I think, in fairness, that you should acknowledge that Photrus is not just a company sponsoring your videos, but is actually _your own_ company, working from software you coded. Very impressive! My only question is exactly how many servers, supercomputers, etc. do you need to generate worldwide coverage at such a fine level of resolution?
Hi @ravenfallsphotography6254, oh yes, it is my baby, of course 😊 In some videos I mention me as the developer, but I'm currently not sure if that is really a good idea, I have to say. I think, it could support that message of "You need to be a technical affine person to become a good photographer". Something many beginners think, in my experience, but that is simply wrong and that is a trend I would like to go again. I don't know if that is the right way, not to mention me as the developer, to be honest. I'm currently just experimenting with that. But your feedback is something I will definitely think about in the future, so thank you therefore, my friend! To your question of servers: The truth is that I don't know the exact number of servers 😊 There are 2 layers: Layer 1 are supercomputers from different weather providers who offer really exact point weather predictions. I don't know their exact server number or cpu-number etc., but they I guess that are many. These predictions are extremely accurate, exactly what we need, but they are also quite expensive to us. And as we - again - seldom photograph just a point in a landscape, we purchase a huge amount of point weather predictions in a radius of 1000 to 1200 kilometers around each single photo location of our clients. So, it is a solution on demand: we purchase them for each single photo spot of our clients, if a clients wants to get an accurate photography weather prediction at a photo location. Based on that there is a second layer, which calculates the photography weather phenomena like red sky, rainbows, golden clouds, etc. That is made by our algorithm on our servers on a high availability cluster. That is a cluster of multiple servers, which we cascade, according to the number of clients and number of photo locations we have to observe. Meanwhile we have clients all over the world, though we anyway don't calculate weather phenomena on each possible photo location on our planet. The number of servers is growing and growing, depending on the number of clients and photo spots. So, no way to tell a fix number of servers. But there are already a lot of CPU cores working and that 24 hours a day 😊 There is a lot going on in the background. Our clients don't see all that of course. For them the only interesting thing is to get an alert email with suggestions where to photograph to get stunning light and weather. And that is exactly what Photrus offers them. I hope that gives at least a little insight for you 😉 Enoy your Sunday, Christian
Very informative video Christian. It's so true with the weather changing quickly. You can only plan so much but the rest is up to nature. Guess we just have adapt to what nature gives us and shoot accordingly. Cheers buddy 🙏
Hi Adrian, thanks a lot, buddy 😊 Yes, it is never a 100% thing, but the mentioned point weather predictions are really extremely accurate. Not comparable to the output of common weather apps. It is just that they are not useful for landscape photographers, as again: we seldom photograph just a point in a landscape 😀 The solution is really to use hundretthousands of point weather predictions in a section of 1000x1000km around each photo spot, as Photrus does to calculate its own models, based on photography weather phenomena. Predicting these photography phenomena just works because of those accurate point weathers. But in the end of the day, as you said: the skill of adapting is anyway really, really important. Because sometimes even the best forecasts fail. Thanks a lot and have a great Sunday, Christian
on my recent dream trip to Iceland I had planned a lot of photographs, but I hadn't counted on bad luck... 5 days out of 6 of rain and zero solar storms 😅😅 We spent every evening checking the weather maps to see if there would be a gap in the clouds on our route, a break from the rain, the possibility of photographing the aurora... And on the last day, thanks to the study of the maps, we decided to bet everything on one location, skipping other stops. And in the end we managed to see the northern Lights ... Now I look at the photographs that I took in non-ideal conditions and they are still beautiful for me because we tried to get the most out of what we had. now a technical question on Photrus: how far in advance of the chosen event do you have a reliable forecast? now I'm going to review your other videos, maybe you've already talked about. Always beautiful content in your videos!
Hi Stefano, so great to hear that it worked with the northern lights in the end 😊 To your question about reliable results: in general, it is not possible to predict weather all too far in the future. I mean, we could use models that offer us forecasts for 16 days in the future, but they are wrong 😊 They just show tendencies. Photrus only uses reliable data, which means up to ~48 hours on very high-resolved weather data, but also here are differences, especially between different phenomena, but in average I would say: predictions 48 hours in the future are likely, predictions 24 hours in the future are very likely, and predictions 12 hours in the future are accurate. The biggest exception is the phenomenon sunrise/sunset with gappy clouds (which uses an additional gappy clouds algorithm for the gappy cloud phenomenon for Premium users) and it predicts the weather in extreme fringe situations. You need quite a lot of things coming together for this one. I made an interesting video about that, by the way: ruclips.net/video/tf0id9TkWtE/видео.html Fog predictions of Photrus are also different from common fog forecasts, as it also takes the difference updrift into account. I explained that in this video: ruclips.net/video/FEISPOtA7P0/видео.html Another difference is also that Photrus looks at real data from the past in some situations. It is not really possible to compare it with other tools, as it calculates phenomena, not only raw weather data. It goes far beyond that. I hope that helps 😉 Have a great weekend, Christian
Hi Fabio, Thank you so much for your kind words, my friend! Means a lot 😊 I don't have experience with the canon pro 300, I have to admit. I decided against Canon and for Epson instead - but that for some good reasons, I mentioned in this video: ruclips.net/video/vbatIn6CznY/видео.html I hope that helps 😉 Nice greetings, Christian
![[#2024MAMA] ROSÉ (로제), Bruno Mars - APT. | Mnet 241122 방송](http://i.ytimg.com/vi/dgGqD28J6aQ/mqdefault.jpg)
I think, in fairness, that you should acknowledge that Photrus is not just a company sponsoring your videos, but is actually _your own_ company, working from software you coded. Very impressive! My only question is exactly how many servers, supercomputers, etc. do you need to generate worldwide coverage at such a fine level of resolution?
