-
stills in browser
-
Thank you, Bill. Interesting post and very relevant to the workflow that I’m trying to develop. Now question: certainly on a small screen there would be very little difference. But what if the exported project was destined for projection on a large screen? I know it will be exported as a 1080p movie, but would the image quality and size within FC make any different to the final product?
I took a look at the various programs that can actually process the raw files and have decided on Aperture. It’s reasonably priced. Lots of tutorials on how to use it. And I’m not buying in to the CC subscription plan.
Also I’m not finding any real problems with using raw files in FC but obviously I’m not reaping any great improvement using them either if the program is not able to process them. When I load the jpegs into my projects there is a slight difference in detail and exposure levels but it would be very similar to your post in terms of a side by side. And once they are exported to a source they are compressed and really I can’t tell the difference.
That said, if i were to process the raw shots in Aperture what do you feel is the sweet spot for exporting to FC? What about a photo slide show on a large screen? Or an 11 x14 print?
Still don’t get what the [Bill Davis] “pixel raster density 300dpi”would have to do with screen resolution. I use these dpi settings when printing but have not heard them discussed in terms of video editing. For an accurate side by side, shouldn’t they be set to the same settings? I guess I’ll learn more about the export choices when I start playing with Aperture.
Mac Pro, macbook pro, Imacs (i7); Camcorders: Panasonic AG-HPX170/AG-HPX250P, Canon HV30/40, Sony Z7U, VX2000, PD170; FCP 6 certified; write professionally for a variety of media; teach video production in L.A.
-
Pngs has benn supported as part of the HTML standard for what, 5, 8 years? And so OS systems and browsers have to support it. Someone would have to be running XP I’d think to have a problem.
-
[Craig Alan] “But what if the exported project was destined for projection on a large screen?”
In projection the size of the screen is almost totally irrelevant.
The projector is putting out a fixed resolution determined by it’s display chip raster (and the projector resolution settings.) The apparent “size” of the screen projection is only affected by type of lens being used to spread and focus the image and the distance of the projector from the screen.
Adding projector to screen distance yields bigger and less bright images, but the image itself has the exact same resolution. Again, that’s fixed by the imaging chip in the projector.
As to the raster density – in the SD era, the typical 640×480 computer raster of the past had a equivalent pixel density of 72dpi. Which is to say if you created a screen that size and at that raster, you’d completely fill the typical TV Screen with a computer image top to bottom, edge to edge.
If you created an image at the exact same size, but with a higher pixel density – say 150 dpi, You would have twice the horizontal resolution and could “blow up” the image twice as wide and twice as high – which yields a 4x zoom at the same screen resolution as the 72dpi image mapped to the same array of pixels. (vertical resolution is more complex since it deals with interlace issues and scanning frequencies, so that’s another topic for another time!)
Now it gets more complex as we move from SD to HD to 4K and beyond, but the relative pixel math stays the same. In a bit mapped raster (which is what we’re discussing – NOT a vector world which works differently) you can increase the overall raster size, OR the pixel density and achieve the same thing – because you’re mapping that pixel array to a presumably fixed screen display setting.
If you fix your still photo raster DIMENSIONS at say, 1280×1024 – then exporting two files, one at 100 dpi, and one at 400 dpi – would net you very different results when you scale up the images in your NLE. It would also yield different file sizes.
The key is to always think of mapping an incoming raster to whatever the raster you’re likely to set your screen at.
Thats the practical process. You have to determine (or at least make a sensible guess at) what raster you’re mapping TO – or you can’t make good decisions about what export raster you’re going to need out of your graphics program in order to zoom into the image in your NLE without significant degradation.
It’s just how pixel arrays work.
It’s tempting to just make your rasters big – imagining that some day you might want to output your work to 4k – but as my photo array shows, it’s super easy to increase the size of pixel rasters and blow up your file sizes WAY beyond what’s required for your actual output – and when you do that, you can bog down your entire production pipeline, since simultaneously increasing both horizontal and vertical (or pixel density!) factors has a massive and escalating effect on increasing file sizes.
I’ve seen countless folks take DSLR full raster files, dump 50 of them into a timeline – and get upset when it takes FOREVER to move them around.
Of course it does. Particularly in X which is built to calculate with and maintain the maximum amount of source material resolution in all it’s image manipulations.
If you work with stills a lot – it’s really important to test and test and test things – rather than just import willy nilly into X.
These days camera native rasters are MASSIVE and ever increasing. And people are messing up right and left by not understanding the basics of how raster math works and how you can unthinkingly create massive file size log jams very, very easily.
BTW, I’m no expert at this stuff. Just a guy who’s been in the trenches for decades reading and trying to figure out what doesn’t work. Anyone who’s got better experience or theory to share, dive right in.
FWIW.
Know someone who teaches video editing in elementary school, high school or college? Tell them to check out http://www.StartEditingNow.com – video editing curriculum complete with licensed practice content.
-
[Bill Davis] “BTW, I’m no expert at this stuff. Just a guy who’s been in the trenches for decades reading and trying to figure out what doesn’t work. Anyone who’s got better experience or theory to share, dive right in.”
