It is okay for you guys to help me out with my FYP?


If it takes 3 seconds to process a single RAW image on a PC then this performance is achieved in a physical implementation where the application has direct access to all CPU cores and can utilize the computing power fully. Most servers offered online (even root servers) are running virtualized in a shared host, sharing resources with other virtual servers. (I'm not referring to Apache vhosts here, but VMware or other forms of virtualization.) On such systems the CPU power is limited to avoid that a single server uses up all resources and all other servers are unusable. Although you might "see" 4 or 8 CPUs in your system reports, their computing power is not fully available to you. This will slow down all your processing further.
I have to agree with others, although technically maybe interesting I see no real value or need for such an application.

Thank you for addressing the physical and technical aspects of my application. The problem could be solved with the help of 'cloud' computing with specific hardware service level agreements as it can be very scalable and providers charge you for what you use. I'm still trying to explore how to proceed and what I can do for my project based on your responses.
 

Just a point to note, your program is banking on broadband internet being universally accessible and fast. All the other technical nitty gritty that other CSers have pointed out aside, I'm curious to know who exactly your target group is. Many, many photographers shoot raw, from new amateurs, serious amateurs to freelancers to professionals. Even within, say the "professional" group, there are many sub-strata such as studio photographers (where such a program might make a little sense) and travel photographers (where such a program will make minimal sense).

Just something for you to consider.

Yes, you are right. I do realize that I have a very broad group of target users after reading your response.

My initial target group are actually photopgrahers who has the intention to share RAW photos with trusted people online but have found it difficult because:
1. No photo sharing site supports it (Although new camera models would have to wait for DCRaw to be updated to do the post processing on my app)
2. They have to manually zip up the files, upload it to a separate file hosting provider, then distribute the links

Is there such a group of photographers that will share their camera RAW (so far the only reason I've thought of for the purpose of sharing is work)?
Is there really no purpose in sharing camera RAW?

Do you mind if I asked for you to explain on why:
1. Such a program might interest studio photographers?
2. And not travel photographers? (My thoughts would be the places that they go to might not have internet access)


Cheers! :cheers:
 

hi i have a suggestion for you. you can try to build a thin client application instead of a fully web application. in this case, what happens is that you can provide a software to download, this software will be very small (<20mb) to do the processing of your raw images, and synchronisation to the server after that.

a scenario for you:
1. you need to edit raw images urgently and sync with your client, but no photoshop on hand.
2. log into the website, download the application and do your processing.
3. when you are done, the application does the upload for you.

in this case you do not need to connect to the server too many times. only once when you are finished with editing.

another option will be to have the thin client application compress to jpeg and upload, thus avoiding other cs-ers view of RAW sharing and bandwidth consumption.

hope this helps :D
 

Yes, you are right. I do realize that I have a very broad group of target users after reading your response.

My initial target group are actually photopgrahers who has the intention to share RAW photos with trusted people online but have found it difficult because:
1. No photo sharing site supports it (Although new camera models would have to wait for DCRaw to be updated to do the post processing on my app)
2. They have to manually zip up the files, upload it to a separate file hosting provider, then distribute the links

Is there such a group of photographers that will share their camera RAW (so far the only reason I've thought of for the purpose of sharing is work)?
Is there really no purpose in sharing camera RAW?

Do you mind if I asked for you to explain on why:
1. Such a program might interest studio photographers?
2. And not travel photographers? (My thoughts would be the places that they go to might not have internet access)


Cheers! :cheers:


I doubt there are many/any photographers that will share their camera RAW. Reason is quite simple. Photographers rarely shoot and edit the same photo together. As far as I know, this doesn't take place simply because it's impractical.

I am not asserting that studio photographers definitely will be interested in an RAW editing software that requires uploading online. I was just using it as an example in contrast with travel photographers who likely not have internet access all the time or may not even bring a computer until they return to their home country if the location requires minimal baggage.
 

No. I would not use the application, for it takes donkey years to upload raw files...and I dun see what you can provide, and a desktop application cannot provide, and probably faster, more user friendly. If I need mobility, I think easier to bring a laptop than to try find a wireless / 3G network and slowly upload the raw to work on the files. Time spend uploading, laptop slower also can finish the job faster.

