The question that will determine if this is good or not is how much Google ends up charging the retailer. If it's under a buck a listing I can't see this being that bad, if it's around $10 a listing this is going to be bad.
If by "listing" you mean "unique product identifier", then $1 is way, way, way too much. I doubt Amazon would pay several hundred million dollars per (week/month/whatever) to be listed in Google product search.
I suppose it would have to be some sort of volume-based scale, where a few hundred item store might pay $0.05/item to list, and Amazon would pay $0.0001/item.
In this case, it kind of is a shopping mall... I would think that as long as Amazon, and other major online e-tailers remain on the listings, I won't complain at all.
If Amazon and other big sites (eBay, Newegg, etc.) fell off the list, I wouldn't care, as I know those sites, and they have good search engines. Google's advantage was in finding smaller legitimate sites with good prices. I think that is who will suffer most from this change.
Are you saying that as a retailer you will pay to have products listed which you no longer sell?
Scammer's might.
Right now, Google Shopping results hit a lot of places that either have much higher prices once you get to the site, don't have the item in stock (but will "order it" if you pay them first), or have other shady practices (not giving Google the correct shipping costs, price is only good for wholesale volume, etc.).
Depending on the cost, some of those places might pay Google to keep them in the listing. Other, more reputable stores might have to either increase prices or drop off the Google list. So, it would appear that this change might lead to worse overall results.
The solution to this is to allow users to mark results as good or bad, and use that as part of the sorting algorithm, but Google hasn't ever had anything like that for shopping results.
I find it highly suspicious they would consistently do that with the most expensive part and would instead look to reduce costs, esp at a firm that has happy hours every day with half price drinks.
I've never understood why employees aren't taught that ice is expensive, either.
Add an ice machine which is basically a sunk cost and then periodically have filters changed, negligible, and electricity is only around 10 cents per kWh so I'd be curious to see those calculations.
The moderately large ice machines are 10-15 amps, giving you nearly 2kW usage. That's $0.20/hour to run at your cheap residential 10 cents number (even Texas averages a bit above that), and you only get around 500 cubes per hour (until you really hit the large machines...that's why empty hotel ice machines are so common). Then, you have a lot of waste. Ice machines use electricity only to make the ice...it is kept cool by insulation only, so a lot melts. You also get a lot of water waste because of the way the machines re-circulate water over the chiller plates...you get evaporation and spillage into the holding bin. You also have waste because you have to run these machines for quite a while before customers show up so that you have enough to make up for the overnight melt. Add in cleaning waste (every cleaning means the machine is completely emptied and the holding bin is no longer chilled by ice), and it really adds up.
Add in maintenance costs (those filters can be hundreds of dollars, and get changed out quite often), and it really does cost that much. That said, I have a small commercial ice machine at home (which is part of the other reason I know about this stuff), and it's worth every penny for the pure, clear ice you get.
The situation is a lot better than it used to be, but VM software tends to have a problem with keeping an accurate clock and that can bite you in some interesting ways, such as:
Pretty much anything to do with authentication; that means Active Directory, Kerberos and LDAP for sure
NTP servers. Just don't go there. It should be obvious why, given the statement above!
I run a trio of physical hosts that run ntpd (which is the minimum for NTP best practices, 5 are recommended) getting authoritative time from [currently 17] different external stratum 1-3 ntpd servers. My Linux machines run ntpd against my local physical hosts and are never more than 20ms off time.
Windows DCs get time from the local physical NTP hosts and pass time down to Windows machines (both physical and VM) using the standard Windows Time services (VMs have "sync to host" in VMware Tools unchecked), and accuracy is better than 100ms at all times. Since Windows allows Kerberos to be off by as much as 5 minutes (that's 300,000ms), even a few thousand milliseconds isn't a big deal.
System logging - you can forget about having accurately time synchronised system logs on VMs.
There probably isn't much that happens on a large computer network where you get better than 1 second accuracy in the logging, either, so even if you need more precision than that, you couldn't get it even with perfect time sync.