Hi @ravenfallsphotography6254, oh yes, it is my baby, of course 😊 In some videos I mention me as the developer, but I'm currently not sure if that is really a good idea, I have to say. I think, it could support that message of "You need to be a technical affine person to become a good photographer". Something many beginners think, in my experience, but that is simply wrong and that is a trend I would like to go again. I don't know if that is the right way, not to mention me as the developer, to be honest. I'm currently just experimenting with that. But your feedback is something I will definitely think about in the future, so thank you therefore, my friend!
To your question of servers: The truth is that I don't know the exact number of servers 😊 There are 2 layers: Layer 1 are supercomputers from different weather providers who offer really exact point weather predictions. I don't know their exact server number or cpu-number etc., but they I guess that are many. These predictions are extremely accurate, exactly what we need, but they are also quite expensive to us. And as we - again - seldom photograph just a point in a landscape, we purchase a huge amount of point weather predictions in a radius of 1000 to 1200 kilometers around each single photo location of our clients. So, it is a solution on demand: we purchase them for each single photo spot of our clients, if a clients wants to get an accurate photography weather prediction at a photo location.
Based on that there is a second layer, which calculates the photography weather phenomena like red sky, rainbows, golden clouds, etc. That is made by our algorithm on our servers on a high availability cluster. That is a cluster of multiple servers, which we cascade, according to the number of clients and number of photo locations we have to observe. Meanwhile we have clients all over the world, though we anyway don't calculate weather phenomena on each possible photo location on our planet. The number of servers is growing and growing, depending on the number of clients and photo spots. So, no way to tell a fix number of servers. But there are already a lot of CPU cores working and that 24 hours a day 😊
There is a lot going on in the background. Our clients don't see all that of course. For them the only interesting thing is to get an alert email with suggestions where to photograph to get stunning light and weather. And that is exactly what Photrus offers them.
I hope that gives at least a little insight for you 😉
Enoy your Sunday,
Christian
Very informative video Christian. It's so true with the weather changing quickly. You can only plan so much but the rest is up to nature. Guess we just have adapt to what nature gives us and shoot accordingly. Cheers buddy 🙏
Hi Adrian, thanks a lot, buddy 😊 Yes, it is never a 100% thing, but the mentioned point weather predictions are really extremely accurate. Not comparable to the output of common weather apps. It is just that they are not useful for landscape photographers, as again: we seldom photograph just a point in a landscape 😀 The solution is really to use hundretthousands of point weather predictions in a section of 1000x1000km around each photo spot, as Photrus does to calculate its own models, based on photography weather phenomena. Predicting these photography phenomena just works because of those accurate point weathers. But in the end of the day, as you said: the skill of adapting is anyway really, really important. Because sometimes even the best forecasts fail.
Thanks a lot and have a great Sunday,
Christian
on my recent dream trip to Iceland I had planned a lot of photographs, but I hadn't counted on bad luck...
5 days out of 6 of rain and zero solar storms 😅😅
We spent every evening checking the weather maps to see if there would be a gap in the clouds on our route, a break from the rain, the possibility of photographing the aurora...
And on the last day, thanks to the study of the maps, we decided to bet everything on one location, skipping other stops. And in the end we managed to see the northern Lights ...
Now I look at the photographs that I took in non-ideal conditions and they are still beautiful for me because we tried to get the most out of what we had.
now a technical question on Photrus: how far in advance of the chosen event do you have a reliable forecast?
now I'm going to review your other videos, maybe you've already talked about.
Always beautiful content in your videos!
Hi Stefano, so great to hear that it worked with the northern lights in the end 😊 To your question about reliable results: in general, it is not possible to predict weather all too far in the future. I mean, we could use models that offer us forecasts for 16 days in the future, but they are wrong 😊 They just show tendencies. Photrus only uses reliable data, which means up to ~48 hours on very high-resolved weather data, but also here are differences, especially between different phenomena, but in average I would say: predictions 48 hours in the future are likely, predictions 24 hours in the future are very likely, and predictions 12 hours in the future are accurate.
The biggest exception is the phenomenon sunrise/sunset with gappy clouds (which uses an additional gappy clouds algorithm for the gappy cloud phenomenon for Premium users) and it predicts the weather in extreme fringe situations. You need quite a lot of things coming together for this one. I made an interesting video about that, by the way: ruclips.net/video/tf0id9TkWtE/видео.html
Fog predictions of Photrus are also different from common fog forecasts, as it also takes the difference updrift into account. I explained that in this video: ruclips.net/video/FEISPOtA7P0/видео.html
Another difference is also that Photrus looks at real data from the past in some situations. It is not really possible to compare it with other tools, as it calculates phenomena, not only raw weather data. It goes far beyond that.
I hope that helps 😉
Have a great weekend,
Christian
Danke für das interessante Video!!!!! 👍👍
Vielen Dank fürs Schauen 😊
Nice tips 👌
Thanks a lot, @lubos-ro3hy, for your appreciation 😊
Have a great Sunday,
Christian
Hi. I love your channel. What do you think canon pro 300 for print my landscape and aerial phoography for sell, have other better print? thanks
Hi Fabio,
Thank you so much for your kind words, my friend! Means a lot 😊
I don't have experience with the canon pro 300, I have to admit. I decided against Canon and for Epson instead - but that for some good reasons, I mentioned in this video: ruclips.net/video/vbatIn6CznY/видео.html
I hope that helps 😉
Nice greetings,
Christian