Generational loss reduces but does not eliminate the greater contrast range of film. Today's contact printing methods can produce excellent results when done properly. I've seen film and DLP versions of the same movie and come away 100% convinced that the film print had much better contrast, and that was with an ordinary 4th-gen release print.
Scratches are not a normal part of film projection. A film can have hundreds or even thousands of showings, and with a good projectionist and good equipment, it will still look brand new afterwards. If your're worried about dirt build up, wet-gate projection can be used to help prevent that (yes, there are theatres that use wet-gate on a regular basis).
Virtually every problem that people point to in film presentation has a solution that does not require replacing it with digital video projection.
Does that surprise you? Your response has a very defensive tone.
I can't just take someone's word for it. I've seen 35mm dye transfer prints, and I've seen DLP several presentations, and going purely by eye, I'm in doubt as to which has a bigger color gamut. Actual measurements are neecessary.
Anyway, thanks for the link -- that's the sort of thing I was looking for. I see that they can approach the color gamut of film, even beat it in some areas, but they are only comparing themselves to an ordinary film print. There is no mention of dye-transfer. That's too bad. But at least it's a start.
By the way, there is no mention of 1920x1080 in that presentation. Are you certain they have finally achieved that resolution?
The color gamut available surpasses that of 35mm film.
Do you have any measurements to back that up? That is something that I've tried to get real, hard data on, but I cannot. The data is available for resolution & contrast ratios, but not color gamut.
By the way, you talk about jitter and weave, but have you ever seen a Kinoton "E" series projector? Instead of a mechanical intermittent, they use electronic stepper motors to do the film pulldown. They produce very steady images with jitter & weave well below that of standard projectors. Many people have only seen cheap, belt-driven projectors from Cristie, and incorrectly assume that that is the best film has to offer. It can actually be much better.
Modern print stocks have a contrast ratio of 10,000:1. Most video projectors are only able to get a ratio of 1,000:1, sometimes 1,100:1. That's not as big a difference as it sounds, since we perceive light logarithmically, but it is still a visible difference.
Film wins when it comes to contrast. It won't always, but it does for the moment.
Current DLP systems max out at 1280x1024. Think about that for a minute.
1280x1024 on a movie sized screen! I've seen it on several occasions, and let me tell you, the pixels are huge. It's ugly. If you've seen digital projection, you're not "lucky".
The resolution of digital projection must be improved if it is to be taken seriously. 1280x1024 doesn't cut it. There are rumors that TI might finally inch their way up to full HDTV resolution (1920x1080) by the time Ep. II premeires, but even that isn't good enough.
Movie special effects are always done at at least 2K resolution, sometimes 4K. There was an article linked from here not too long ago which quoted some people who were working on FX for LOTR that some of their shots were done at 4K resolution. What good is that if movies are going to be projected at a puny resolution of 1280x1024?
Fortunately, there are more technologies on the horizon besides DLP. JVC's D-ILA is already at 2K resolution, and may reach 4K before too long. Sony is working on laser projection which should also be able to reach 4K.
The bottom line is -- don't be in a hurry! They should take the time to do things right, and make sure theaters are not stuck with equipment that is obsolete overnight. It's a good thing that there are currently only 20 digital screens in the US. The technology needs more time to mature first!
Yes, I've found that Kodak film pretty much sucks for color. Many years ago, I started using Fuji exclusively, and have never looked back. I'm very happy with the color I get from it.
Damn, I could have sworn there were some instructions in there that operated on two 64-bit doubles, but I see now that the best AltiVec can do is to operate on four 32-bit floats. Bummer.
Actually, the only thing they've mandated is that broadcasts become digital. They've said nothing about resolution. And since there are modes in the ATV standard for "Standard Definition" (NTSC-like 480 pixel tall images), that's what broadcasters are going to use.
Boy, would I love to have movie channels that show their stuff in HDTV coming in over a regular cable TV hookup, but it ain't gonna happen.
History has shown that given a choice between transmitting the same number of channels at higher quality and transmitting a larger number of channels at the same quality, broadcasters will choose the latter every time, because they make more money that way.
We will never, ever see widespread HDTV in the US. We'll be stuck at NTSC resolution for the rest of our lives. Heck, if they could convince people to live with 100x100 digital video streams, they would, just so they could squeeze even more channels out of the same bandwidth. They drool at the idea of 50 million channels of shopping and other crap. Picture quality? What the heck is that?
"We know that Photoshop runs better on the G4, what about your operation?"