Granted if you fill it yourself you can add little or no ice, but for any to go cups where the store fills it, I think you'd be surprised at just how little drink is in the rather large cup when you take away all the (cheap filler) ice.
Having seen the accounting for several places I worked that sold carbonated soft drinks, the ice was actually the most expensive part.
For each gallon of soda at the fountain, it costs about $1 for the syrup, $0.40 or so for the carbonated water (cost of CO2 plus filtration), $0.10-$0.50 for the cups (depends on number used to serve the gallon and type of cup), and about $1.25 for the ice (cost of filtration plus electricity for freeziing the water).
That said, you can see that at $2 for a drink, you only need to keep the amount consumed to around 64oz per drink to make money.
Still a hell of a lot of cash, but probably not unaffordable to Google.
It would be overall better for Google's bottom line to just buy one of each of the movie and recording companies and break them off from the MPAA/RIAA.
They might sort the videos into two piles: those with no third-party material, and those with some
This won't work, as essentially every video made is copyrighted somewhere, even if there isn't any "third-party material".
The question is whether the uploader has the right to distribute what they uploaded, and it is completely impossible to determine based on content alone. As an example, behind-the-scenes footage of a highly anticipated movie might or might not be infringing, yet we know the MPAA would request a takedown.
The workers on the next level might sort those with some third party material into those which probably constitute fair use
This also will not work, as fair use is merely a defense. If the uploader is not authorized to distribute the material in the video, then it is infringing, and would have to be taken down. At that point, it is technically a matter for a person fully-trained in copyright law.
The whole point of pre-screening would be to avoid having to deal with the huge number of DMCA takedown notices and move the staff from that group to the pre-screening group. If videos that have been historically DMCA'd get through (i.e., fairly obvious fair use like dancing toddlers), then Google gains nothing by pre-screening. Seriously, even if Google got the big media companies to pay for this pre-screening, they wouldn't do it unless it overall helped their bottom line, and I can't see any way that it would, especially if you take into account goodwill.
Even my stupid cellphone has an audio detection feature. You let it record a few seconds of music and it sends the data to a server that replies with the name of the song, the album and the band.
Just because there are a few seconds of a recognizable song in the audio of an uploaded video does not mean that it is infringing.
There will always be false positives
With automated scanning where a few seconds is enough to trigger, there will be nothing but false positives.
Thank you for making the same point that the article makes...only a human can actually determine infringement.
What's a reasonable cost? Netflix, Lovefilm and Amazon Video are all pretty cheap.
If you want to maybe be able to see a movie, then all these work fine. But, if you absolutely want to watch a particular movie, downloading a torrent is better.
With an already downloaded torrent, you don't have to worry about an internet connection right now. You also don't have to worry about losing access to the movie at some future date. And, although none of the services you list have commercials (that I know of) before/during your viewing, that could change, too.
Also, although Netflix and Amazon have reasonable support for a wide variety of devices (don't know about Lovefilm), there are still a lot of devices they don't cover. This is not a problem with a torrent download.
Self-promotion isn't considered advertising though, which is part of the problem, and I'm certain cuts into your 21-22 minutes number.
If by "self-promotion", you mean pop ups that cover part of the show, then I suppose you could count that as "not show", but where do you draw the line? Is the station logo "self-promotion"? If so, then 100% of TV is "non-show".
Otherwise, my numbers are correct, as everything that is created by the show producers ends up at about 21.5 minutes per half hour almost universally. And, this hasn't changed a lot in the past five years (maybe down 45 seconds or so). You can easily check this by downloading a torrent of a show, and add between 20 and 40 seconds for the end credits that the uploaders cut.
The average half hour show is only about 17-20 minutes long anymore. The rest is ads and self promotion.
I hate commercials as much as anybody (probably more, since every bit of television I have watched in nearly 20 years has been recorded and commercials skipped or edited out), but a scripted half hour show on a broadcast network is between 21 and 22 minutes after the non-show parts are removed. This does include the credits, though. Cable networks are about the same.
Some shows (like Mythbusters) have a lot of in-show recaps that bring down the amount of real content by a lot, but it's still "show" and not any kind of ad. Part of the reason they do this is that some people might actually forget something after a 5-6 minute break.
