OK here goes, according to DA's Print Image Specifications & Requirements, Link to Chart http://help.deviantart.com/132/ (http://help.deviantart.com/132/)
to get a 18"x24" High Quality Print I need to render an output Image of 5400x7200 Image Size(pixels).
Wondering how long it will take to render with the TG TP version we are using now? (JavaJones?) And does anyone (JavaJones?) know what we can expect from the Final TG2 when it comes out?
At this point I'm ready to throw the Towel in! :'(
Hi Jay!
print resolution is a much discussed issue...there are many factors that affect what resolution you should print at, like how far from the print will it be viewed?....it is possible to print at 200ppi/dpi at larger sizes as the human eye cannot differentiate what it cannot see! There is plenty about on the web, here is just one view:
http://www.crhfoto.co.uk/crh/filesizeres.htm (http://www.crhfoto.co.uk/crh/filesizeres.htm)
I hope this helps somehow....you may get a high enough quality print at 18"X24" with a 3600X2400 image?
good luck!
needless to say i will only be printing wallet sized prints at the moment!!
churz
J
:)
Quote from: inkydigit on January 24, 2007, 01:16:09 PM
Hi Jay!
print resolution is a much discussed issue...there are many factors that affect what resolution you should print at, like how far from the print will it be viewed?....it is possible to print at 200ppi/dpi at larger sizes as the human eye cannot differentiate what it cannot see! There is plenty about on the web, here is just one view:
http://www.crhfoto.co.uk/crh/filesizeres.htm (http://www.crhfoto.co.uk/crh/filesizeres.htm)
I hope this helps somehow....you may get a high enough quality print at 18"X24" with a 3600X2400 image?
good luck!
needless to say i will only be printing wallet sized prints at the moment!!
churz
J
:)
Thanks Jason, but I am aware of all the info on the web including the link I provided to DA. My point is, I don't want to settle for lesser quality in my prints because of incredibly ridiculous render times from TGTP. "Believe it or not" people do buy some of my work and I'm NOT willing to offer lesser quality prints because TG now or may never be optimized to the point of being a productive tool. I was trying not to be to rash, but I don't know how else to put it. I understand that the full version is scheduled to be out in 2007 and it's very difficult to believe that Planetside will make a productive tool from what we have at this point in time. I also wish that the Testers would be
willing and
allowed to share some of the knowledge that they have learned in the past year (or what ever the time has been) that they have been using TG TP. What's with all the secrets anyway?
:-\
one possible solution you might want to consider is using a RIP to uprez the output. Basically it's a bit of software/hardware that attaches to a printer and does all sorts of voodoo on your image and voila out it comes at a much more reasonable size.
These things are rather expensive to buy but a lot of pro printing labs will have one or more of 'em. These things are regularly used by photographers to get billboard size images.
Quote from: FrankThomas on January 24, 2007, 02:13:56 PM
one possible solution you might want to consider is using a RIP to uprez the output. Basically it's a bit of software/hardware that attaches to a printer and does all sorts of voodoo on your image and voila out it comes at a much more reasonable size.
These things are rather expensive to buy but a lot of pro printing labs will have one or more of 'em. These things are regularly used by photographers to get billboard size images.
Thanks for the info Frank and good suggestion! I just read some info on ImagePrint and you are right! Very Pricey @ $495 for the light version. Up to $2500 for large format printers.
Still hoping to get some answers from Planetside though.
@Jay...that is exactly the point I outlined towards the end of my thread on water rendering times!!!!!
"Really, what I am saying is that to produce a print quality(3200X2400) of the same image that took me 70 hrs (for 800X600)...would take 16 times as long...in other words, 1120 hours (or about 47 days!)!!!!
Even IF there was a 25% optimization success...it would still take 840 hours!
Based on the render time benchmark database...even IF a CPU was 3x as fast as my machine....that same render would still take 210 hours! Keep in mind, my render had NO clouds and NO objects, few shaders....it was technically very simple!
Unless SIGNIFICANT optimization is achieved in future builds of TG2, render times will be unacceptably long for "casual" users....and, I think that may affect registered program purchase decisions."
I'm not truly convinced that Up-resing software is the answer....I've played with the Alienskin blow-up demo...It's not bad in some situations.....But I don't think a 4X blowup of a TG2 image would produce the print quality results you ( or your customers) expect. I don't care what kind of magic they do, the bottom line is you only have so much data to begin with...to make more data means INTERPOLATION...which means FAKING data.
If they were really that great people would just be buying 3MP cameras and then up-resing the images with $100 software to get a 12MP image....instead of shelling out $8000 for the camera.
Anyway, all this takes us from the topic of discussion. The bottom line is that unless Planetside can pull some rabbits out of their hat when it comes time to render time optimization, they'll be left with a pretty but unwieldy beast that few can operate. I am hopeful, but not necessarily optimistic, that they can achieve that.