If it can be optimized for AltiVec, almost nothing will be faster than a G4.
Just take a look at these RC5 stats (mid-way down the page). G4s smoke everything, because the RC5 client is optimized for AltiVec, thus it can compute four keys in a single clock cycle. By comparison, Athlons do one key per clock cycle, and Pentium 4s do one key every four clock cycles.
So if you've got an operation that can benefit from the G4's SIMD capabilities, Macs are your best bet.
How about one of those iRock FM modulators? Small, light, and battery powered. You can get 'em for $30 at CompUSA. They sound pretty good, better than the old cassette adapters. One guy I know swears it is actually better than the expensive modulator that hooks from his CD changer directly into his stereo's FM antenna lead.
So does this mean they are finally going to take down the Quake 2 stuff from the front page of their website and replace it with something a little more recent?
Michael, if you want people to stop being suspicious of you and the other editors, you're going to have to make the site more open.
That means making things so that you no longer have to "guess" what percentage of moderation is being done by the editors, nor having to "think" how many mod points you've spent in a given day.
Make that information publicly available. Have moderation history done by the editors listed in their user info. And fix the notification so that it makes a distinction between moderation done by editors and non-editors.
Carbon is more than just a 'stop-gap' measure, it is a full featured API set. And like you said, it's to port to/from it than with Cocoa.
Also, the Carbon APIs are actually a lot more robust when it comes to error handling. Nearly all of the Carbon APIs return meaningful error codes, so that your app can do the right thing in case of failure. Cocoa makes poor use of such mechanisms, opting for a simple boolean pass/fail in the best cases, and nothing at all in the worst. It also fails to take advantage of language-level exceptions for error handling. Basically, if something goes wrong in your Cocoa app, you're screwed when it comes to getting enough information to actually do something meaningful about it. And some of the newly developed software that is inherited from NeXT is even worse. CoreFoundation will simply call abort() on your app in many cases. How rude!
I know Apple's marketing material hypes Cocoa and downplays Carbon, but in the end, apps are better off with Carbon. Even major developers who are new to the Mac are choosing to use it (i.e., Maya).
Besides, a lot of the most interesting stuff in Mac OS X is down below the Carbon/Cocoa level. Quicktime, ColorSync, ATS, etc. are all in the Application Services level, and are thus accessible to both Carbon and Cocoa apps.
Actually, Carbon and Cocoa live at the same level in the system, as shown by this diagram.
Cocoa requires an Objective-C runtime to work, but Carbon does not, as it is just a library. A little known fact is that both APIs allow the use of InterfaceBuilder to layout the GUI elements. Carbon has added support for.nib files in addition to the old-style resource manager stuff. And the new Carbon event structure is actually very nice.
Re:Bad screenshots for showing anti-aliasing
on
KDE 3.0 Screenshots
·
· Score: 1
Not only that, they didn't even save their JPEGs in progressive mode, so they don't give you a good overall view while they are loading.
<RANT>
Why the hell don't people use progressive mode for their JPEGs? Just look at this list of browsers that support progressive JPEG. Just look at it! Do you know anyone who is still using a pre-Netscape 2.0 browser?! Me neither. Must we stay in the stone age forever? Sheesh. People who insist on non-progressive JPEGs are almost as bad as those people who go into conniptions when you send them non-ASCII e-mail.
I'm pretty sure the previous poster was not talking about the e-mail exchange with apex, but instead was talking about this, which really does look like a bad attitude. Flaming someone in all caps because they forgot to include the version number in the bug report is just plain out of line.
Re:Will we have to revise unicode?
on
XML for Ancients
·
· Score: 2
That article is complete crap. I can't believe anyone takes it seriously.
The author of that article doesn't seem to understnad the fact that Unicode is a character set, not a font. He also doesn't seem to understand how Unicode's surrogate pairs work (which allow for encoding of more than 1 million characters). He doesn't seem to understand that Unicode is an evolving standard (i.e., 3.1 is hardly the final version). And he doesn't seem to understand that UTF-8, UTF-16, UTF-32, etc. are all just different formats, and they actually represent the exact same character set.
But most importantly, he is flat-out wrong about how and why the decisions were made regarding encoding of East Asian languages. He needs to learn about the history of Han unification for CJK characters. If he did, he would know that linguists and computer scientists from East Asian countries have been involved in Unicode since the beginning. The unification of East Asian characters was done on purpose, and has the full support of linguists, scholars, and computer scientists from those countries.