One thing that really makes me glad I edit out commercials is the front loading that some shows (like Supernatural) do. In the first half hour of show, there are only about 5 minutes of commercials, which puts a whopping 15 minutes in the last half hour, with about 10 minutes of that in the last 15 minutes. This is done because they assume you are hooked at that point and will stay around until the end of the show regardless of the annoyance.
In this case, if the definition of 'infringement' is stretched far enough to save our poor, beleaguered, broadcasters it is stretched far enough to allow near-total control over any device that handles rendering of copyrighted material, which is virtually anything.
The next obvious step is that you are "infringing" if you turn off Crappy Show to watch something else. After that, the act of not turning on Crappy Show will be infringement.
Even having used them myself, I don't get the appeal of what basically amounts to a poor-man's alternative to having multiple monitors.
For me, the problem is that multiple monitors don't give me enough real estate compared to virtual desktops. I run the Microsoft-written "TopDesk" at 11x3, so that means my total desktop space is 21120x3600 pixels.
I can have a dozen programs maximized without having to hunt through them (one keystroke plus one mouse click gets me to any open window). In addition, it's easy to group sub-tasks together onto one desktop. So, I can have 3-4 terminals open to a Linux machines to configure nfs client and server, and video and audio editing software also open, yet neither group of windows interferes with the other. In addition, my e-mail client, web browser, and a spreadsheet are also open without getting in the way of any other tasks.
I can also easily configure windows to always open in the same location, which can be a problem with multiple monitors. Then, too, moving windows around from one desktop to another is much easier, as I have the overview of the whole workspace, and can move the window using that (and that has shortcuts that allow me to snap the window to special places).
There's nothing wrong with multiple monitors (although it can be an issue when you use a KVM as I do), but adding virtual desktops gives you another whole level of window management tricks to employ.
PeerBlock is probably great for all the Neanderthals who still use Windows. It isn't available to anybody else.
The lists are available for download by anybody, and are in a format that can easily be used as a source for whatever sort of filtering software you want to use (like iptables, or the system built in to your BitTorrent client).
But it might be other kind of fake information (maybe give a lot of fake sources through peer exchange?).
I think that might have an effect on torrents with few peers, but if you have a lot of real peers, it doesn't take long to connect to enough that your speeds for that torrent are maxed out (or at least "way fast enough").
Some of the description seems to indicate that they would do IP spoofing of legitimate peers to send invalid data and get that peer banned. I'm not sure how well that would work in reality, and even if it does work, it seems like a minor fix to the protocol would solve that issue.
I've been transcoding for at least 10 years (and everything but ffmpeg sucks).
For H.264 encoding, ffmpeg just spawns x264, but you don't have the ability to control all x264 options through ffmpeg, so using x264 directly is generally better.
For the life of me, I can't figure out how YIFY does it...
Lots of pre-filtering to remove hard-to-encode detail. Even a relatively gentle temporal smoother can cut bitrates by half on very grainy material, and not affect quality much unless you freeze-frame.
Single pass has always produced garbage results for me unless you knock the bitrate WAY up, but why even bother recoding, then?
Two-pass x264 with an average bitrate just has the first pass compute the CRF value that will give you the target bitrate. After just a few encodes of similar material, all the rest of your encodes can use one-pass with the CRF set to near the value that was computed for the two-pass encodes.
One of the things this does is that every movie gets exactly the bitrate it deserves, so easy to compress (less movement, less film grain, etc.) movies get a lot lower bitrate at the same CRF value. This is how I end up with some Blu-Ray rips at 2Mbps, despite using the same "high quality" settings on every encode. Some movies just don't need as much bitrate. The lowest bitates come from clean tradition animation (since you only have 12 pictures per second, and easy to determine movement), but the latest Harry Potter movies are just so freaking dark that they also end up at really low average bitates.
