flickering

frustrated with Sony wrote on 11/15/2009, 11:26 PM
From what I read on other posts a high resolution causes flickering to occur? i.e. I did Pedigree charts and have lots of face photo's in Photoshop Slideshow. Then saved at 10 resolution. When brought into Vegas I got immediate flickering which I resolved with Reduce Interlace Flicker PLUS a Gaussian Blur. But, here is my discrepancy. When I brought the rendered movie into DVD Architiect, the flickering returned big time and nothing (even Reduce Interlace Flicker could resolve it.)

Question: the suggetion on the forum is to take the slideshows into a this "Infranview" program (which I gather must be bought) to resize as a large batch file OR go back into Photoshop and reduce the resolution??

Comments

Sierra Nomad Photography wrote on 11/16/2009, 12:09 AM
Don't blame Sony. I discovered this problem when I was using another NLE, so I think it is universal.

I don't know about the recommendation that you read; but here's what I finally came up with (based on a lot of advice from others and hours of personal experiments):

Resize your photos to 720x480 (if NTSC) and save at 300 dpi. That's it. My experience was that this eliminated the flicker in about 90% of the images; and in the rest the small amount of flicker that remained could be taken care of with a slight blur.

Hope this helps.
TOG62 wrote on 11/16/2009, 3:34 AM
Actually, I don't think dpi is relevant, as this only affects printed output. If you intend to use any pan or zoom (Ken Burns effect) you'll need the images to be rather larger.

Mike
musicvid10 wrote on 11/16/2009, 5:09 AM
Are you reporting flickering in the DVDA Preview, or on a burned DVD? If the latter, what kind of player and screen are you viewing it on?

Have you tried rendering in Vegas at Progressive/Best? What are the results?
Sierra Nomad Photography wrote on 11/16/2009, 9:50 AM

Re: the DPI. I don't understand why it made a difference; but in my tests it did. I compared 96 dpi with 300.

Many that were flickering badly at 96 were solid as a rock at 300.
TOG62 wrote on 11/16/2009, 10:11 AM
Re: the DPI. I don't understand why it made a difference; but in my tests it did. I compared 96 dpi with 300.

Even at the same image pixel size?

Mike
Sierra Nomad Photography wrote on 11/16/2009, 6:23 PM
Yes, all were 720x480. Also, Iin the test, I had about 20 images that I prepped both ways. Had the "same" image play back to back so I could really compare the differences.

Someone on Ulead's forum recommended it; and I thought I would notice little or no change - but the improvement was dramatic.
Former user wrote on 11/16/2009, 7:21 PM
To clarify,

Your images started at 720 x 480 at 96dpi,

You converted to 300 dpi, but kept the images at 720 x 480.

Is that right? Cause normally when you change the DPI in most photo programs, they will automatically scale the image to a higher resolution unless you change the resolution back to 720 x 480.

Otherwise, I find it hard to believe you would see any difference because DPI has no use in video. But I will test this just to see what you might be seeing.

Dave T2
Sierra Nomad Photography wrote on 11/16/2009, 7:40 PM
Both sets of images, when completely prepped, were 720x480. The only difference was the dpi. :)

In prepping both I started with the original file of about 2,000x3,000. I copied the folder so that I had 2 of each and batch processed each folder.

Former user wrote on 11/16/2009, 8:05 PM
In theory there should be no difference.

But my guess is the quality difference has something to do with how the software you are using is scaling the image down. and this is just a wild guess, but it might make a difference if it changed the DPI first and then set the resolution as opposed to setting the resolution first and then changing the DPI.

What software are you using to resize your images?

Dave T2
Sierra Nomad Photography wrote on 11/17/2009, 7:08 AM
Photoshop CS.

I don't know. Like I said, I can't explain it. Maybe there was something unrelated that affected the results. But that's my standard prep for stills now.
MPM wrote on 11/17/2009, 11:39 AM
FWIW...

The image you see in each frame has to change in order to see flickering. There are 2 things that can cause it to change on a frame by frame basis --> panning/zooming, & interlacing.

RE: Panning/zooming, the video format & any restricted band width makes it [flicker, stutter etc] worse, but the cause is the encoder -- with mpg the encoder finds changes in the frames looking ahead in the stream, then records those changes in P & B frames... with this type of motion some do better than others.

RE: Interlacing --> the nature of the beast is each field (2 per frame) is incomplete, with no data every other horizontal line... it's like looking through partly opened blinds, where you can see through the gaps. At 50 or 60 fields per second [PAL/NTSC] the phosphors in the picture tube are supposed to have enough of an afterglow that, combined with our vision [persistence], we don't notice. In practice, any stationary horizontal edges &/or lines need to be faded over more than one scan line.

When it comes to software including DVDA, handling interlaced video can & does sometimes get muddled. The display you're looking at has an effect, as do the graphics drivers since they may auto-convert interlaced to progressive (i.e. the absence of interlacing) -- how &/or if this is involved depends on the app. The source format when it's interlaced can matter, as some conversions can botch the job, and having the wrong field order set can be disastrous. And lastly, the software itself though trying to be helpful, can botch the display &/or do a poor job of trying to split progressive frames into fields, sometimes shifting the image by a scan-line or more.

DVDA's anti-flicker switch can help. You can also render your stills as 24p with pulldown, same as film-based video. And/or you can render your video as progressive. When playing a std Video DVD, stand-a-lone players are going to send an interlaced signal to interlaced TVs... software players & often graphics drivers are going to try & convert interlaced to progressive. Either type of player handles progressive video just fine, or there wouldn't be too many std video DVDs sold retail. ;-)

DVDA doesn't have a lot of control when it comes to encoding, so if you have problems, encode elsewhere & import the video. That said, for menus past discussions have been divided with some folk getting better results with progressive, others with interlaced -- personally I go with the source.

When it comes to stills, if/when they're going to be seen on a regular TV screen, colors have to be tweaked, as bad things happen if the TV can't handle them, even leading to audio problems. Sharp lines & edges are OK interlaced, ***as long as they're perfectly vertical***. If you've got anything other than full color photos, objects need to be faded -- text & similar should have both fading & shadows. Personally I prefer to handle any image mangling myself, rather than have video-related software try to be overly helpful & botch the job...

Adobe has their recommended *pre-video* still sizes in their knowledge base. Sony apps I believe process stills a bit differently. Vegas does it's mangling, then the MainConcept encoder mangles stills a little bit more -- do not *proof* any stills until they've been encoded. But before you're tempted to get too precise -- especially with 16:9, the DVD player (software & hardware) displays how it sees fit, as will a std TV. Take a picture of the display & compare if you wish -- I have, & you might be surprised how much it can vary from the original.

As far as still image resolution goes... it's a difficult subject that can maybe be compared to an infinitely flexible bean bag chair. It can be spread out wide with little depth, or narrow & deep [tall], but either way it has the same number of pellets (i.e. pixels) inside. Jpg images are harder to deal with, as different software interprets different file's rez differently. At any rate, when you mangle an image by re-sizing it, software or hardware has to make an educated guess on which pixels to toss &/or which ones to add -- sometimes you're even dealing with sub-pixels. More data [like chocolate] is better. That's why many photographers take pictures with much higher rez than they'll ever use -- or why recording studios will work with much higher rez than the CD format wav files they'll eventually produce.