Even if your application is best in the market, and I'm tempted to use do I trust my raw files with you? What if your servers crashes without backup? Of cos if you do a great job, flicker just need to provide 90% of what you doing, I think most people end up with flicker instead...

So I dun see any value add that is compelling to use the service.

Thank you for revisiting my thread :)

My app would like to allow photographers to edit, manage and share their photos in one place.

I understand that camera RAW are the proprietary work of photographers and concerns you have like what wdEvA has brought up. I will keep in mind to protect the rights of the users. I believe this is something that can be assured via a statement on the website,

I agree that servers do crash, but they have redundancy for storage with the help of RAID technology.

From your reply, do you use Flickr to share your photos?

And is there any problems you might have with regards to post processing of camera RAW? As I am essentially trying to understand how you guys work and what kind of problems you face during the process.

Thanks!
 

Sorry, I'm just curious, what exactly are the academic requirements of your project?
 

Thank you for addressing the physical and technical aspects of my application. The problem could be solved with the help of 'cloud' computing with specific hardware service level agreements as it can be very scalable and providers charge you for what you use. I'm still trying to explore how to proceed and what I can do for my project based on your responses.
Check these SLA very carefully. Even in the foggiest cloud there must enough computing power for you, be it on demand or via pre-allocated resources. Check how the application works, how the CPU is addressed. I'm not an expert here, but not all applications can easily be virtualized. The spreading of computing power via virtualization layer can cause further problems under certain conditions. Similar to some older applications who cannot utilize multiple CPU's / cores.
Secondly: how long do you expect people to wait for the results to show up? If Rashkae's example shows 3 seconds (based on physical implementation) and your cloud gives you only computing power for 6 seconds, plus various other factors that could delay the performance then the user experience will be quite bad. There are already surveys and figures out about how user perceive performance and delay with websites. Afaik, everything slower than a few seconds to load (at least the main parts of the page) is perceived as slow.
I have my doubts about a broad market that would use your application. But in schools and other locations, being short on budget and equipment I can imagine your application could be helpful for the first steps in picture editing. Basic functions, preloaded images (e.g. done by the teacher), server equipment somewhere in the school, LAN connection to all PC, low level PCs with just a browser .. that's an environment that could use your idea.
 

some people might use it, whether they are willing to pay for it is another big question.

i think you are referring to the big question? :bsmilie:

i don't think we have reach stage of printing option (to print yourself or to outsource) for this scenario.
 

hi i have a suggestion for you. you can try to build a thin client application instead of a fully web application. in this case, what happens is that you can provide a software to download, this software will be very small (<20mb) to do the processing of your raw images, and synchronisation to the server after that.

a scenario for you:
1. you need to edit raw images urgently and sync with your client, but no photoshop on hand.
2. log into the website, download the application and do your processing.
3. when you are done, the application does the upload for you.

in this case you do not need to connect to the server too many times. only once when you are finished with editing.

another option will be to have the thin client application compress to jpeg and upload, thus avoiding other cs-ers view of RAW sharing and bandwidth consumption.

hope this helps :D

Hello~ :) Thanks for sharing your ideas!

Thin clients will work but that would be something similar to Picasa.

As part of the course requirements, they want something that has not been done before (Crazy isn't it? lol)

To build on your idea, I'm thinking that if it is possible to build a thin client that runs on the web browser with JavaScript and HTML5. There's a website called Cloud Canvas that is built entirely using HTML5 canvas which works like Photoshop or Illustrator but with a very limited feature set.
 

Sorry, I'm just curious, what exactly are the academic requirements of your project?

Oh its no problem at all! You guys have been of great help actually :D

They just want us to go through the entire system development process from Analysis -> Design -> Development -> Testing

We choose from a list of broad topics and then zoom in on a specific area to focus on. The broad topic I've chosen is actually titled 'Develop a personal photoblog', which has to do with photography then with the help of my project supervisor. I was introduced to camera RAW and decided to focus on it to explore ways to share them online. But it seems like I've hit a roadblock as there is no requirement for such an app :sweat:

Sorry for the lengthy paragraph :X

Right now, I would say I'm still at the analysis stage doing my user research on this fantastic community :D
 

Thanks for the welcome. You guys are really helpful :)

I have the same question for you as above:

Since you mentioned that it might be feasible when the NGNBB picks up, is it alright to say that if bandwidth is fast, cheap and abundant, there might just be a chance that you might use it? I know this is out of context but limited bandwidth capacity is not stopping people from uploading to YouTube right? :X

I'm really glad you brought up concerns with regards to copyright issue pertaining to camera RAW. From the responses I've received so far, I understand the importance of it and will keep in mind to ensure that rights belong to the users.

