to capture or write the best stuff? I've had photos where a facial expression of someone on the sideline or someone raising their arms for a touchdown told the story of the receiver running after a catch, or the fact that this was that 1000th strikeout - whatever, it would take a lot to program the AI for all those stats and background awareness to ensure it's actually used. I'm sure we'll get to the point where it is usable before it's optimal. And I don't look forward to those times.
Great to see another video! I think most of your suppositions make sense, and agree with most of it. I do, however, disagree that if AI saves money in the post processing portion, that money will be reallocated. I can see this money being 'saved' for sure, but then kept as profit, or just less overhead.
This is indeed a fear, and admittedly there will be some money-saving spreadsheet-crunching organizations that choose to work this way. But I can also say with high level of certainty that there is more content out that many companies wish they can cover but cannot due to existing budget. I guess we'll see what prevails!
I'm happy with just advanced denoise. That'll do AI, thanks!
Interesting video. Not a fan of AI, I think it’s going to widen the gap between rich and poor as administrative jobs tend to pay more than in the field jobs. Especially with me as a Systems Admin, I’m positive AI could do much of my “preventative maintenance” type work. I guess the positive is just like any other computer program, there will need to be patches and updates to AI machines.
@JerryTLai, as a sports photo editor for one of the major agencies, I'm curious on your opinion as to the acceptable use of AI tools today for editorial photos. In the past, cropping and basic toning and color correction were generally the limits. However, AI based tools have significantly increased the capabilities to improve an image. I think we can safely assume that adding/removing content is still unacceptable for editorial photos, so no generative fill. However, there are AI tools to improve photo quality such as removing noise or blur (e.g. Lightroom, Topaz, etc). Is the use of those tools acceptable for editorial photos? Are your photographers allowed to use these AI based capabilities to improve their images? I appreciate you insights.
Haha, you're a step ahead of me. These are issues that rack my brain these days, and will probably be discussion points in my next video. The short answer is currently these tools are not acceptable for editorial photos. While I don't doubt some of them do a very good job and can fool editors, some of them do an extremely rough and poor job and are indeed flagged and spiked. It's only going to continue... anyways, thanks for watching & stay tuned!
A question on a different subject if it's OK. On some of the shots you showed of shooters with big glass at games a lot of them don't use lens shades. I see that a lot and no one could clue me in as to why.
In tight packed-in spots, shooting without telephoto lens hood gives a reduction of about 8 inches in total lens length (e.g., for a Canon EF 400 mm 2.8 V2 + EF/RF adapter for mirrorless camera). In the same situation, greater rotation visibility is offered to the adjacent photographers without the lens hood when side-by-side. Dealing with light flares without the hood will be an issue when locked in a fixed position. Otherwise find an area with space to not interfere with the view of other photographers or spectators.
@par4dcourse45 is right on. Taking off the lens hood is a courtesy to other photographers if shooting in packed locations. It gives everybody a chance rotate and pivot the way they need to based on what plays they want to follow. If I have plenty of room I will have the hood on but if it's tight I will take it off to play nice with other people near me.
The AI industry will not stop at anything. Photographers and photography completely changed and for the most part obsolete as it stands now. Resolution will be increased to where you could place 2-3 cameras around the field with an entire view, then start them shooting before the game begins.Then have AI pick and and edit keepers from the total footage, send them out, etc.. I hope I'm full of hooey but this doesn't seem to be out of the question to me.
In theory this would be the quickest way to achieve automation of the actual photography, but shooting everything wider just doesn't have the same effect of photographing things with a long lens due to depth of field effects. I think more advanced actual robots will have to come into existence before sideline photographers are obsolete. Fortunately, I think we are at least still a ways off from this scenario from coming to play.
First video in one year Jerry
Can we see how you organize your pelican case with a 400mm in it , please?
I'm a bit sad, because of our country, photographers usually are not paid well... so you have to do some other jobs with it.... you are mostly outta free time... is that the same in other countries as well?
i'm a semi retired fitter / machinist and amateur photgrapher. Over my 45year working life I've seen CNC machines and computerised manufacturing systems take over many manual processes. I feel professional sports photography is now beginning to be on a similar path.. 🐯