What size/aspect ratio's should photos be for 16x9

PeterDuke wrote on 4/4/2011, 8:01 AM

In their wisdom, SCS has adopted a pedantic approach to picture aspect ratios for SD, which is out of step with most other people. That is why some people get black bars when they import images.

Q1:

16/9 = 1.78, but SCS uses 1.82, with a pixel aspect ratio of 1.2121, for NTSC wide.

480x1.82 = 873, so you should have your images at least 873x480 square pixels. (873/1.2121=720, giving 720x480 pixels as per the SD standard).

If you plan to zoom or pan, the images should be bigger so that in the worst case after the pan/zoom you still have at least 873x480 pixels in the frame.

PNG is best because there is no compression loss compared to JPEG, and TIFF has to be converted using Quicktime.

Q2 Choose NTSC wide screen profile if that is what you want.

Q3 pass

Former user wrote on 4/4/2011, 8:04 AM

You can use the native PSD format to save your pics.

Dave T2

craftech wrote on 4/4/2011, 8:30 AM

If you are grabbing a frame for a .Psd still make sure you have Full resolution rendering quality set to Best and Deinterlace method set to None in Project Properties.

John

Chienworks wrote on 4/4/2011, 8:36 AM

Way easier is to just let Vegas do all the thinking for you.

Create an NTSC widescreen project.
Drop the stills on the timeline.
Open up Pan/Crop on the first still, right-mouse-button click in the cropping frame and choose "match output aspect".

If you have more than one still, press Ctrl-C to copy, then select the rest of the stills, right-mouse-button click on one, and from the popup menu choose "Paste event attributes".

Note that i didn't mention a single number in any of my instructions, and you don't have to know any numbers at all. Vegas just does it for you. Why strain your brain when the computer already does the work?

rs170a wrote on 4/4/2011, 9:11 AM

1. Convert them from RAW to PNG.
This is an NTSC SD project so use a free tool like IrfanView to batch resize them to (approx.) no more than 1440 x 960. This assumes you're not doing deep zooms otherwise make them a bit bigger.
2. 1.2121 (NTSC DV Widescreen)
3. Let Vegas handle it.

Mike

Matt L wrote on 4/4/2011, 10:15 AM

You can add .PSD files into Vegas 10?

musicvid10 wrote on 4/4/2011, 10:18 AM

All versions of Vegas Pro import PSD files.
Vegas Pro 10 has the added advantage of layer support.

rmack350 wrote on 4/4/2011, 11:07 AM

Yep, PSD can be used on the timeline. That can be an advantage when you have text or vector layers because you can go back to photoshop and make changes, and Vegas will pick them up right away.

As for aspect ratios, Photshop's settings now match Vegas. Adobe got religion on this. Even so, don't use Photoshop's video templates for stills going to Vegas.

Kelly's suggestions are fine and allow you to avoid thinking, but it's actually good to have thought things through so that you understand what you're doing. So here's yet another version of what to do:

1.a: File types - Convert from RAW to PNG or PSD (PNG might be easier on Vegas than PSD but PSD has a lot of flexibility.) I've not tried to do batch actions on RAW through Photoshop but it might be possible. Let us know how that goes!
1.b Image size - You could just drop pictures on the timeline and then "Match output aspect". The one drawback is that high resolution photos might eventually hang Vegas when it runs out of memory. It's good to be conservative with your size at 2x or 3x project size. Or 1x if you never plan to resize.

Vegas automatically conforms square pixel stills to the project's PAR. In fact, its hard to get around this, so don't bother with Photoshop's PAR correction modes. They won't help you.

One way to find out exactly what Vegas needs is to set up an NTSC DV Widescreen project, put anything on the timeline, and export a still. That still will be the size that Vegas wants. It turns out that size is 873x480. (BTW, DON'T try to do math with frame aspect ratios. Things like 16:9 and 4:3 are nominal in the SD realm and will only screw you up. SD was an analog world and those frame sizes didn't include the entire width of the signal, just what the viewer should see of the frame.)

So if you must crop and resize try a multiple of 873x480, but you don't absolutely have to do this if you follow Kelly's (chienworks) instructions.

2: Just use the NTSC Widescreen profile.

3: PAR - Vegas always assumes stills use a 1.0 Pixel Aspect Ratio. Just leave it alone and don't do any squeezing of stills in Photoshop. Don't use Photoshop's video templates.

Rob

JJKizak wrote on 4/4/2011, 11:41 AM

Just drop them on the timeline like they said and don't worry about it. If your doing a 10 x 1 zoom keep them at least 1920 x 1920.
JJK

bdg wrote on 4/4/2011, 2:14 PM

In V10
Tick:
Preferences/Editing/ "Automatically crop still images added to timeline"

Then alls you have to do is add your png's to the timeline and they are automatically cropped to 16:9.
Saves a bunch of messing around.

PeterDuke wrote on 4/4/2011, 6:21 PM

"Just drop them on the timeline like they said and don't worry about it. If your doing a 10 x 1 zoom keep them at least 1920 x 1920"

If you start off with 1920x1920 and zoom in 10x then you will be left with 192x192 to put in your frame. Not good when SD NTSC widescreen requires 873x480 square pixels for full resolution.

You can start of with huge images and let Vegas worry about the maths but you will find that many large images makes Vegas (and other video editors too) unstable.

For best results and performance, get your hands dirty and put reasonably sized images on the timeline.