Thanks for the list of software for my reference! Btw, is it possible for me to get the bundled software from DSLRs without actually purchasing one? I went to Canon support site but it only lets me download the updater :(

hmm hosting images and videos is one thing, but processing them is another.
personally, even with high speed broadband, I would still prefer to process the images on my own system. It will always be faster no matter how fast broadband goes.
And I've the full flexibility on the controls as well as no worries on the status of the server, as well as the status of my broadband.

As for the software, you can download the 30-days trial of photoshop and lightroom to try it out.
Not sure if Nikon / Canon provides a trial download, but ya, the software is free only when you purchase a camera from them.
 

Wow, your PC must be pretty good! Anyway I tried it post processing a Sony A900 RAW of about 37MB (the biggest I can find) resize to 1024x768 on a quad core server processor and I got about the same results as you.. It took around 2.8 to 3.2s.

Nope, normal quad-core. Desktop chips are actually frequently faster than server chips - most server chips are running at a lower clock with ECC RAM in order to maximize reliability instead of those last 5% of performance, and I have full CPU availability. As previously mentioned, unless you fork out huge $$$ for a dedicated server box, you will only get CPU slices.

What software do you use for post processing? Photoshop, lightroom or the bundled software from your camera?

Depends on which gives better results - right now CS5 and the bundled Sony software (version 3) both give good results. I would say I use photoshop more as I usually do smoothening, blemish/eyebag correction, etc. as well.


No, I do not want to get fired :bigeyes: The office was just an example, maybe a better example would be a cyber cafe, friend's house :bsmilie:

Cyber cafe, where they are on a cheap plan, and it's usually hogged by facebook gamers, bittorrent, etc? A friend's house? Why? I'd just go to my own house.


When you say:

1. Sharing low-res shots with customers, do you mean images that are in 1024x768 resolution as mention earlier in your workflow?

Correct.

2. You share via YouSendIt or your personal FTP, do you also use Flickr, Picasa or other photo sharing sites to share your photos?

Not with customers, only smaller travel snapshots.

If we take bandwidth issues (cost and capacity) out of the picture, would this application actually be:
1. An app that supports both tweaking and sharing capabilities that could complement your workflow? I mean rather than having you to manually zip up the selected photos, upload and ensure that it is actually the correct file, it does all of that for you?

Nope, because my actual workflow could be far more complex - There's all the retouching as well.

2. A good collaboration tool if you are on a team? You save the time in setting up your LAN and synchronisation issues with regards to who actually have the most updated post processed image?

No, because we do not edit the same files at the same time.

Also, is it okay for you to discuss the workflow of how you actually work in a team? How is the LAN setup? Do they work via a NAS? How you divide the work, how you put it altogether later on?

We each work on our own shots, using an agreed-on whitebalance and specific post-processing saved actions that are shared.

So is it alright to say that if bandwidth is fast, cheap and abundant, there might just be a chance that you might use it? ;p

Nope, sorry. Too many limitations beyond just bandwidth.
 

Oh its no problem at all! You guys have been of great help actually :D

They just want us to go through the entire system development process from Analysis -> Design -> Development -> Testing

We choose from a list of broad topics and then zoom in on a specific area to focus on. The broad topic I've chosen is actually titled 'Develop a personal photoblog', which has to do with photography then with the help of my project supervisor. I was introduced to camera RAW and decided to focus on it to explore ways to share them online. But it seems like I've hit a roadblock as there is no requirement for such an app :sweat:

Sorry for the lengthy paragraph :X

Right now, I would say I'm still at the analysis stage doing my user research on this fantastic community :D

No worries.

I am just concerned that from the looks of things.. well, if the project does not require you to do a camera-RAW specific program then you might want to KIV the idea and brainstorm for other possibilities. Besides the thin app that has already been suggested, I honestly can't think of anything wrong with what is already avaliable for RAW processing, so if you focus too much on forcing out a project relating to RAW, then you might end up with a lacklustre piece that has too little practical applications and too many limitations.
 

