Pre Loading Graphics

Hi Richard

I am doing a lot of work where still images are used as part of a corporate presentation and as sponsor loops i.e. crossfade between different company logos.

Clients love that I can have a moving background as the graphics change in front. The only problem is that there is a noticable jitter in the video playback as each new image loads up.

The jitter obviously gets worse the larger the file size but even the slimmest jpegs cause some form of jump.

My machines have 1.5 gig of Ram and Raptor drives for content.

Is there a possibility of a preload feature or can you suggest a possible work around.

Cheers

Toby Harding

I have experienced the same. What i did was to convert them to quicktime files, to avoid the "hickup" that occurs when loading pictures.

The catch to converting to quicktime is when you want to use the transparency of a photoshop file or you want the high res of a jpeg.

Most good quality image files are much bigger than movie files. It takes a palpable amount of time to:
a) Load the image into RAM
b) Decode

The way around this is to convert your stills in to Photo-Jpeg Quicktime movies. This is actually what the Photo-Jpeg codec was designed to do in the first place. The advantage of using Photo-Jpeg over DV, is that you are not restricted to using Standard Definition. Photo-Jpeg will successfully encode movies at whatever size you set. The compression ratio is very impressive for the quality you get. You can then use Inframe to select the relevant still.

What about transparency?

Also the beauty part of using image files is that it allows for quick updates when the client throws that last minute graphic at you.

I have considered the Photo Jpeg option before but requires a little more pre-production than I have time for sometimes.

I was thinking more of a way of nominating a folder to be pre loaded into ram that you might include as part of a mark cue. Obviously you are limited by how much ram you have installed.

get yourself a copy of cleaner, or even a copy of quicktime pro, outputting the still image to a photojpeg quicktime file wont take you more than a minute or two more.

Transparency is always a problem, as only a few image codecs support it.

The Animation codec for movies, TIFF files, photoshop files for still images.
There may be others.

The problem with transparency is that it is done with an uncompressed alpha channel.

So this kills performance.

I have used the animation codec on real shows with graphical image movie files - with some success - but animation codec files can give quite poor performance.

----

Jitter, you really really have to convert your stills to something more manageable.

Still image files, tend to be much larger than movie frames - and take a long time to load.
the hard disc in a computer can only do one thing at a time....

Thankyou all

I will do a bit of experimenting

Cheers

Toby

Quote:

---

Originally Posted by tharding

What about transparency?

Also the beauty part of using image files is that it allows for quick updates when the client throws that last minute graphic at you.

I have considered the Photo Jpeg option before but requires a little more pre-production than I have time for sometimes.

I was thinking more of a way of nominating a folder to be pre loaded into ram that you might include as part of a mark cue. Obviously you are limited by how much ram you have installed.

---

This doesn't seem too practical to me. Catalyst specifically does not use buffering techniques, so adding more RAM doesn't make any difference.

I tend to use .psd files for transparency stuff. If you have a nice Ultra 320 SCSI RAID array, then using mulitple layers of Animation codec movies becomes a possibility, and Animation supports an alpha channel.

Hugh

To get transparency to work, you need to render your animation files with 'millions of colour +', the '+' means it adds the alpha channel

Quote:

---

Originally Posted by Spam Butterfly

This doesn't seem too practical to me. Catalyst specifically does not use buffering techniques, so adding more RAM doesn't make any difference.

I tend to use .psd files for transparency stuff. If you have a nice Ultra 320 SCSI RAID array, then using mulitple layers of Animation codec movies becomes a possibility, and Animation supports an alpha channel.

Hugh

---

using a ram disc, is different from using more ram.

ram discs might be useful.

i might add buffering strategies in the future. the problem was- how to manage them, and not spend minutes loading up images....

I am used to using the Macromedia Director preload function which can take some time but if you plan for it always meant a sequence ran smoothly.

I did my show last night using the Animation Codec file as suggested and it was a vast improvement. Using maximum resolution as I was projecting across multiple screens there was only the slightest perceptible halt as I changed frames. Most times I didn't even notice once we were into the throws of the show.

This was a major difference form using the Photoshop files that I had in the past. Also Drop shadows work so much better via Final Cut as you can set the Alpha type.

The workflow of compiling still graphics via Final Cut was fine and is almost as quick as just dumping Photoshop files into a library folder.

Thankyou for the help and advice!!

Cheers

Toby

Quote:

---

Originally Posted by tharding

I did my show last night using the Animation Codec file as suggested and it was a vast improvement. Using maximum resolution as I was projecting across multiple screens there was only the slightest perceptible halt as I changed frames. Most times I didn't even notice once we were into the throws of the show.

---

What resolution did you render your animation files at?

Do you know how big the file was, and how big each frame was?

R

The longest dimension of any frame was 2048 - more importantly in width as I was projecting across 2 screens. All of the Logos I was using were landscape. I find that if I go with the highest resolution it gives me more versatility with slow zooms and pans across the image.

Some files were the full 2048 x 2048 whereas others were 2048 x 768.

A quicktime with 18 different images in it was only about 6-8MB with each logo taking up the default 12 sec. This was far more efficient than the original photshop files with their layering. The main thing was having to do a save as with layers switched off in Photoshop if you want to keep layer styles like drop shadows.

Cheers

Toby

those are pretty huge...

Richard

Just wondering if you had a chance to experiment with the Ram Disk at all?

I tried it on my system and managed to get 7 layers all running at 25 fps. The only reason I didn't try 8 was that I only created a 512 Mb Disk (I have 1.5 gig on my machines).

Obviously the biggest bootleneck is going to be the 8gig limit on RAM in the G5 but still, the ability to place some content on a RAM Disk can help. Especially if it is a consistent background file or the large graphic files I mentioned earlier in this thread.

To have this functionality built into the software would be awesome.


Cheers

Toby