Now, if you need to fit to absolute file size because of limitations like file system (4GB file size limit for FAT32) or device (CD-R, etc.), then two-pass might be the right way to go, but for something like a phone with relatively large storage compared to the file size once you re-size the source, you're still better off with a single-pass CRF encode, where 800x480/24p will take up between 500MB and 1GB per hour of movie for very good quality. If that's too big for your taste, change the CRF value to lower the quality. Once you find a CRF value that works for you, stick with it for encodes for the same target.
if you want to go ahead and do work that, chances are, has already been done, and is constantly being done for you by others that create far better quality transcodes.
In general, there are two kinds of rips available on torrents: too large so that quality stays high, and really small to fit on phones or similar devices. It's very hard to find something in-between, where you the encode is close to transparent, but as small as possible. Most rips still use two-pass average bitrate mode, which is basically "find the CRF value that gives me this bitrate", which is the wrong way to maintain quality. The right way is to pick a CRF value and let x264 figure out the bitrate needed for that quality.
Although recent rips are getting the word on how to use x264 correctly, there's an awful lot out there that don't use the right "tune", don't set the correct colorspace, and generally tinker too much with settings that don't matter, while ignoring the ones that do. The forums at doom(9|10).org are a great wealth of information, much of it direct from the x264 developers.
In addition, things like the green tint on "Fellowship of the Ring" can be removed if you do the transcode yourself. And, let's not talk about missing audio or subtitle tracks, glitches in the movie, or the fact that I own at least 10 movies on DVD and Blu-Ray that just aren't available anywhere on the Internet.
That's why video professionals and tv stations rely on hardware based transcoding, and this solutions tend to be expensive.
x264 can encode 1080p in realtime on a modern Intel CPUs (Sandy Bridge, etc.) with pretty much as good a quality for the same bitrate as most hardware solutions. For non-HD, x264 just smokes hardware, as it can do better than realtime encodes at very high quality on those same CPUs.
With QS, I can rip an entire 50GB Blu-Ray in 12 minutes to a 1080p MKV @ 8000kbps. It takes about 16 hours doing the same task with a normal x264 encoder such as Handbrake even though the quality is a little bit better.
Even using the "slower" preset on x264, 1080p encodes take about 3 times as long as the movie, so no more than 8 hours. This is on a slower CPU (since you have QuickSync) than you use, and end up at about 4Mbps
If I used a less-intensive preset, I would get encodes at about the same bitrate as yours, but taking just a little more than the running time of the movie to do it. QuickSync may be even faster, but 3 hours to encode most movies is good enough.
With enough bitrate, anything looks good.
In general, this is true, but very poor encoders can still screw up a high bitrate encode. That's why I use x264...it's going to give me the absolute best quality and lowest bitrate for the amount of encoding time. Since I only encode my movies one time, taking 6-8 hours to do it and getting bitrates as low as 2Mbps for 1080p with high picture quality is worth the extra time. It's also nice to be able to use the full power of AVISynth during the encode. My Blu-Ray rip of "A New Hope" doesn't have the useless "Jabba in the hangar" scene, and I'm working on getting Han shooting first.
The problem with transcoding is that it exists at all. Strongarm the holdout encoders into using h264 or mp4v with mp4 wrappers, and transcoding will be like... well, like anything no one does anymore.
There will always be transcoding, since you can't fit the 20GB H.264 stream from a Blu-Ray on a phone. And, why would you want to? Resize the 1920x1080 to 800x480 or so, and it will look great on every phone.
For tablets or other devices with more resolution, you still don't need all the bits that most Blu-Ray encodes use. Most are essentially constant bit rate around 25-30Mbps. For movies that are essentially "talking heads" (courtroom dramas like A Few Good Men and Presumed Innocent are the best examples), most of those bits aren't actually adding anything to the picture quality. Even action movies can easily get by with 10Mbps average on a full-resolution transcode, as long as the action scenes get enough bits.
Slashdot Mirror
The question that will determine if this is good or not is how much Google ends up charging the retailer. If it's under a buck a listing I can't see this being that bad, if it's around $10 a listing this is going to be bad.
