About textures and render quality: A recent post by Anton got my simple mind cranking. Okay, I put a 2000 x 2000 body texture on DAZ mike. Does the size of my final render effect how crisp or blurred the appearance of the texture will look? if it does, is it a substantial difference the farther the final render size is compared to the texture size? Like if I render below or above 2000 x 2000 does it make a difference or does Poser (or whatever software one uses) downsize it automatically regardless of the render size I choose? One more question linked to this: Let's say I have a painted body texture and a photorealistic texture, both at 2000 x 2000. Does this 'sizing' issue affect the two types of textures differently? What I mean is, is there an optimal ratio for painted textures and a separate optimal ratio for photographed textures? About render settings: Is there a threshhold point where too much DPI (say over 300) doesn't add anything significant to render quality? Rendering huge: If I render a Poser scene real huge (say 5000 x 5000) does it make a sharpness difference when downsizing in the render in photoshop for webpages? Thanks, Curious ramster

.

Thanks tasquah. I will deliberate upon your wisdom. That's the kind of answer my Sensei often gives me. I often wonder if it's his answer or if his hearing aid isn't turned on. And we sit there quietly drinking our tea. Got your hearing aid in :D ram

Is there a threshhold point where too much DPI (say over 300) doesn't add anything significant to render quality? Depends on what you intend to do with the final result. If you intend to print it in a magazine, find out the LPI of the magazine's screen. For instance, if it's going on newsprint (web press) it doesn't make any sense to make it over 160 dpi. If it's intended for onscreen only, make it 72dpi (or 96dpi if you expect it to be viewed on some high-res Mac monitors, though this will make it display at 'full size' bigger on a PC or lower-res Mac) If you're making a poster, go with 1.5 times the screen LPI and you'll be safe. For T-shirts, 60 dpi is fine because the screen weave of a standard silkscreen is 42 threads per inch. The idea is that you want to be a little over (in dpi) the LPI of the screen in question to be used in production, if the production uses a screen. If you're printing to film, there's an equivalent measure of film grain size which corresponds to the film wuality and speed. If you're printing a movie, have fun waiting on the render, even in Poser. If you're outputting NTSC for television, the screen is really low and is interlaced (meaning that to maintain a smooth animation quality each frame only changes the even or odd fields), and the pixels aren't square. Most printers, even bubblejets, nowadays, go much hogher than 300 ppi output. Rendering huge: If I render a Poser scene real huge (say 5000 x 5000) does it make a sharpness difference when downsizing in the render in photoshop for webpages? Nope, but it does allow you to show a detail or a link to a larger copy. Let's say I have a painted body texture and a photorealistic texture, both at 2000 x 2000. Does this 'sizing' issue affect the two types of textures differently? What I mean is, is there an optimal ratio for painted textures and a separate optimal ratio for photographed textures? There's a logical threshold above which more detail does no good. To put it in simple terms, lets say you apply a 72x72 pixel texture to a square and render that square from the front (non-perspective) camera zoomed so the square is 1 inch by 1 inch in size and the result is intended for only on-screen viewing (a web graphic or screen saver for instance) so you use 72dpi. That's the optimal size for that texture at that output doppage. If you zoom in the image will lose quality (because the texture pixels will be rendered at bigger than 1 for 1 size), but if you zoom out it won't gain quality (because at 72dpi you cannot render a pixel smaller than one pixel). If you zoom out so that the resulting size is 1/2 inch x 1/2 inch, you only need a 36x36 pixel texture. Of course, since things are usually in perspective, the resolution should match the relative size of the closest part. Note that while this is often also true of bumpmaps, there are a few things to keep in mind: 1) bumpmaps in poser (yes, even 5) work at a lower resolution than you think -- about 1/3 the resolution you make them. Of course, this isn't always a bad thing because things should generally not look as bumpy as they technically really are. 2) sometimes you want an even higher bumpmap because of efects you can do with them, like, for instance, you can use 'noise' type bumpmaps to make the highlight on a surface gradually fade from a sharp one to a fuzzy one (somehting that just doesn't work with material IDs, unless you make it one poly per molecule) - - - - - -

Modelers should note that these same principles apply to modelling. Polylicious models aren't an end-all for models. Just being a higher resolution mesh won't mke a model better. If you're generating a crowd of people in the stands at an Eagles game, you want low-poly meshes -- even simplified human-like forms. They wouldn't even need all the body parts -- for instance, posable fingers would be absurd and even posable wrists might be unnecessary. Facial details could be texture mapped in and conforming clothing is a waste -- just texture and bumpmap it on. On the other hand, if you want a picture of the rolling hills of breast from the viewpoint of a mosquito alighted on a nipple (ouch!) even Vicki 3's mesh isn't going to be enough, as you will see a really rather low-res nipple at so close a viewpoint.

Ram: Tasquah was just bookmarking the thread because he wanted to be informed of any replies to it so that he could also see those answers (indicating that he was wondering the same thing). It's sort of like a more silent 'Mee too!'

Example: This is part of a recent render of my Cpt. Panaka's Blaster prop based on the one in The Phantom Menace and Attack of the Clones. This is intended for scren viewing only, to show off the prop in here. Thus it's aimed at 72dpi output. The selected area is about 173 pixels high (ish, I made the square aftyer measuring it). That's about 2.4 inches. The grip uses its own texture template that fills the image. A 72dpi 2.4 inch tall (ie, 173 pixel tall) texture is all that's needed tomake this render at this quality. Anything more would be overkill. Half that and the texture will start looking like a wall you just ran into in Quake. The actual texture used was 512 pixels tall, so it was way overkill, but it's the one I supplied with the prop because I didn't know how close people would render at (someone might want to do a 'laser-time' effect following the laser blast out of the barrel from an extreme closeup, for instance) so I gave it good leeway.