You might want to check out other “theories” about dpi in video, e.g. this from Rich Harrington:
https://www.richardharringtonblog.com/files/026a3349ffd567f3cb9b30742914c51e-913.php
Simon Ubsdell
http://www.tokyo-uk.com -
[Bill Davis] “If you fix your still photo raster DIMENSIONS at say, 1280×1024 – then exporting two files, one at 100 dpi, and one at 400 dpi – would net you very different results when you scale up the images in your NLE. It would also yield different file sizes.”
Pixels are data — the actual little dots that make up the image. Pixels have no inherent size themselves. Physical image size is irrelevant since different displays have different physical sizes. A 1920×1080 display could be 10″ or it could be 60″ on the diagonal, but they are both 1920×1080.
DPI, or dots per inch, is metadata. Since pixels do not have any inherent size, but print does, DPI metadata tells a print system how to relate the indeterminate size in pixels to the intended size on the page.
Since DPI is metadata, not image data, you can change DPI without changing the actual image itself. If you have a 100×100 image at 300 dpi, its nominal print size is 0.3 inches square (100 dots divided by 300 dots per inch). You could change the DPI to 72, and that would change its nominal print size to 1.38 inches square (100 dots divided by 72 dots per inch). These two different print-size images are really identical, and are only 100×100 pixels in Photoshop, FCPX, Pr, Ae, or any other app that measures in pixels instead of inches.
Print people ask for high-DPI images because they are thinking in print size, not in pixels. A 6-inch image at 300 DPI is 1800 pixels (300 dots per inch times 6 inches), whereas a 6-inch image at 72 DPI is only 432 pixels (72 dots per inch times 6 inches). This matters a great deal for print, but again because print considers physical size and video does not, this is irrelevant for video.
Where it gets confusing is that some apps like Photoshop have resampling turned on by default when changing image size. This means that when you adjust the DPI, Photoshop will also scale the image to preserve print size. If you uncheck resampling, then you’ll see that changing the DPI without scaling the image keeps the same resolution (pixels) but change the print size (inches).
TL;DR — for video, pixel resolution is all that matters. DPI is a meaningful measure of resolution for print only because they think in print size, not in pixels. DPI is irrelevant for video because while pixels are the “dots” in DPI, video has no notion of physical size (the “per inch” part). Some resizing tools have an option that uses DPI as a scaling factor, but disabling this to avoid resampling shows the true relationship between resolution and DPI.
Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog – What I’m thinking when my workstation’s thinking
Creative Cow Forum Host: Live & Stage Events -
[Bret Williams] “Yes, images of that size are very taxing on FCP. Since they’re probably 10x the file size of a jpeg, they require more memory and processing power. Because they’re twice the resolution they need to be, they require more memory and processing power.”
In general, I have found that when using oversize stills in FCP X, the best route is to make them into compounds before trying to do anything else with them.
This seems to speed up any image processing you do to them as well as creating a frame that is the “expected” size for any transform filters.
Simon Ubsdell
http://www.tokyo-uk.com -
[Walter Soyka] “Some resizing tools have an option that uses DPI as a scaling factor, but disabling this to avoid resampling shows the true relationship between resolution and DPI.”
Thank you Walter. That’s much clearer.
I suspect my notions of how DPI relates to scalability came from the old days when this scaling connection was common.
I have in my thinking lots of experiences from my learning years where increasing the DPI would yield a larger raster for scaling purposes. Thanks for clarifying the modern way things work.
(And on a personal note – happy new year – hope we get to see each other, perhaps at NAB this year?)
Know someone who teaches video editing in elementary school, high school or college? Tell them to check out http://www.StartEditingNow.com – video editing curriculum complete with licensed practice content.
-
[Craig Alan] “I’d love to add some of the touch up ability of photoshop to my edits. Not really wanting to go that cloud route.”
As I mentioned before, Pixelmator is a good cheap alternative to Photoshop which includes probably all the features you’d need for touch-up for video.
Its Curves (identical to Photoshop) are a very powerful way of colour correcting and you have Levels as well. You can do all the expected dodge and burn, clone, blur, sponge, patch healing, sharpening, etc, etc, that you could want. And of course you have Blend Modes for elaborate compositing type enhancements.
I’d be surprised if you didn’t find everything you needed there. Plus there’s a free 30 day trial so it’s well worth giving it a run for its money.
https://www.pixelmator.com/try/
(Having said that, I don’t use it much myself and certainly have no vested interest in it, and I still prefer Photoshop for the really super-tricky stuff.)
Simon Ubsdell
http://www.tokyo-uk.com -
Thanks Simon,
I looked into Pixelmator by reading reviews and checking out their site. As well as other apps. I went with Aperture. Aperture not only gets good reviews specifically about processing raw files but for its organizational ability. And it interfaces with iPhoto and can share their libraries. Its a pro app and was only $80.
Mac Pro, macbook pro, Imacs (i7); Camcorders: Panasonic AG-HPX170/AG-HPX250P, Canon HV30/40, Sony Z7U, VX2000, PD170; FCP 6 certified; write professionally for a variety of media; teach video production in L.A.
-
Reply to this Discussion! Login or Sign Up