If by "listing" you mean "unique product identifier", then $1 is way, way, way too much. I doubt Amazon would pay several hundred million dollars per (week/month/whatever) to be listed in Google product search.
I suppose it would have to be some sort of volume-based scale, where a few hundred item store might pay $0.05/item to list, and Amazon would pay $0.0001/item.
In this case, it kind of is a shopping mall... I would think that as long as Amazon, and other major online e-tailers remain on the listings, I won't complain at all.
If Amazon and other big sites (eBay, Newegg, etc.) fell off the list, I wouldn't care, as I know those sites, and they have good search engines. Google's advantage was in finding smaller legitimate sites with good prices. I think that is who will suffer most from this change.
Are you saying that as a retailer you will pay to have products listed which you no longer sell?
Scammer's might.
Right now, Google Shopping results hit a lot of places that either have much higher prices once you get to the site, don't have the item in stock (but will "order it" if you pay them first), or have other shady practices (not giving Google the correct shipping costs, price is only good for wholesale volume, etc.).
Depending on the cost, some of those places might pay Google to keep them in the listing. Other, more reputable stores might have to either increase prices or drop off the Google list. So, it would appear that this change might lead to worse overall results.
The solution to this is to allow users to mark results as good or bad, and use that as part of the sorting algorithm, but Google hasn't ever had anything like that for shopping results.
I find it highly suspicious they would consistently do that with the most expensive part and would instead look to reduce costs, esp at a firm that has happy hours every day with half price drinks.
I've never understood why employees aren't taught that ice is expensive, either.
Add an ice machine which is basically a sunk cost and then periodically have filters changed, negligible, and electricity is only around 10 cents per kWh so I'd be curious to see those calculations.
The moderately large ice machines are 10-15 amps, giving you nearly 2kW usage. That's $0.20/hour to run at your cheap residential 10 cents number (even Texas averages a bit above that), and you only get around 500 cubes per hour (until you really hit the large machines...that's why empty hotel ice machines are so common). Then, you have a lot of waste. Ice machines use electricity only to make the ice...it is kept cool by insulation only, so a lot melts. You also get a lot of water waste because of the way the machines re-circulate water over the chiller plates...you get evaporation and spillage into the holding bin. You also have waste because you have to run these machines for quite a while before customers show up so that you have enough to make up for the overnight melt. Add in cleaning waste (every cleaning means the machine is completely emptied and the holding bin is no longer chilled by ice), and it really adds up.
Add in maintenance costs (those filters can be hundreds of dollars, and get changed out quite often), and it really does cost that much. That said, I have a small commercial ice machine at home (which is part of the other reason I know about this stuff), and it's worth every penny for the pure, clear ice you get.
The situation is a lot better than it used to be, but VM software tends to have a problem with keeping an accurate clock and that can bite you in some interesting ways, such as:
I run a trio of physical hosts that run ntpd (which is the minimum for NTP best practices, 5 are recommended) getting authoritative time from [currently 17] different external stratum 1-3 ntpd servers. My Linux machines run ntpd against my local physical hosts and are never more than 20ms off time.
Windows DCs get time from the local physical NTP hosts and pass time down to Windows machines (both physical and VM) using the standard Windows Time services (VMs have "sync to host" in VMware Tools unchecked), and accuracy is better than 100ms at all times. Since Windows allows Kerberos to be off by as much as 5 minutes (that's 300,000ms), even a few thousand milliseconds isn't a big deal.
System logging - you can forget about having accurately time synchronised system logs on VMs.
There probably isn't much that happens on a large computer network where you get better than 1 second accuracy in the logging, either, so even if you need more precision than that, you couldn't get it even with perfect time sync.
Granted if you fill it yourself you can add little or no ice, but for any to go cups where the store fills it, I think you'd be surprised at just how little drink is in the rather large cup when you take away all the (cheap filler) ice.
Having seen the accounting for several places I worked that sold carbonated soft drinks, the ice was actually the most expensive part.
For each gallon of soda at the fountain, it costs about $1 for the syrup, $0.40 or so for the carbonated water (cost of CO2 plus filtration), $0.10-$0.50 for the cups (depends on number used to serve the gallon and type of cup), and about $1.25 for the ice (cost of filtration plus electricity for freeziing the water).