Quote from: gradient on January 24, 2007, 03:43:47 PM
@Jay...that is exactly the point I outlined towards the end of my thread on water rendering times!!!!!
"Really, what I am saying is that to produce a print quality(3200X2400) of the same image that took me 70 hrs (for 800X600)...would take 16 times as long...in other words, 1120 hours (or about 47 days!)!!!!
Even IF there was a 25% optimization success...it would still take 840 hours!
Based on the render time benchmark database...even IF a CPU was 3x as fast as my machine....that same render would still take 210hours! Keep in mind, my render had NO clouds and NO objects, few shaders....it was technically very simple!
Unless SIGNIFICANT optimization is achieved in future builds of TG2, render times will be unacceptably long for "casual" users....and, I think that may affect registered program purchase decisions."
'm not truly convinced that Up-resing software is the answer....I've played with the Alienskin blow-up demo...It's not bad in some situations.....But I don't think a 4X blowup of a TG2 image would produce the print quality results you ( or your customers) expect. I don't care what kind of magic they do, the bottom line is you only have so much data to begin with...to make more data means INTERPOLATION...which means FAKING data.
If they were really that great people would just be buying 3MP cameras and then up-resing the images with $100 software to get a 12MP image....instead of shelling out $8000 for the camera.
Anyway, all this takes us from the topic of discussion. The bottom line is that unless Planetside can pull some rabbits out of their hat when it comes time to render time optimization, they'll be left with a pretty but unwieldy beast that few can operate. I am hopeful, but not necessarily optimistic, that they can achieve that.
Yeah I know and I'm thinking exactly as you. I guess we will just have to wait and see what we get for a final version. The good side of this is we don't have to wait for years because it will be out this year, right. ::)
Well Just to be the optmistic one I'll just say that im sure Terragen 2 will be fast and efficent....But since you need it now you could try a render farm....very expencive but you can rent time for reltivly cheap but I don't know of any that currently support terragen 2.
Regards,
Will
Yes it's true that currently it would take way too long to render anything larger than regular desktop resolution size images. But given that there's still potential in optimizing the renderer for speed, plus multi-threading, plus availability of quad-core processors probably this year, we might have good reason for hope that the render times with larger images become more accessible to us.
On the other hand side, I have a TG0.9 render on 30" x 20" in my room. The source image was rendered with 4000x2600 pixels, and upscaled to match the 30" x 20" at 300 dpi. Together with its passepartout and frame, the whole image takes quite some space on the wall. However, you actually NEVER get so close to image that pixels become apparent - the picture is just too large to view it closer than 1 meter away from it.
Now, my point is, that TG0.9 full quality renders are comparable to a TG2 render at 0.5 quality, or less. With TG2 it's perfectly acceptable for most scenes to render at 0.5 quality, maybe up to 0.7 if you're ambitious. For carefully upscaled prints, that should do.
All the above - of course - pending that renderer speed optimizations really will happen. And I'm sure they will.
Best regards,
Frank
@FrankB...sorry, but I find your post somewhat confusing...
First you say "Yes it's true that currently it would take way too long to render anything larger than regular desktop resolution size images"....
Then you say "Now, my point is, that TG0.9 full quality renders are comparable to a TG2 render at 0.5 quality, or less. With TG2 it's perfectly acceptable for most scenes to render at 0.5 quality"
Are you suggesting that Jay render his 5400X7200 images at 0.5 quality....or....are you pinning TG2's render hopes on a "quad" processor that isn't even available yet?
Quote from: gradient on January 25, 2007, 02:51:53 AM
@FrankB...sorry, but I find your post somewhat confusing...
First you say "Yes it's true that currently it would take way too long to render anything larger than regular desktop resolution size images"....
Then you say "Now, my point is, that TG0.9 full quality renders are comparable to a TG2 render at 0.5 quality, or less. With TG2 it's perfectly acceptable for most scenes to render at 0.5 quality"
Are you suggesting that Jay render his 5400X7200 images at 0.5 quality....or....are you pinning TG2's render hopes on a "quad" processor that isn't even available yet?
Well, both are true at the same time, my intention was to point out two factors that are relevant to making posters from renders. On the one side, the renderer itself will improve in efficiency and future hardware will even boost that. On the other side, it's viable to reduce render times by not using quality 1 for everything. Using my own experience with making posters from TG0.9 images, I wanted to share that TG2 renders at 0.5 quality can still make good posters. Hope that helped wipe-out some of the confusion :-)
Regards,
Frank
Just in case anyone has more money than sense, Dell are currently selling servers with dual Quad core Xeon processors :)
A touch on the expensive side though and not a lot of use till we get the multithreading in TG2 but then they should be amazing for rendering
Well,
I render for fun.