Last edited:
Check these SLA very carefully. Even in the foggiest cloud there must enough computing power for you, be it on demand or via pre-allocated resources. Check how the application works, how the CPU is addressed. I'm not an expert here, but not all applications can easily be virtualized. The spreading of computing power via virtualization layer can cause further problems under certain conditions. Similar to some older applications who cannot utilize multiple CPU's / cores.
Secondly: how long do you expect people to wait for the results to show up? If Rashkae's example shows 3 seconds (based on physical implementation) and your cloud gives you only computing power for 6 seconds, plus various other factors that could delay the performance then the user experience will be quite bad. There are already surveys and figures out about how user perceive performance and delay with websites. Afaik, everything slower than a few seconds to load (at least the main parts of the page) is perceived as slow.
I have my doubts about a broad market that would use your application. But in schools and other locations, being short on budget and equipment I can imagine your application could be helpful for the first steps in picture editing. Basic functions, preloaded images (e.g. done by the teacher), server equipment somewhere in the school, LAN connection to all PC, low level PCs with just a browser .. that's an environment that could use your idea.

Wow, this is really informative. From your response, are you a network engineer, system analyst, web developer in disguise? :bsmilie:

Yeah, your idea could work in schools which conduct photography workshops for students that teach them how to edit their photos from camera RAW. They could save up on the expensive license from Adobe.

Without Adobe software, it also addresses the compatibility issues faced across all camera manufacturers. They don't have to install every single bundled software and worry about the different functionalities across the different programs.
 

some people might use it, whether they are willing to pay for it is another big question.

i think you are referring to the big question? :bsmilie:

i don't think we have reach stage of printing option (to print yourself or to outsource) for this scenario.

No worries.

I am just concerned that from the looks of things.. well, if the project does not require you to do a camera-RAW specific program then you might want to KIV the idea and brainstorm for other possibilities. Besides the thin app that has already been suggested, I honestly can't think of anything wrong with what is already avaliable for RAW processing, so if you focus too much on forcing out a project relating to RAW, then you might end up with a lacklustre piece that has too little practical applications and too many limitations.

Yeah I see your concerns.. So far the response is bad and the outlook is pretty bleak.. :sweat: haha
 

Yeah, your idea could work in schools which conduct photography workshops for students that teach them how to edit their photos from camera RAW. They could save up on the expensive license from Adobe.

Without Adobe software, it also addresses the compatibility issues faced across all camera manufacturers. They don't have to install every single bundled software and worry about the different functionalities across the different programs.

1 Schools with a media lab use the educational licensing - that's SGD 99 for the whole suite, all shouldered by MOE. Other than that, if the lab uses a Nikon camera or a Canon or a Sony, etc, for it's classes, they have the free software that came with the camera. Usually they will only use one program, or a RAW file provided by the teacher and they all do the same edits.

2. School labs do not have the bandwidth for your app either.

Are you an SMU SIS student? Sounds like their projects.
 

Also, having done extensive work with the Microsoft Azure team, I can assure you that cloud-eneabling an application is not easy - and the server costs not cheap.
 

1 Schools with a media lab use the educational licensing - that's SGD 99 for the whole suite, all shouldered by MOE. Other than that, if the lab uses a Nikon camera or a Canon or a Sony, etc, for it's classes, they have the free software that came with the camera. Usually they will only use one program, or a RAW file provided by the teacher and they all do the same edits.

2. School labs do not have the bandwidth for your app either.

Are you an SMU SIS student? Sounds like their projects.

99? That is really cheap man compared to commercial licenses..

Thanks for letting me know how the schools conduct their workshop.

Do you mean that the workshops are conducted with respect to the brand of the camera? So if I own a Olympus DSLR, I wouldn't be able to benefit as much as say a workshop that uses specifically Canon DPP?
 

1 Schools with a media lab use the educational licensing - that's SGD 99 for the whole suite, all shouldered by MOE. Other than that, if the lab uses a Nikon camera or a Canon or a Sony, etc, for it's classes, they have the free software that came with the camera. Usually they will only use one program, or a RAW file provided by the teacher and they all do the same edits.

2. School labs do not have the bandwidth for your app either.

Are you an SMU SIS student? Sounds like their projects.

Oops I forgot to answer your last question.. No.. I'm a SIM student.. :p