That said, you can see that at $2 for a drink, you only need to keep the amount consumed to around 64oz per drink to make money.
Still a hell of a lot of cash, but probably not unaffordable to Google.
It would be overall better for Google's bottom line to just buy one of each of the movie and recording companies and break them off from the MPAA/RIAA.
They might sort the videos into two piles: those with no third-party material, and those with some
This won't work, as essentially every video made is copyrighted somewhere, even if there isn't any "third-party material".
The question is whether the uploader has the right to distribute what they uploaded, and it is completely impossible to determine based on content alone. As an example, behind-the-scenes footage of a highly anticipated movie might or might not be infringing, yet we know the MPAA would request a takedown.
The workers on the next level might sort those with some third party material into those which probably constitute fair use
This also will not work, as fair use is merely a defense. If the uploader is not authorized to distribute the material in the video, then it is infringing, and would have to be taken down. At that point, it is technically a matter for a person fully-trained in copyright law.
The whole point of pre-screening would be to avoid having to deal with the huge number of DMCA takedown notices and move the staff from that group to the pre-screening group. If videos that have been historically DMCA'd get through (i.e., fairly obvious fair use like dancing toddlers), then Google gains nothing by pre-screening. Seriously, even if Google got the big media companies to pay for this pre-screening, they wouldn't do it unless it overall helped their bottom line, and I can't see any way that it would, especially if you take into account goodwill.
Even my stupid cellphone has an audio detection feature. You let it record a few seconds of music and it sends the data to a server that replies with the name of the song, the album and the band.
Just because there are a few seconds of a recognizable song in the audio of an uploaded video does not mean that it is infringing.
There will always be false positives
With automated scanning where a few seconds is enough to trigger, there will be nothing but false positives.
Thank you for making the same point that the article makes...only a human can actually determine infringement.
What's a reasonable cost? Netflix, Lovefilm and Amazon Video are all pretty cheap.
If you want to maybe be able to see a movie, then all these work fine. But, if you absolutely want to watch a particular movie, downloading a torrent is better.
With an already downloaded torrent, you don't have to worry about an internet connection right now. You also don't have to worry about losing access to the movie at some future date. And, although none of the services you list have commercials (that I know of) before/during your viewing, that could change, too.
Also, although Netflix and Amazon have reasonable support for a wide variety of devices (don't know about Lovefilm), there are still a lot of devices they don't cover. This is not a problem with a torrent download.
Self-promotion isn't considered advertising though, which is part of the problem, and I'm certain cuts into your 21-22 minutes number.
If by "self-promotion", you mean pop ups that cover part of the show, then I suppose you could count that as "not show", but where do you draw the line? Is the station logo "self-promotion"? If so, then 100% of TV is "non-show".
Otherwise, my numbers are correct, as everything that is created by the show producers ends up at about 21.5 minutes per half hour almost universally. And, this hasn't changed a lot in the past five years (maybe down 45 seconds or so). You can easily check this by downloading a torrent of a show, and add between 20 and 40 seconds for the end credits that the uploaders cut.
I understand why Fox and Friends wouldn't like this kind of feature, but what kind of legal ground do they have here?
Fox has standing to sue if they are suing on behalf of one of their fully-owned broadcast TV stations (like in New York, DC, Los Angeles, etc.).
The stations that do the final broadcast hold copyright over the entire "stream", which includes commercials.
The average half hour show is only about 17-20 minutes long anymore. The rest is ads and self promotion.
I hate commercials as much as anybody (probably more, since every bit of television I have watched in nearly 20 years has been recorded and commercials skipped or edited out), but a scripted half hour show on a broadcast network is between 21 and 22 minutes after the non-show parts are removed. This does include the credits, though. Cable networks are about the same.
Some shows (like Mythbusters) have a lot of in-show recaps that bring down the amount of real content by a lot, but it's still "show" and not any kind of ad. Part of the reason they do this is that some people might actually forget something after a 5-6 minute break.