If the author of that article had just spent a few minutes reading the a copy of The Unicode Standard, he would not have made those mistakes. He didn't even have to read the whole thing! Just the Introduction and Appendix A would have set him straight on the issues I just mentioned. The fact that he didn't means this guy really shouldn't be doing work for a company with the word "Research" in the title.
Oh, and even though that page says the article has not been modified since June 4, you can see from the google cache that they have since removed their promise of responding to criticism.
And one more thing: Since he derides those mean old Westerners on the Unicode committee for being insensitive towards the peoples of East Asian countries, perhaps he should ask himself if it is considered impolite or insensitive to sweepingly refer to such peoples as "Oriental", which he does in the first few paragraphs.
Taco, I've been reading since the site was run off the server you were adminning at work, and had expectations consistent with your scale of operations. But if you're implementing paid subscriptions, you might also want to apply some of the standards normally expected of professional journalism.
Even if he's not implementing paid subscriptions, he should still apply professional standards, considering that he is, by definition, a professional (i.e., he's getting paid to do this).
I also see that the story has now been changed to remove the inflammatory statement, yet there is no "UPDATE" tag or other indication that it has been edited. That is even less professional. Ugh.
Yeah, yeah, it's a free site, blah blah. Nearly all corporate news sites are free too, but if any of them did something like this, people would be yelling, no doubt about it.
Yes, people should use VPNs, but the point of the article was that they're not.
Actually the article does not mention VPN even once.
It mostly talks about encryption, which is pretty useless on wireless networks. It stops only the most casual snoopers (like those "hackers" described in the article). Using WEP doesn't mean your network is secure.
It's hardly a secret that your laptop will see something when you're standing out in the parking lot near any company with an 802.11 network. That doesn't mean it's insecure. A company with even a smidgen of security sense will put the wireless network outside their firewall, and require employees to use VPN to access internal stuff. People on the outside may be able to get a little free internet access, but that's it.
The article is very light on details, gives no information as to what "wide open" means (just because you can see the network, that does not mean it is insecure). There is only one mention of the word "firewall" in the whole thing, and even then it's very vague.
I think this reporter has been duped by a couple of script kiddies. The supposed terms "war driving", "war pedalling", and "war walking" sound like something the kiddies made up on the spot, and later snickered at the reporter for believing.
The installer script was in the.pkg file, hence the words "buggy pkg file" in my above post.
You started this thread by saying iTunes shouldn't need an installer, but several people rightly pointed out why one is necessary. You then implied that putting the system-level frameworks iTunes depends on into a separate installer would somehow have prevented the problem, which it wouldn't.
You finally got something right by saying it was a bug in the install script, but you still don't seem to grasp the facts that 1) The install script is necessary and 2) Making it a separate install from iTunes would not prevent bugs in said install script.
It seems you're the one who wasn't paying attention.
Almost every movie comes with at least one trailer that is physically attached. The difference in this case is that the theaters were forbidden from intercutting their normal logo trailer between it and the movie, which they are normally allow to do.
Slashdot Mirror
Scratches are not a normal part of film projection. A film can have hundreds or even thousands of showings, and with a good projectionist and good equipment, it will still look brand new afterwards. If your're worried about dirt build up, wet-gate projection can be used to help prevent that (yes, there are theatres that use wet-gate on a regular basis).
Virtually every problem that people point to in film presentation has a solution that does not require replacing it with digital video projection.
Does that surprise you? Your response has a very defensive tone.
I can't just take someone's word for it. I've seen 35mm dye transfer prints, and I've seen DLP several presentations, and going purely by eye, I'm in doubt as to which has a bigger color gamut. Actual measurements are neecessary.
Anyway, thanks for the link -- that's the sort of thing I was looking for. I see that they can approach the color gamut of film, even beat it in some areas, but they are only comparing themselves to an ordinary film print. There is no mention of dye-transfer. That's too bad. But at least it's a start.
By the way, there is no mention of 1920x1080 in that presentation. Are you certain they have finally achieved that resolution?
Do you have any measurements to back that up? That is something that I've tried to get real, hard data on, but I cannot. The data is available for resolution & contrast ratios, but not color gamut.
By the way, you talk about jitter and weave, but have you ever seen a Kinoton "E" series projector? Instead of a mechanical intermittent, they use electronic stepper motors to do the film pulldown. They produce very steady images with jitter & weave well below that of standard projectors. Many people have only seen cheap, belt-driven projectors from Cristie, and incorrectly assume that that is the best film has to offer. It can actually be much better.
