4 ms·

This seems correct, but probably leaves most readers with the wrong impression. Theora is a codec that is free and open, and the patents that cover it do not en

This seems correct, but probably leaves most readers with the wrong impression. Theora is a codec that is free and open, and the patents that cover it do not encumber it (since they are also freely licensed), so I'd consider it patent unencumbered. https://www.theora.org/benefits/ https://www.theora.org/benefits/

scarface_74 3y ago
So do you think it would be a good use of die space on a processor to support a substandard by today’s standard codec that was last used a decade ago? It is a complete non starter for Apple or any mobile device maker to build in support for a non hardware optimized codec on a mobile device. How does that benefit users besides the ones who bemoan not being able to run Linux on the Apple Pencil?
- rpdillon 3y ago
  I was simply responding to the query about a non-patent-encumbered codec.
- jchw 3y ago
  > So do you think it would be a good use of die space on a processor to support a substandard by today’s standard codec that was last used a decade ago? There's a specific reason why Theora was last used a decade ago... Apple. Apple is the reason. There was some other opposition, but in 2009, Apple drove the hardest against Theora being included in the HTML5 spec, and they ultimately won, leaving us with no standard codec everyone could actually agree on. Because the web is supposed to be open, and h264 is not. It being excluded from the spec, you could not rely on Theora support in the browser. > It is a complete non starter for Apple or any mobile device maker to build in support for a non hardware optimized codec on a mobile device. Given that a RPi 5 takes less than a full core to decode h264, I'm not even sure this is a real concern. In my tests, ogv.js is able to decode VP9 at modern resolutions on an iPad without even really making the device warm; it's undoutedly inefficient, but the fact that it can even be done using WebAssembly gives me a feeling that this isn't really the huge problem it's made out to be. For Theora... I mean. It's just not an issue. You can just do it in software. Maybe not in 2009, but certainly today, right now, on any hardware made in the past few years, on one core, without even driving the clocks up. For AV1, yes, sure, this is a real issue; but Apple has been a part of the AOM for years, so I have no idea why they are building machines without hardware AV1 acceleration today. That's a question that only Apple can answer, given that it has real adoption already. > How does that benefit users besides the ones who bemoan not being able to run Linux on the Apple Pencil? Yes sure, let's take a random unrelated detour to make fun of tinkerers who like to be able to reads notes run their own software on hardware they purchased. I'm disappointed, although unfortunately not surprised at all.
  - scarface_74 3y ago
    > There's a specific reason why Theora was last used a decade ago... Apple. Apple is the reason The last stable release was in 2009. The iPhone was still a bit player then, only available via a few carriers and still behind RIM in market share. No one cared about Apple’s opinion in 2009. > For AV1, yes, sure, this is a real issue; but Apple has been a part of the AOM for years, so I have no idea why they are building machines without hardware AV1 acceleration today. That's a question that only Apple can answer, given that it has real adoption already. https://bitmovin.com/av1-playback-support https://bitmovin.com/av1-playback-support > Yes sure, let's take a random unrelated detour to make fun of tinkerers who like to be able to reads notes run their own software on hardware they purchased. I'm disappointed, although unfortunately not surprised at all. How is not supporting an “open standard” stopping anyone from “tinkering”? There are open source implementations of the encoders and they will be decoded in hardware once you install whatever you are tinkering with on your phone.

scarface_74 3y ago

So do you think it would be a good use of die space on a processor to support a substandard by today’s standard codec that was last used a decade ago? It is a complete non starter for Apple or any mobile device maker to build in support for a non hardware optimized codec on a mobile device. How does that benefit users besides the ones who bemoan not being able to run Linux on the Apple Pencil?

rpdillon 3y ago
I was simply responding to the query about a non-patent-encumbered codec.
jchw 3y ago
> So do you think it would be a good use of die space on a processor to support a substandard by today’s standard codec that was last used a decade ago? There's a specific reason why Theora was last used a decade ago... Apple. Apple is the reason. There was some other opposition, but in 2009, Apple drove the hardest against Theora being included in the HTML5 spec, and they ultimately won, leaving us with no standard codec everyone could actually agree on. Because the web is supposed to be open, and h264 is not. It being excluded from the spec, you could not rely on Theora support in the browser. > It is a complete non starter for Apple or any mobile device maker to build in support for a non hardware optimized codec on a mobile device. Given that a RPi 5 takes less than a full core to decode h264, I'm not even sure this is a real concern. In my tests, ogv.js is able to decode VP9 at modern resolutions on an iPad without even really making the device warm; it's undoutedly inefficient, but the fact that it can even be done using WebAssembly gives me a feeling that this isn't really the huge problem it's made out to be. For Theora... I mean. It's just not an issue. You can just do it in software. Maybe not in 2009, but certainly today, right now, on any hardware made in the past few years, on one core, without even driving the clocks up. For AV1, yes, sure, this is a real issue; but Apple has been a part of the AOM for years, so I have no idea why they are building machines without hardware AV1 acceleration today. That's a question that only Apple can answer, given that it has real adoption already. > How does that benefit users besides the ones who bemoan not being able to run Linux on the Apple Pencil? Yes sure, let's take a random unrelated detour to make fun of tinkerers who like to be able to reads notes run their own software on hardware they purchased. I'm disappointed, although unfortunately not surprised at all.
- scarface_74 3y ago
  > There's a specific reason why Theora was last used a decade ago... Apple. Apple is the reason The last stable release was in 2009. The iPhone was still a bit player then, only available via a few carriers and still behind RIM in market share. No one cared about Apple’s opinion in 2009. > For AV1, yes, sure, this is a real issue; but Apple has been a part of the AOM for years, so I have no idea why they are building machines without hardware AV1 acceleration today. That's a question that only Apple can answer, given that it has real adoption already. https://bitmovin.com/av1-playback-support https://bitmovin.com/av1-playback-support > Yes sure, let's take a random unrelated detour to make fun of tinkerers who like to be able to reads notes run their own software on hardware they purchased. I'm disappointed, although unfortunately not surprised at all. How is not supporting an “open standard” stopping anyone from “tinkering”? There are open source implementations of the encoders and they will be decoded in hardware once you install whatever you are tinkering with on your phone.