One thing that really makes me glad I edit out commercials is the front loading that some shows (like Supernatural) do. In the first half hour of show, there are only about 5 minutes of commercials, which puts a whopping 15 minutes in the last half hour, with about 10 minutes of that in the last 15 minutes. This is done because they assume you are hooked at that point and will stay around until the end of the show regardless of the annoyance.
In this case, if the definition of 'infringement' is stretched far enough to save our poor, beleaguered, broadcasters it is stretched far enough to allow near-total control over any device that handles rendering of copyrighted material, which is virtually anything.
The next obvious step is that you are "infringing" if you turn off Crappy Show to watch something else. After that, the act of not turning on Crappy Show will be infringement.
Even having used them myself, I don't get the appeal of what basically amounts to a poor-man's alternative to having multiple monitors.
For me, the problem is that multiple monitors don't give me enough real estate compared to virtual desktops. I run the Microsoft-written "TopDesk" at 11x3, so that means my total desktop space is 21120x3600 pixels.
I can have a dozen programs maximized without having to hunt through them (one keystroke plus one mouse click gets me to any open window). In addition, it's easy to group sub-tasks together onto one desktop. So, I can have 3-4 terminals open to a Linux machines to configure nfs client and server, and video and audio editing software also open, yet neither group of windows interferes with the other. In addition, my e-mail client, web browser, and a spreadsheet are also open without getting in the way of any other tasks.
I can also easily configure windows to always open in the same location, which can be a problem with multiple monitors. Then, too, moving windows around from one desktop to another is much easier, as I have the overview of the whole workspace, and can move the window using that (and that has shortcuts that allow me to snap the window to special places).
There's nothing wrong with multiple monitors (although it can be an issue when you use a KVM as I do), but adding virtual desktops gives you another whole level of window management tricks to employ.
PeerBlock is probably great for all the Neanderthals who still use Windows. It isn't available to anybody else.
The lists are available for download by anybody, and are in a format that can easily be used as a source for whatever sort of filtering software you want to use (like iptables, or the system built in to your BitTorrent client).
But it might be other kind of fake information (maybe give a lot of fake sources through peer exchange?).
I think that might have an effect on torrents with few peers, but if you have a lot of real peers, it doesn't take long to connect to enough that your speeds for that torrent are maxed out (or at least "way fast enough").
Some of the description seems to indicate that they would do IP spoofing of legitimate peers to send invalid data and get that peer banned. I'm not sure how well that would work in reality, and even if it does work, it seems like a minor fix to the protocol would solve that issue.
I've been transcoding for at least 10 years (and everything but ffmpeg sucks).
For H.264 encoding, ffmpeg just spawns x264, but you don't have the ability to control all x264 options through ffmpeg, so using x264 directly is generally better.
For the life of me, I can't figure out how YIFY does it...
Lots of pre-filtering to remove hard-to-encode detail. Even a relatively gentle temporal smoother can cut bitrates by half on very grainy material, and not affect quality much unless you freeze-frame.
What about slice-based parallel processing?
According to the wiki, you're still better off using normal multithreading even if you are using slices (as you must if you are encoding for Blu-Ray).
Single pass has always produced garbage results for me unless you knock the bitrate WAY up, but why even bother recoding, then?
Two-pass x264 with an average bitrate just has the first pass compute the CRF value that will give you the target bitrate. After just a few encodes of similar material, all the rest of your encodes can use one-pass with the CRF set to near the value that was computed for the two-pass encodes.
One of the things this does is that every movie gets exactly the bitrate it deserves, so easy to compress (less movement, less film grain, etc.) movies get a lot lower bitrate at the same CRF value. This is how I end up with some Blu-Ray rips at 2Mbps, despite using the same "high quality" settings on every encode. Some movies just don't need as much bitrate. The lowest bitates come from clean tradition animation (since you only have 12 pictures per second, and easy to determine movement), but the latest Harry Potter movies are just so freaking dark that they also end up at really low average bitates.