If I would do it for profession or If I sold images, I would invest the earned money into new hardware. Any old machine could do the rendering - it does not even need a screen. And two machines could render simultanously cropped parts of a render.
My main machine, the one I do the development needs to be fast, as I need quick results für quick changes.
Everything else is a matter of patience.
Considering the pricing of Vue or Cinema4D (which are comparable slow in high quality large renders, too) you can spend 200 Bucks for Terragen and everything left for hardware.
This program is a stunning beauty. And anything else has to deal with its capabilities. Even time.
Just my two cents.
Volker
Another point comes into my mind.
Buzzzzz, the pictures you sold - were those made with TG0.9?
If so, did you downsampled these before printing?
If not, then there is no need to use the highest detail-settings in TGTP for equal large renders or you render higher quality in a smaller size.
Regards,
Volker
Quote from: FrankB on January 25, 2007, 03:20:24 AM
Quote from: gradient on January 25, 2007, 02:51:53 AM
@FrankB...sorry, but I find your post somewhat confusing...
First you say "Yes it's true that currently it would take way too long to render anything larger than regular desktop resolution size images"....
Then you say "Now, my point is, that TG0.9 full quality renders are comparable to a TG2 render at 0.5 quality, or less. With TG2 it's perfectly acceptable for most scenes to render at 0.5 quality"
Are you suggesting that Jay render his 5400X7200 images at 0.5 quality....or....are you pinning TG2's render hopes on a "quad" processor that isn't even available yet?
Well, both are true at the same time, my intention was to point out two factors that are relevant to making posters from renders. On the one side, the renderer itself will improve in efficiency and future hardware will even boost that. On the other side, it's viable to reduce render times by not using quality 1 for everything. Using my own experience with making posters from TG0.9 images, I wanted to share that TG2 renders at 0.5 quality can still make good posters. Hope that helped wipe-out some of the confusion :-)
Regards,
Frank
Quote from: Voulge on January 25, 2007, 04:46:38 AM
Well,
I render for fun.
If I would do it for profession or If I sold images, I would invest the earned money into new hardware. Any old machine could do the rendering - it does not even need a screen. And two machines could render simultanously cropped parts of a render.
My main machine, the one I do the development needs to be fast, as I need quick results für quick changes.
Everything else is a matter of patience.
Considering the pricing of Vue or Cinema4D (which are comparable slow in high quality large renders, too) you can spend 200 Bucks for Terragen and everything left for hardware.
This program is a stunning beauty. And anything else has to deal with its capabilities. Even time.
Just my two cents.
Volker
Quote from: Voulge on January 25, 2007, 05:40:29 AM
Another point comes into my mind.
Buzzzzz, the pictures you sold - were those made with TG0.9?
If so, did you downsampled these before printing?
If not, then there is no need to use the highest detail-settings in TGTP for equal large renders or you render higher quality in a smaller size.
Regards,
Volker
@ Frank B- OK, let me render 2700x3600 vs. 5400x7200 and also drop back from 1.0 quality to 0.5 quality and see what I get for print quality and render time. Thanks for your Input! Appreciated!
@ Voulge- I also render for fun and I love Terragen. However my Wife Beats me and makes me sell Prints to justify the many hours I spend playing with Terragen. ;)
@ Voulge_ On you second point- With TG 9.43 I render 5400x7200 pixels for large prints. Then I convert from 72ppi (which is the output of TG) to 266ppi which enlarges the image to 19,950x26,600 pixels and a document size of 75"x100". Now this I downsize to a document print size of 18"x24" which is then 4788x6384 pixels. So to answer your question, yes a slight downsize.
So this is what I am going to try, see my reply to FrankB.
Thanks for your Suggestions,
Jay
Hello Jay,
sorry that I forgot the " ;D :D ;)"s
If I get you right, you have a 5400x7200 original data, which is converted twice to make a 266dpi print while the original data would make a 300dpi of the same size. ;)
And I must admit that I was too lazy to read Franks's posts, I agree with him.
I used to downsample my TG0.9 renders at least to 50% to get a good screen-quality. With TGTP it is not necessary anymore. Thus your decision to render quarter size is very fine with me ,-)
Regards,
Volker
Quote from: Voulge on January 25, 2007, 10:04:54 AM
Hello Jay,
sorry that I forgot the " ;D :D ;)"s
If I get you right, you have a 5400x7200 original data, which is converted twice to make a 266dpi print while the original data would make a 300dpi of the same size. ;)
And I must admit that I was too lazy to read Franks's posts, I agree with him.