Modern print stocks have a contrast ratio of 10,000:1. Most video projectors are only able to get a ratio of 1,000:1, sometimes 1,100:1. That's not as big a difference as it sounds, since we perceive light logarithmically, but it is still a visible difference.
Film wins when it comes to contrast. It won't always, but it does for the moment.
1280x1024 on a movie sized screen! I've seen it on several occasions, and let me tell you, the pixels are huge. It's ugly. If you've seen digital projection, you're not "lucky".
The resolution of digital projection must be improved if it is to be taken seriously. 1280x1024 doesn't cut it. There are rumors that TI might finally inch their way up to full HDTV resolution (1920x1080) by the time Ep. II premeires, but even that isn't good enough.
Movie special effects are always done at at least 2K resolution, sometimes 4K. There was an article linked from here not too long ago which quoted some people who were working on FX for LOTR that some of their shots were done at 4K resolution. What good is that if movies are going to be projected at a puny resolution of 1280x1024?
Fortunately, there are more technologies on the horizon besides DLP. JVC's D-ILA is already at 2K resolution, and may reach 4K before too long. Sony is working on laser projection which should also be able to reach 4K.
The bottom line is -- don't be in a hurry! They should take the time to do things right, and make sure theaters are not stuck with equipment that is obsolete overnight. It's a good thing that there are currently only 20 digital screens in the US. The technology needs more time to mature first!
Yes, I've found that Kodak film pretty much sucks for color. Many years ago, I started using Fuji exclusively, and have never looked back. I'm very happy with the color I get from it.
Damn, I could have sworn there were some instructions in there that operated on two 64-bit doubles, but I see now that the best AltiVec can do is to operate on four 32-bit floats. Bummer.
Actually, the only thing they've mandated is that broadcasts become digital. They've said nothing about resolution. And since there are modes in the ATV standard for "Standard Definition" (NTSC-like 480 pixel tall images), that's what broadcasters are going to use.
History has shown that given a choice between transmitting the same number of channels at higher quality and transmitting a larger number of channels at the same quality, broadcasters will choose the latter every time, because they make more money that way.
We will never, ever see widespread HDTV in the US. We'll be stuck at NTSC resolution for the rest of our lives. Heck, if they could convince people to live with 100x100 digital video streams, they would, just so they could squeeze even more channels out of the same bandwidth. They drool at the idea of 50 million channels of shopping and other crap. Picture quality? What the heck is that?
If it can be optimized for AltiVec, almost nothing will be faster than a G4.
Just take a look at these RC5 stats (mid-way down the page). G4s smoke everything, because the RC5 client is optimized for AltiVec, thus it can compute four keys in a single clock cycle. By comparison, Athlons do one key per clock cycle, and Pentium 4s do one key every four clock cycles.
So if you've got an operation that can benefit from the G4's SIMD capabilities, Macs are your best bet.
How about one of those iRock FM modulators? Small, light, and battery powered. You can get 'em for $30 at CompUSA. They sound pretty good, better than the old cassette adapters. One guy I know swears it is actually better than the expensive modulator that hooks from his CD changer directly into his stereo's FM antenna lead.
So does this mean they are finally going to take down the Quake 2 stuff from the front page of their website and replace it with something a little more recent?
That means making things so that you no longer have to "guess" what percentage of moderation is being done by the editors, nor having to "think" how many mod points you've spent in a given day.
Make that information publicly available. Have moderation history done by the editors listed in their user info. And fix the notification so that it makes a distinction between moderation done by editors and non-editors.
The more open things are, the better.
Also, the Carbon APIs are actually a lot more robust when it comes to error handling. Nearly all of the Carbon APIs return meaningful error codes, so that your app can do the right thing in case of failure. Cocoa makes poor use of such mechanisms, opting for a simple boolean pass/fail in the best cases, and nothing at all in the worst. It also fails to take advantage of language-level exceptions for error handling. Basically, if something goes wrong in your Cocoa app, you're screwed when it comes to getting enough information to actually do something meaningful about it. And some of the newly developed software that is inherited from NeXT is even worse. CoreFoundation will simply call abort() on your app in many cases. How rude!
I know Apple's marketing material hypes Cocoa and downplays Carbon, but in the end, apps are better off with Carbon. Even major developers who are new to the Mac are choosing to use it (i.e., Maya).
Besides, a lot of the most interesting stuff in Mac OS X is down below the Carbon/Cocoa level. Quicktime, ColorSync, ATS, etc. are all in the Application Services level, and are thus accessible to both Carbon and Cocoa apps.