Matt L wrote on 4/11/2011, 7:28 AM

1. I set my .png files to 1620 x 1080p for pan/zoom purposes.
2. Should I set my luminance to 0-255?
3. When encoding 16:9 480p, should I also set: Inverse 3:2 pulldown?
4. Do I also select LetterBox when encoding?
5. I noticed some of my photos have a pulsing/flicker on the screen. I have flicker checked on each photo using Ultimate S from VASST. How do I fix this?

Also, do I sharpen my photos in Photoshop, before I create the .png file, or should I sharpen them using Vegas?

Thanks!

Matt L wrote on 4/11/2011, 7:29 AM

1. I set my .png files to 1620 x 1080p for pan/zoom purposes.
2. Should I set my luminance to 0-255?
3. When encoding 16:9 480p, should I also set: Inverse 3:2 pulldown?
4. Do I also select LetterBox when encoding?
5. I noticed some of my photos have a pulsing/flicker on the screen. I have flicker checked on each photo using Ultimate S from VASST. How do I fix this?

Also, do I sharpen my photos in Photoshop, before I create the .png file, or should I sharpen them using Vegas?

Thanks!

dennisWAP wrote on 4/11/2011, 8:52 AM

I have a question regarding PNG files. When you Save As, in Photoshop and you choose PNG file type, it asks you PNG options Interlace. Should you choose None or Interlace?

robwood wrote on 4/11/2011, 8:58 AM

i'll answer a couple of 'em...

2) "Should I set my luminance to 0-255?"
your photo stills should already have a range of 000-255. you *may* want to shift the values to 016-235 if you're rendering to or working in a YUV (broadcast) format rather than RGB (computer); usually i put all YUV or RGB media on separate tracks and assign RGB > YUV or YUV > RGB filters to each track.

3) "When encoding 16:9 480p, should I also set: Inverse 3:2 pulldown?"
i must have missed something. i thought the media involved was 29.97?
if it's 24fps, keep the project settings in 24, encode using one of the 24p presets in Vegas mpeg-2 custom settings
(they'll have a frame-rate setting that says "23.976 +2-3 pulldown")
if it's 29.97 use the "DVD Architect NTSC Widescreen video stream" or variant. *

6) "When do I sharpen/blur?"
general rule is
use Blur BEFORE scaling (helps reduce moires, buzz and other harshness that can happen when scaling down).
use Sharpen AFTER scaling (helps recover some crispness to line edges that scaling muddies up.

a popular tweak with Vegas is to use the Sharpen filter with a setting of 0.00...
this gives a moderate amount of sharpening with little damage.
and just to geek a little: sometimes i find even Sharpen of 0.000 too much so i made a softer version with Unsharp Mask

Unsharp Mask = Sharpen
set Radius to 0.001, leave Threshold at 0.000, set Amount to 1.25
(same as using Sharpen filter with Amount of 0)

Unsharp Mask = soft Sharpen
leave Radius at 0.001 and Threshold at 0.000, set Amount between .250 and .750
clean by eye or scope but with subtle lift to highlights/edges.
(a setting of 0.500 - 0.666 is probably the range i use most)

---

*caveat: u may have to gear down 24fps footage to 23.976fps to avoid frame-blending;
worst case scenario, you'll have to render a 24fps lossless master, re-import, change the Playback Rate to .999, then render as 23.976.
or you can change the fps setting in Project Properties and manually change the frame-rate for each clip one-at-a-time and make sure no gaps or offsets have occurred with the time adjust.

Arthur.S wrote on 4/11/2011, 11:29 AM

"What size should my images be set to in Photo-shop, and which format (jpg,png) should I use?"

In Photoshop there are templates for TV and media (HD 16:9 etc) under the 'new' dialogue.

JJKizak wrote on 4/11/2011, 1:01 PM

What I should have said was 1920 would be the minimum for a 10 x 1 zoom but that the quality would be marginal. It would be prudent if you knew the pictures you were going to zoom were 6000 x 6000 but most of the time you don't know untill you start messing with them on the timeline.
JJK

Matt L wrote on 4/11/2011, 8:11 PM

<< Should you choose None or Interlace? >>

I chose NONE.

Matt L wrote on 4/14/2011, 10:39 AM

<< 2) "Should I set my luminance to 0-255?"
your photo stills should already have a range of 000-255. you *may* want to shift the values to 016-235 if you're rendering to or working in a YUV (broadcast) format rather than RGB (computer); usually i put all YUV or RGB media on separate tracks and assign RGB > YUV or YUV > RGB filters to each track.>>

What is this filter called? All I see is Studio to RGB, and RGB to Studio.

musicvid10 wrote on 4/14/2011, 10:55 AM

Those are the correct filters.

Chienworks wrote on 4/14/2011, 11:19 AM

"In their wisdom, SCS has adopted a pedantic approach to picture aspect ratios for SD, which is out of step with most other people. That is why some people get black bars when they import images.

I think the confusion is based on people assuming that "normal" SD video is 4:3 and "widescreen" SD video is 16:9. This isn't true at all. Both formats are slightly wider than 4:3 and 16:9. The extra width is there to help allow for the old analog signals that had poor quality edges so there was some run-off room for the scanning beam that was hidden behind the front bezel over the screen. However, when editing full-frame these extra edges are visible and must be taken into account.

People mistakenly use those ratios interchangeably with the names "normal" and "wide". Real "16:9" didn't actually come out until HD 1280x720. However, since it was widescreen 16:9, folks started referring to SD widescreen as 16:9 too.

In this case, SCS Vegas actually has it exactly right. Any other NLE that shows ratios that work out to 4:3 or 16:9 is wrong.