An rmra atom holds pointers ( stored in rmda atoms) to other movies located somewhere else. Typically they are only found in Reference Movies.
Is the file with the rmra atom only a couple of
k in size? If so, you probably ended up with a reference movie. Try doing what I described here (or use this handy script) to track down the real movie file.
Ok - when you look at the page, you want to find the EMBED tag that holds the movie, so try the following command: curl http://www.apple.com/trailers/columbia/charliesang elsfullthrottle/large_trailer.html | grep EMBED
Looking at the embed tag, you will see that it references two quick time movies, one as a SRC argument, and one as a HREF argument. If both are given, you want the HREF, otherwise, you want the SRC. here is the link from the HREF option. It is the reference file for the movie.
Now, download it and run strings on it. You will see several instances of "url" on one line, and a url on the next line. These are the movies that the reference movie points to. Almost always, the first several will be to inform users that they have too old of a version of quick time (they usually have a quicktime version embedded in the name. ) skip over them.
at the end of the file, you will see this: ca2_fullthrottle-tlr_m480.mov - that is the relative path to the real movie file.
Take the URL for the reference file, and replace ca2_fullthrottle-tlr_481.mov with ca2_fullthrottle-tlr_m480.mov, and you will have this - the url for the actual movie data file.
This technique should work, even if Apple changes their convention of just adding a m before the size of the movie, to indicate the real file.
This also only works if the movie is being served over http. If it is over rtsp, you will need some extra tools.
I hope this helps. If you get confused, here is a decompiled version of the reference file, so that you can see in plain text what all of the embedded urls are for.
As an afterthought, if anyone wants to see the decompiled version of the reference file (to see what it checks for and what it does) I've put a decompiled version here.
Basically, it checks for old versions of QuickTime (anything before 5.0.2) and redirects them to a page telling them that they need to upgrade. Versions 5.02 and later are allowed through to the movie.
At 20K, almost everything is going to sound like ass. MP3, Vorbis, and AAC all sound pretty bad (with AAC sounding best of a bad set)
At that rate, for a mix of music and voice I feel that it's a close race between WMA, and Qdesign Music 2 Pro; with Qdesign edging out WMA for stereo audio at that rate (WMA has less high range - it sounds like the encoder filters out higher pitches to reduce noise before it encodes).
For a mix of music and voice, both Speex and Qualcomm Pure Voice are out, as while they would do a good job on voices they would not do well with the non-voice audio.
Didn't I hear of some Canadian company making a caffinated beer years ago? ( I think I found it - Rethink Beer (warning - annoying flash site)).
When I was in grad school, and I needed to concentrate for an all nighter to whip the last bits of a project into shape for a demo in the morning, I would usually have a big mug full of of espresso on one side of my desk, and a bottle of scotch on the other. If I was feeling tired, I would sip from one, if I was too stressed out and twitching, I would sip from the other. It worked quite well in the short term, but really burned me out over the long term...
It's good to know that in the second case, had the stress caused a stroke, I might have had a little protection.
As a test, I put the PDF file of the review of the hardware RNG up here (The summary is here).
If you have bittorrent installed, feel free to try to download from me.
Re:Anyone interested in extending this concept?
on
Geocoding All Content
·
· Score: 1
Or, you could use a database that supports positions. In Postgres, I use the point data type for holding locations.
That lets me do things like
select blurb,file, date, id, (geodist_mi(location,(select location from pictures where id=27 ) ))::numeric(7,3) as distance from pictures where location @ (select circle(location,5) from pictures where id=27) order by distance asc, date desc limit 5;
to obtain a list of up to 5 of the closest pictures within 5 degrees of picture #27 sorted with the nearest first, and breaking ties with the date the picture was taken.
additionaly, geodist_mi computes the great circle distance in miles between two points, not just the planar distance as your system uses.
Those guys have to run the most annoying relay tester I've seen. Every time it tests you, it sends a burst of 30 messages or so, all with return addresses on the box they are testing so they don't have to deal with bounces.
Now, some people may feel it's my own fault for taking advantage of the part of RFC 2821 which states that if a mailserver defers checking to see if it can relay or deliver the mail then "These servers SHOULD treat a failure for one or more recipients as a "subsequent failure" and return a mail message as discussed in section 6.".
But, I guess they feel that everyone runs sendmail, so every time they test my mailserver, I end up with another batch of relay rejected messages intended for them sitting in my postmaster mailbox.