Cocoa requires an Objective-C runtime to work, but Carbon does not, as it is just a library. A little known fact is that both APIs allow the use of InterfaceBuilder to layout the GUI elements. Carbon has added support for .nib files in addition to the old-style resource manager stuff. And the new Carbon event structure is actually very nice.
Canada is getting very close to having its own version of the DMCA. It may not "seem much better" for long.
<RANT>
Why the hell don't people use progressive mode for their JPEGs? Just look at this list of browsers that support progressive JPEG. Just look at it! Do you know anyone who is still using a pre-Netscape 2.0 browser?! Me neither. Must we stay in the stone age forever? Sheesh. People who insist on non-progressive JPEGs are almost as bad as those people who go into conniptions when you send them non-ASCII e-mail.
</RANT>
(Er, I meant to link to this bug, actually. But that other one was pretty bad, too.)
I'm pretty sure the previous poster was not talking about the e-mail exchange with apex, but instead was talking about this, which really does look like a bad attitude. Flaming someone in all caps because they forgot to include the version number in the bug report is just plain out of line.
The author of that article doesn't seem to understnad the fact that Unicode is a character set, not a font. He also doesn't seem to understand how Unicode's surrogate pairs work (which allow for encoding of more than 1 million characters). He doesn't seem to understand that Unicode is an evolving standard (i.e., 3.1 is hardly the final version). And he doesn't seem to understand that UTF-8, UTF-16, UTF-32, etc. are all just different formats, and they actually represent the exact same character set.
But most importantly, he is flat-out wrong about how and why the decisions were made regarding encoding of East Asian languages. He needs to learn about the history of Han unification for CJK characters. If he did, he would know that linguists and computer scientists from East Asian countries have been involved in Unicode since the beginning. The unification of East Asian characters was done on purpose, and has the full support of linguists, scholars, and computer scientists from those countries.
If the author of that article had just spent a few minutes reading the a copy of The Unicode Standard, he would not have made those mistakes. He didn't even have to read the whole thing! Just the Introduction and Appendix A would have set him straight on the issues I just mentioned. The fact that he didn't means this guy really shouldn't be doing work for a company with the word "Research" in the title.
Oh, and even though that page says the article has not been modified since June 4, you can see from the google cache that they have since removed their promise of responding to criticism.
And one more thing: Since he derides those mean old Westerners on the Unicode committee for being insensitive towards the peoples of East Asian countries, perhaps he should ask himself if it is considered impolite or insensitive to sweepingly refer to such peoples as "Oriental", which he does in the first few paragraphs.
Even if he's not implementing paid subscriptions, he should still apply professional standards, considering that he is, by definition, a professional (i.e., he's getting paid to do this).
I also see that the story has now been changed to remove the inflammatory statement, yet there is no "UPDATE" tag or other indication that it has been edited. That is even less professional. Ugh.
Yeah, yeah, it's a free site, blah blah. Nearly all corporate news sites are free too, but if any of them did something like this, people would be yelling, no doubt about it.
Actually the article does not mention VPN even once.
It mostly talks about encryption, which is pretty useless on wireless networks. It stops only the most casual snoopers (like those "hackers" described in the article). Using WEP doesn't mean your network is secure.
It's hardly a secret that your laptop will see something when you're standing out in the parking lot near any company with an 802.11 network. That doesn't mean it's insecure. A company with even a smidgen of security sense will put the wireless network outside their firewall, and require employees to use VPN to access internal stuff. People on the outside may be able to get a little free internet access, but that's it.
The article is very light on details, gives no information as to what "wide open" means (just because you can see the network, that does not mean it is insecure). There is only one mention of the word "firewall" in the whole thing, and even then it's very vague.
I think this reporter has been duped by a couple of script kiddies. The supposed terms "war driving", "war pedalling", and "war walking" sound like something the kiddies made up on the spot, and later snickered at the reporter for believing.
You started this thread by saying iTunes shouldn't need an installer, but several people rightly pointed out why one is necessary. You then implied that putting the system-level frameworks iTunes depends on into a separate installer would somehow have prevented the problem, which it wouldn't.
You finally got something right by saying it was a bug in the install script, but you still don't seem to grasp the facts that 1) The install script is necessary and 2) Making it a separate install from iTunes would not prevent bugs in said install script.
It seems you're the one who wasn't paying attention.
Almost every movie comes with at least one trailer that is physically attached. The difference in this case is that the theaters were forbidden from intercutting their normal logo trailer between it and the movie, which they are normally allow to do.