Visual Effects and QuickTime Extension In an earlier posting to this list, I wrote in part:
> In Rev 1.0, the above effect is as smooth as silk. In 1.1B1, the >leading edge of the effect paints very unevenly and jerkily. >It appears to be painting in about 20-pixel chunks. Kevin replied: >I'm not getting this here. Is it possible that you have some other process >open that is taking up the CPU? Internet Explorer is really bad at this. >The effect seems to take about the correct amount of time here too now. Each time that I did this test, a single version of Revolution was the only open program. Since the above report, I did more testing to compare the visual effects in 1.0 and 1.1. I made a single stack with a single wipe effect and built standalones from it in both versions. I used this updated recipe: Create a 640 x 480 stack with two cards. Place a 600 x 400 dark blue square on the second card. Place a button on each card with this script: on mouseUp set the effectRate to 4000 --so that the effect is easily seen lock screen go next card unlock screen with visual wipe left very slow end mouseUp I went to computer labs and tried the 1.0 and 1.1 standalones on different Macs running different OS versions, and I made sure that the conditions were as equal as possible for the testing of the versions (no other open programs). In each case, on the same computer under the same conditions, the effects in Revolution 1.1 were quite inferior to those in 1.0. By inferior, I mean that the effects have a moving stair-step appearance along the leading edge as they paint. This has a ragged, jumpy look. In comparison, the effects in Rev 1.0 are quite smooth. When an effect is painting in Rev 1.0, the leading edge of the effect is very even--at the most, it "stair-steps" in about 1-pixel increments, and this appears very smooth. However, when an effect is painting in Rev 1.1, the leading edge stair-steps in approximately 20-pixel increments, and this is very noticeable. In one test, I checked the Revolution 1.0 stack on an approximately ten-year-old Mac. As expected, it did take a longer time for each card transition to start, but once it did, the visual effects were smooth. In fact, Revolution 1.0 on a 25 MHz Mac had smoother visual effects than Revolution 1.1 running on a 233 MHz iMac (running OS 9.0) or Rev 1.1 running on a 350 MHz G3 (running OS 9.1). Because version 1.0's effects ran so smoothly on such a slow computer, insufficient megahertz did not appear to be a factor. And because I had no other open programs in any of these tests, there was no interference there. That led me to look at the extensions. I found that with all extensions off, the visual effects in 1.1 were very close to those in 1.0. After some extension troubleshooting, it all narrowed down to the QuickTime extension. With the QuickTime extension off, Revolution 1.1's effects are smooth and nearly identical to those in 1.0. On my programming computer, I reloaded with the latest version of QuickTime to ensure that my copy wasn't corrupt, and the results were the same. In short, Rev 1.1 may be having some sort of adverse interaction with the QuickTime extension that impacts the visual effects--an interaction that does not occur with Rev version 1.0. With 1.0, the QuickTime extension can be on or off, and the effects are smooth. This problem first showed up in 1.1B1. Pre-1.0 betas and 1.0 itself were fine. Kevin, you mentioned that you were not seeing this at your end. I cannot explain this... I have been able to get repeatable results on different Macs. Offhand, I have only two ideas: 1. Might I have been using computers of less speed than yours--ones that more readily show these differences? I did these tests on Macs only up to 350 MHz, which constitutes the majority of my current target market. On these computers, the difference in effects between Rev 1.0 and 1.1 shows up very handily. I thought that I had ruled out megahertz as a factor, when, as I mentioned, I noted that Rev 1.0's effects ran smoothly on a 25 MHz Mac, while 1.1 looked inferior on a computer more than 10 times that fast. But I don't know how these tests would compare at today's cutting edge on, say, a dual 800 MHz G4. 2. Since I found that the QuickTime extension is definitely a factor at my end, could it be that my tests involved different versions of QuickTime from yours? In the early stages of testing, I was not keeping track of QuickTime versions, because I didn't yet suspect it. But in later tests during which I disabled the QuickTime extension and determined that it was involved, either version 4.0.3 or 5.0.2 was active. If neither of these was a factor at your end, and you are still not able to reproduce this, then a piece of the puzzle is still missing. After the Mac tests, I then built 1.0 and 1.1 standalones for Windows (on a Mac). I tested on a single Windows 98 computer. The Rev 1.0 standalone worked fine, but the 1.1 standalone would not display visual effects at all. For example, when using a four-second effectRate, four seconds would pass with no action, and then the screen would suddenly unlock with a straight transition-- no effect at all. To find out if any part of this was related to the Mac authoring environment, I built a 1.0 stack and a 1.1 stack, each in its respective Windows Revolution environment. With Revolution open and running in Windows 98 (these were not standalones), the 1.0 stack would display visual effects, but the 1.1 stack would not. Ted _______________________________________________ improve-revolution mailing list [EMAIL PROTECTED] http://lists.runrev.com/mailman/listinfo/improve-revolution