There are two parts of this that bug me:
If a mail server does not relay mail, it is rude for a test to result in mail to the administrators of that server
It is possible for the username they use in their test to actually deliver mail to a real user. I consider it as bad as spamming if their test drops dozens of messages in the account of an innocent user with no idea of what is happening, or control over the mail server.
iPix claims to own the concept of flattening out an image taken through a fisheye lens into a panorama. They are rather aggressive in going after anyone they thing might infringe.
This is part of the reason that iPix has a rather bad reputation among many people (the other reasons are the low quality of the images their software generates, and the per-image-generated fees they charge for the use of their software).
Check the archives of Apple's QTVR Mailing List - this question comes up all the time, and the pros and cons of every existing image stitching program for every operating system have been discussed at great depth many times.
You might want to try Pano Tools if you can track down a copy, I've heard good things about them, but I think there was a problem with iPix claiming they infringed on a patent related to unwrapping spherical images.
Personally, I use Vr Worx 2.1 for stitching together panoramas, and making simple QTVR scenes. I've been very happy with it. It does only run under OSX, so it may not be suitable for your needs.
Microsoft file sharing is the most secure in the world. In fact, you don't even need to use a firewall with Windows.
That's right - all you need to do is leave your box hooked up to the network with no firewall, and in less than 5 minutes, one of a large number of dedicated volunteers will scan your system for any security flaws. If any are found, this tireless worker will log into your box, and install any necessary patches for you.
Don't worry if the disk thrashes from time to time, or if there is a lot of network activity, these are just symptoms of the high level of careful service you are receiving from your unknown friend.
To ensure the best service, be sure to tip him, by putting your credit card number, zip code, expiration date, SSN, and a suggested tip amount in a file called c:\tip.txt. A little gesture like this can go a long way!
caching associativity and memory interleaving anyone?
I think the main reason is probably that vDSP only uses vectorized functions on split data, if it has to stride over the data set, it uses a scalar version of the function.
Take a look at the docs for vDSP - for each function it lists the critera used to select the vector/scaler version of each function. You'll see the requirement that "Stride is equal to 1" pop up all the time - those functions only use altivec if the data has been split.
I've updated it to say: as much as twice as fast as vBigDSP on split data (but about the same on interleaved data).
There's a rather obvious reason for this, if you read the docs on vDSP.
All of the functions are provided in vector and scalar forms, with a set of criteria listed that must be met for the vector form of the function to be used. You'll see the requirement "Stride must be equal to 1" quite often. In other words, the function is only vectorized for split data, and runs a slower scalar version for interleaved data.
So, by using interleaved data, you are benchmarking a scalar vDSP against your other scalar algorithms.
Could someone please clarify why 2 people on one bastard PC would be better than 2 people on two normal PC's with some Cat-5 between them
Software Licensing - or at least that's why I was asked to look into some bizarre korean multihead cards back in 1999.
At the place I was working at that time, they used some expensive software that was licensed per processor. If one user used the software on a dual cpu machine, 2 licenses would be checked out. If he was on a single CPU machine one license would be checked out. Multiple instances of the software running on the same machine did not check out any additional licenses.
The idea was, to stick some of these strange cards into a bunch of machines, and put the users who were in training on these boxes. The resule: 2 users per license, saving up some licenses for the people who needed to run the program on the multicpu boxes.
The cards were completey unsuitable, and burned out quickly, so the project never went anywhere.
Back in early 1999, a couple of jobs ago, I saw a strange device to do this under windows.
A manager had been somewhere in southeast asia, and had picked up a few sets of a very strange device. They were pci cards that packed a low end video card, and some logic into a single package. Each one had a RJ-45 jack on the back, that connected to a little pod which hooked to a keyboard, monitor, etc to allow a second (third, etc...) user to have their own workspace and desktop on the host machine.
They only worked under Windows 98 (not NT) and were rather unstable - machines would crash often, the picture quality on the pods was rather bad, and the little suckers kept burning out. I think all of the samples died on me over a couple of days when I was testing them out.
I can't remember the name - I don't remember there being much english on the packaging.
The ones I saw didn't have a valid prefix. they both had the same hex digit for all values in the mac address - I think it was EF:EF:EF:EF:EF:EF or some other odd value like that.
I think it may have been a marker, to indicate where the MAC should go in the actual production run, but whoever made the cards skipped that step.
Now that I think about it, there is another part to the story.
There was a dos utility that came with them that claimed to do diagnostics, and let you change the MAC address. When you ran it (even the diagnostics), it would report "no card detected" and that card would be wiped - totally dead, not detectable by any OS in any way whatsoever.
Slashdot Mirror
An rmra atom holds pointers ( stored in rmda atoms) to other movies located somewhere else. Typically they are only found in Reference Movies.