Now, if you need to fit to absolute file size because of limitations like file system (4GB file size limit for FAT32) or device (CD-R, etc.), then two-pass might be the right way to go, but for something like a phone with relatively large storage compared to the file size once you re-size the source, you're still better off with a single-pass CRF encode, where 800x480/24p will take up between 500MB and 1GB per hour of movie for very good quality. If that's too big for your taste, change the CRF value to lower the quality. Once you find a CRF value that works for you, stick with it for encodes for the same target.
if you want to go ahead and do work that, chances are, has already been done, and is constantly being done for you by others that create far better quality transcodes.
In general, there are two kinds of rips available on torrents: too large so that quality stays high, and really small to fit on phones or similar devices. It's very hard to find something in-between, where you the encode is close to transparent, but as small as possible. Most rips still use two-pass average bitrate mode, which is basically "find the CRF value that gives me this bitrate", which is the wrong way to maintain quality. The right way is to pick a CRF value and let x264 figure out the bitrate needed for that quality.
Although recent rips are getting the word on how to use x264 correctly, there's an awful lot out there that don't use the right "tune", don't set the correct colorspace, and generally tinker too much with settings that don't matter, while ignoring the ones that do. The forums at doom(9|10).org are a great wealth of information, much of it direct from the x264 developers.
In addition, things like the green tint on "Fellowship of the Ring" can be removed if you do the transcode yourself. And, let's not talk about missing audio or subtitle tracks, glitches in the movie, or the fact that I own at least 10 movies on DVD and Blu-Ray that just aren't available anywhere on the Internet.
That's why video professionals and tv stations rely on hardware based transcoding, and this solutions tend to be expensive.
x264 can encode 1080p in realtime on a modern Intel CPUs (Sandy Bridge, etc.) with pretty much as good a quality for the same bitrate as most hardware solutions. For non-HD, x264 just smokes hardware, as it can do better than realtime encodes at very high quality on those same CPUs.
With QS, I can rip an entire 50GB Blu-Ray in 12 minutes to a 1080p MKV @ 8000kbps. It takes about 16 hours doing the same task with a normal x264 encoder such as Handbrake even though the quality is a little bit better.
Even using the "slower" preset on x264, 1080p encodes take about 3 times as long as the movie, so no more than 8 hours. This is on a slower CPU (since you have QuickSync) than you use, and end up at about 4Mbps
If I used a less-intensive preset, I would get encodes at about the same bitrate as yours, but taking just a little more than the running time of the movie to do it. QuickSync may be even faster, but 3 hours to encode most movies is good enough.
With enough bitrate, anything looks good.
In general, this is true, but very poor encoders can still screw up a high bitrate encode. That's why I use x264...it's going to give me the absolute best quality and lowest bitrate for the amount of encoding time. Since I only encode my movies one time, taking 6-8 hours to do it and getting bitrates as low as 2Mbps for 1080p with high picture quality is worth the extra time. It's also nice to be able to use the full power of AVISynth during the encode. My Blu-Ray rip of "A New Hope" doesn't have the useless "Jabba in the hangar" scene, and I'm working on getting Han shooting first.
No, but it is using it for post-processing such as deinterlacing, noise reduction, etc.
I use the GPU to do FFT noise reduction before some encodes, and it's essentially "free" as it's faster than the 8 threads used by x264 for encoding.
The problem with transcoding is that it exists at all. Strongarm the holdout encoders into using h264 or mp4v with mp4 wrappers, and transcoding will be like... well, like anything no one does anymore.
There will always be transcoding, since you can't fit the 20GB H.264 stream from a Blu-Ray on a phone. And, why would you want to? Resize the 1920x1080 to 800x480 or so, and it will look great on every phone.
For tablets or other devices with more resolution, you still don't need all the bits that most Blu-Ray encodes use. Most are essentially constant bit rate around 25-30Mbps. For movies that are essentially "talking heads" (courtroom dramas like A Few Good Men and Presumed Innocent are the best examples), most of those bits aren't actually adding anything to the picture quality. Even action movies can easily get by with 10Mbps average on a full-resolution transcode, as long as the action scenes get enough bits.