I used to downsample my TG0.9 renders at least to 50% to get a good screen-quality. With TGTP it is not necessary anymore. Thus your decision to render quarter size is very fine with me ,-)
Regards,
Volker
Lets not confuse ppi (Pixels per inch) which is monitor resolution and dpi (Dots per inch) which is printing resolution. Totally two different things. With 266 ppi screen resolution of the image and I normally print at 600 dpi which my print software handles. This most times gives me really good print viewing even up close.
well the print you plan is very huge so i think you can go with 200 dpi. How long will it render? easy to count just render 800x600 and then multiply to get the results.
@Jay
I'm trying to understand your workflow........you said;
"With TG 9.43 I render 5400x7200 pixels for large prints. Then I convert from 72ppi (which is the output of TG) to 266ppi which enlarges the image to 19,950x26,600 pixels and a document size of 75"x100". Now this I downsize to a document print size of 18"x24" which is then 4788x6384 pixels. So to answer your question, yes a slight downsize."
If I understand correctly....First you are upsizing your image...then you are downsizing your image. Both of these processes result in image degradation. If you want to achieve an 18X24 print @266dpi...why don't you do the conversion in PS all in one step?
BTW, I don't think you'll be too happy with 2700X3600 at 0.5 quality....
Oh...and for those who didn't do the math....a 2700X3600 render will take 20.25 times as long as the render at 800X600 for the same quality setting....
@Voulge;
Sorry I just read your thread....
you said;
"I used to downsample my TG0.9 renders at least to 50% to get a good screen-quality."
Yes, I agree that was a typical 0.9xx workflow.....and, it included some measure of sharpening between downsizes...But, for printing, you could take a 5400X7200 render, sharpen it...take it to a print shop and get a decent 18X36 inch print (regardless of ppi setting btw...just take your image file and ask for a 18X36 print!))
What it seems folks are trying to do now is take a 2700X3600 TG2 render (which, granted is of better quality than a comparable 0.9xx)...then UPSIZE it with software to 5400X7200...which degrades the quality and get that same 18X24 inch print.
Which method is more efficient in terms of time and quality?.....guess you guys will have to try it out....
We do expect a great deal of optimization between now and the final version, but it is virtually impossible to make any accurate estimate of just how much faster it will be.
Ultimately you have the chance here to work with and learn the TG2 system and perhaps also get some good renders out of it in the process. The alternative is to simply wait and start from scratch when the final is released. I would say then it would be better to experiment and learn now and simply lower your expectations to a more realistic level, taking into account the state of the application as a "preview".
It may be that you simply can't reasonably expect to use TG2 as a stand-in for TG 0.9 at this time *for your purposes*. Print resolution is something I would agree TG2 is not particularly good with at this time, but this will be improved in many regards. The real question is whether the final release will be appropriate for your needs and I feel confident saying that it will be.
The bottom line on all of this is that you're working with a Technology Preview. It is capable of being used for production, which has already been proven by several companies, but the intention of this release is not that it be specifically used for those purposes, and we make no warranty that it is suitable at this time for such. If you feel the render times and stability are acceptable for your uses, as many others seem to, then that's excellent and you can get a lot of value out of TG2 even in pre-release. On the other hand if you feel your needs are more demanding than the current release comfortably allows, the simple reality is you will have to wait for the final release to see if it truly meets your needs.
- Oshyan
When I made print sized renders in Terragen 0.9 I usually made them in 6000x9000 or something. This seemed to make it detailed and large enough for DeviantArt to accept them for prints. Even quite large sized prints.
But when trying to render in that size with Terragen 2 I get a error message... first this:
"trImage:unable to allocate pixelmap: size 648000000 bytes"
When I click ok I get "Unable to start render!"
Not sure why... but I did manage to find a threshold where I could render... About 4000x2667 or so...
Talking of production rendering (Motion Pictures) what is the industry standard resolution for rendering for this as I'd like to find out weather, I can do it on my system?
What this program really needs is to be taken in house at a major production facility (Can you say Dream Works Pictures and ILM) and have feed back form them as to what they feel is missing in terms of features and put the functionality into the software there is a president in the industry all ready with The Brazil Render System.
TG2 need to have a well known animator do some work with it and while they are working with the software they would be suggesting new features and improvements to the software, the resulting animation would be shown at the major animation festivals and Terragen would in that way become a house hold name.
If you look at this history of CGI you will find that this has happen before with Alias Power Animator (Now Maya) with the world famous Luxo Jr animation.
Regards to you.
Cyber-Angel
The standard film resolution render target is about 2000x2000, last I heard.
We do already have people from several well-known effects houses working with the current test versions. Certainly their feedback is very valuable, but the major issues like high resolution rendering and general workflow are well-known to us. I think feedback from industry professionals will actually be much more useful a bit later in development when we're more in the refining stage. Right now there are so many low-level, basic issues to work on that detailed expert feedback isn't needed to identify them or determine solutions.
- Oshyan