Is the file with the rmra atom only a couple of k in size? If so, you probably ended up with a reference movie. Try doing what I described here (or use this handy script) to track down the real movie file.
scripting.slashdot.org can be broken down into:
and many others!
Well, if you want a simpler interface, you can download one.
It even comes with source code...
If someone tried to protect the file against conversion to another format, Apple provides a handy tool for unprotecting it.
It's called dumpster. It breaks down the atom structure of a quick time file, and lets you edit any aspect of it.
Ok - when you look at the page, you want to find the EMBED tag that holds the movie, so try the following command:g elsfullthrottle/large_trailer.html | grep EMBED
curl http://www.apple.com/trailers/columbia/charliesan
Looking at the embed tag, you will see that it references two quick time movies, one as a SRC argument, and one as a HREF argument. If both are given, you want the HREF, otherwise, you want the SRC.
here is the link from the HREF option. It is the reference file for the movie.
Now, download it and run strings on it. You will see several instances of "url" on one line, and a url on the next line. These are the movies that the reference movie points to. Almost always, the first several will be to inform users that they have too old of a version of quick time (they usually have a quicktime version embedded in the name. ) skip over them.
at the end of the file, you will see this: ca2_fullthrottle-tlr_m480.mov - that is the relative path to the real movie file.
Take the URL for the reference file, and replace ca2_fullthrottle-tlr_481.mov with ca2_fullthrottle-tlr_m480.mov, and you will have this - the url for the actual movie data file.
This technique should work, even if Apple changes their convention of just adding a m before the size of the movie, to indicate the real file.
This also only works if the movie is being served over http. If it is over rtsp, you will need some extra tools.
I hope this helps. If you get confused, here is a decompiled version of the reference file, so that you can see in plain text what all of the embedded urls are for.
As an afterthought, if anyone wants to see the decompiled version of the reference file (to see what it checks for and what it does) I've put a decompiled version here.
Basically, it checks for old versions of QuickTime (anything before 5.0.2) and redirects them to a page telling them that they need to upgrade. Versions 5.02 and later are allowed through to the movie.
You can just run strings on the reference file to see all of the Urls in it.
.qtrm source file.
Or, if you are feeling brave (and have a mac), you could use movie to xml to decompile it to a
At 20K, almost everything is going to sound like ass. MP3, Vorbis, and AAC all sound pretty bad (with AAC sounding best of a bad set)
At that rate, for a mix of music and voice I feel that it's a close race between WMA, and Qdesign Music 2 Pro; with Qdesign edging out WMA for stereo audio at that rate (WMA has less high range - it sounds like the encoder filters out higher pitches to reduce noise before it encodes).
For a mix of music and voice, both Speex and Qualcomm Pure Voice are out, as while they would do a good job on voices they would not do well with the non-voice audio.
I was home all weekend, and no one called me. Not even a salesman.
I know how you can get lots of phone calls.
Just put your phone number in every message you post.
- Didn't I hear of some Canadian company making a caffinated beer years ago? ( I think I found it - Rethink Beer (warning - annoying flash site)).
- When I was in grad school, and I needed to concentrate for an all nighter to whip the last bits of a project into shape for a demo in the morning, I would usually have a big mug full of of espresso on one side of my desk, and a bottle of scotch on the other. If I was feeling tired, I would sip from one, if I was too stressed out and twitching, I would sip from the other. It worked quite well in the short term, but really burned me out over the long term...
It's good to know that in the second case, had the stress caused a stroke, I might have had a little protection.I'm playing around with bittorrent.
As a test, I put the PDF file of the review of the hardware RNG up here (The summary is here).
If you have bittorrent installed, feel free to try to download from me.
That lets me do things like
to obtain a list of up to 5 of the closest pictures within 5 degrees of picture #27 sorted with the nearest first, and breaking ties with the date the picture was taken.
additionaly, geodist_mi computes the great circle distance in miles between two points, not just the planar distance as your system uses.
Now, some people may feel it's my own fault for taking advantage of the part of RFC 2821 which states that if a mailserver defers checking to see if it can relay or deliver the mail then "These servers SHOULD treat a failure for one or more recipients as a "subsequent failure" and return a mail message as discussed in section 6.".
But, I guess they feel that everyone runs sendmail, so every time they test my mailserver, I end up with another batch of relay rejected messages intended for them sitting in my postmaster mailbox.
There are two parts of this that bug me:
That was Panotools - you can read the story here.
iPix claims to own the concept of flattening out an image taken through a fisheye lens into a panorama. They are rather aggressive in going after anyone they thing might infringe.