Wow! Thanks, Dodger! That's very helpful. I think I'll have to read this a couple times and play with Poser to make it sink in. Many thanks for your help! Tasquah - many apologies. My reply was meant to show my consternation, not any animosity. I thought your reply was meant to go over my head - obviously it did - LOL Now to consider and apply Dodger's advice. ram

No worries at all!

Most competent LCD screens nowadays come out at 110 to 125 dpi resolution (basically dpi is equivalent to a monitor's "pixels per inch" rating). I've seen high-grade images look nice on CRT monitors, but look absolutely bad-assed on a good LCD screen. Also, some of the higher-end CRT monitors (like my NEC Accusync 25" monitor) has 88x108ppi as standard.Dodger is right in that 72dpi will cover the vast majority of monitors, though... "Generic Monitor" in Windows assumes it as a matter of course, and unless you have special drivers, odds are good that "Generic Monitor" is the monitor driver most folks have installed... In my contributions to themes.org (a site for Linux graphics), I always use 124dpi in my final... you don't notice it in most CRT monitors, but those with LCD monitors can tell the difference very readily (X-Windows monitor drivers are sharp as Hell if you match it up to your make and model of monitor... that's why I much prefer doing my postwork in Linux.) /P

The smaller details you want the larger texture you need. You texture should have enough resolution to portray the detail level you want to show. The larger the final render is the more details will show and the more resolution you need. A lot of painted textures are much larger than they have to be, sins people think large size = quality, and sometimes it does :) If you want to print your image you should render it to the same size you want to print it in, where one pixel on screen = one dot on you printer, or you should interpolate the image to that size before printing. Max Printer rez * wanted size on paper = render size / interpolation size. You should calculate the axis ( x or y) most in conflict with the paper size so your printer doesnt print it on two pages instead of one. The DPI has no effect on you texture nor on the render, its just a number used to size the image when printing, but sins most modern printer drivers allows you to size the image before printing the DPI is not used. You can change the DPI to anything you like at anytime without affecting the quality of a render. (But dont set the dpi to zero as it freaks out some drawing programs). A X1000 * Y1000 image has 1 000 000 picture elements (pixels) regardless of DPI. Its better to downsize than upsize an image, so if you have the computer power then makeem large. When you downsize many pixels are merged into one, so they can give a better result than just rendering at the wanted size sins they portray a more real version of the image when you apply efx or do other things to the image before downsizing.

ramster I was interested in seeing what others had to say about your questions. Dodgers answer were great though i useally have to read them a few times to really digest them. I do want to add that a render should always be done in Tiff or BMP or PSD if you have it. Never in jpg , besides the fact poser jpg algorthms blow chuncks the loss ratio is very bad and the more you play with it the worse it gets. As far as size goes rendersing bigger than your eventual out put is always a good idea. Though i have come to undestand that resizing will decay a picture as well so to big and you will have some loss as well. How much depends upon a few things like what paint program you are useing. Adobe in my opinion does it the best. I have been told by experts to scale a 100 at a time and every 4 or 5 times add a tiny tiny amount of noise. The DPI thing is a issue in the poser form . Part of the reason is we all come from different backgrounds . Some of us worked with profesional printers and have always been told to submit things at 300- 600 DPI so we always worked in that but as you see from the answers its not nessisary as much as it used to be. I hope this helped you some or better than my original anwer :)

If you want to print your image you should render it to the same size you want to print it in Not sure about bubblejet printers which can basically put a dot almost anywhere, but this is not true of screened 4-colour printing and other commercial printing processes. For these, you want your image about 120%-150% (there's an equation but I do not know it, I just remember the ballpark since I don't want to open calc every time I save an image) the size of your output screen. And never use the same doppage or an even multiple of it. If you try them at the same DPI as the screen LPI or if you use an even multiple (i.e., using 200dpi for a 100 LPI screen), you will often encounter horrifying amounts of moire (weird patterns that are basically equivalent to a visual 'chord') As a note, Photoshop 5.5 and up has an image export feature in the help menu, I think, that automatically determines your optimum image resoution based on your target media (allowing for multiple screen resolutions and stuff). If you make an image large and follow its guidelines it will be just right. Unless you then stick it in your Quark doc and hit Shft-Ctr-Alt-Dot and zoom it up before printing it B^)

"If you want to print your image you should render it to the same size you want to print it in" that is true if you are using a "dye-sublimation printer" not an inkjet printer...

Thanks everyone for the advice. It was more than I had anticipated. I'll have to get my scientific calculator out now and some graph paper - LOL Seriously, Thanks!! This was a big help. ram

Don't worry about the specific equations -- just go 135% of the final target resolution and you'll be fine (except for web graphics, where you wan to go exactly the target resolution)

Yah Ram Most of us have been doing this for a bit so we have our own different sizes and what not. I post on the internet mostly on variations of 800 x 600 up to 1600 x 1200 and they both work for printing as well ( 8 1/2 X 11 and 5 x 7 ). but its what works for me depends upon what your end product is. Mine is internet or printed. The 135 % is right on the marks for final image. It can be shrunk to fit or slightly cropped to get a bit better enhancement one way or the other.

Thanks Dodger and Tasquah. My aim was to get the best render set-up for web galleries/pages. ram