This is part of the reason that iPix has a rather bad reputation among many people (the other reasons are the low quality of the images their software generates, and the per-image-generated fees they charge for the use of their software).
Check the archives of Apple's QTVR Mailing List - this question comes up all the time, and the pros and cons of every existing image stitching program for every operating system have been discussed at great depth many times.
You might want to try Pano Tools if you can track down a copy, I've heard good things about them, but I think there was a problem with iPix claiming they infringed on a patent related to unwrapping spherical images.
Personally, I use Vr Worx 2.1 for stitching together panoramas, and making simple QTVR scenes. I've been very happy with it. It does only run under OSX, so it may not be suitable for your needs.
...or reverse-engineer it from the MS-DOS binaries...
It's been done - the book Dissecting Dos came out in 1994.
So if you wait a year, and send the request again, do you get a copy of your first request back?
here's a mention from reuters.
He was buried last tuesday in Kodiakanal India.
Microsoft file sharing is the most secure in the world. In fact, you don't even need to use a firewall with Windows.
That's right - all you need to do is leave your box hooked up to the network with no firewall, and in less than 5 minutes, one of a large number of dedicated volunteers will scan your system for any security flaws. If any are found, this tireless worker will log into your box, and install any necessary patches for you.
Don't worry if the disk thrashes from time to time, or if there is a lot of network activity, these are just symptoms of the high level of careful service you are receiving from your unknown friend.
To ensure the best service, be sure to tip him, by putting your credit card number, zip code, expiration date, SSN, and a suggested tip amount in a file called c:\tip.txt. A little gesture like this can go a long way!
caching associativity and memory interleaving anyone?
I think the main reason is probably that vDSP only uses vectorized functions on split data, if it has to stride over the data set, it uses a scalar version of the function.
Take a look at the docs for vDSP - for each function it lists the critera used to select the vector/scaler version of each function. You'll see the requirement that "Stride is equal to 1" pop up all the time - those functions only use altivec if the data has been split.
I've updated it to say: as much as twice as fast as vBigDSP on split data (but about the same on interleaved data).
There's a rather obvious reason for this, if you read the docs on vDSP.
All of the functions are provided in vector and scalar forms, with a set of criteria listed that must be met for the vector form of the function to be used. You'll see the requirement "Stride must be equal to 1" quite often.
In other words, the function is only vectorized for split data, and runs a slower scalar version for interleaved data.
So, by using interleaved data, you are benchmarking a scalar vDSP against your other scalar algorithms.
Could someone please clarify why 2 people on one bastard PC would be better than 2 people on two normal PC's with some Cat-5 between them
Software Licensing - or at least that's why I was asked to look into some bizarre korean multihead cards back in 1999.
At the place I was working at that time, they used some expensive software that was licensed per processor. If one user used the software on a dual cpu machine, 2 licenses would be checked out. If he was on a single CPU machine one license would be checked out. Multiple instances of the software running on the same machine did not check out any additional licenses.
The idea was, to stick some of these strange cards into a bunch of machines, and put the users who were in training on these boxes. The resule: 2 users per license, saving up some licenses for the people who needed to run the program on the multicpu boxes.
The cards were completey unsuitable, and burned out quickly, so the project never went anywhere.
Back in early 1999, a couple of jobs ago, I saw a strange device to do this under windows.
A manager had been somewhere in southeast asia, and had picked up a few sets of a very strange device. They were pci cards that packed a low end video card, and some logic into a single package. Each one had a RJ-45 jack on the back, that connected to a little pod which hooked to a keyboard, monitor, etc to allow a second (third, etc...) user to have their own workspace and desktop on the host machine.
They only worked under Windows 98 (not NT) and were rather unstable - machines would crash often, the picture quality on the pods was rather bad, and the little suckers kept burning out. I think all of the samples died on me over a couple of days when I was testing them out.
I can't remember the name - I don't remember there being much english on the packaging.
The ones I saw didn't have a valid prefix. they both had the same hex digit for all values in the mac address - I think it was EF:EF:EF:EF:EF:EF or some other odd value like that.
I think it may have been a marker, to indicate where the MAC should go in the actual production run, but whoever made the cards skipped that step.
Now that I think about it, there is another part to the story.
There was a dos utility that came with them that claimed to do diagnostics, and let you change the MAC address. When you ran it (even the diagnostics), it would report "no card detected" and that card would be wiped - totally dead, not detectable by any OS in any way whatsoever.
Any cell phone, even one without service, can call 911.
The old beat up cell phone (with an auto adaptor, or a single use emergency battery) in your glove box may save your life someday.