I would agree that the initial effort was to recreate existing unix kernels; there's a lot of baseline functionality that is common to most any OS, so there will be nothing innovative about that. However, as the kernel grows and the complexity of various problems increases, the level of original thought (such as that put into the VM system) increases, which is what i was trying to get at with the 'innovative' comment.
True, i'm not about to mail timothy personally and tell him to change his ways, i just get annoyed and deterred when i see the GNU/Linux name. It's like it wasn't enough for RMS that the linux kernel was released under the GPL and was a major catalyst for open source in general.
I'm sorry, i can't stand it when people call Linux GNU/Linux. I fail to see why people feel the need to bow to RMS's ego; the GNU utilities are primarily rewrites of existing utilities, not innovative new technologies. All the comparisons of lines of code are pointless - the kernel is one single chunk, making it much much more complex to work with than any of the hundreds of GNU utilities packages in the standard distro.
Seriously, if you are going to start referring to it as GNU/Linux, you should change your website's name to Apache/Slashdot; maybe you should start telling people who use slashcode that they have to have 'Slash/' at the beginning of their website name. It's just as retarded.
The strongarm is 206mhz but other than that you are right on. The cirrus processor is used in rio portables, nomad portables, etc - it's basically _the_ chip right now.
As far as vorbis, there are companies willing to pay development time for an integerized decoder, but there are a limited number of people who can do it (i would guess 3-6 people, monty being one of them, and everybody else would be error prone due to the lack of a spec).
The 9312 isn't released yet, so i dont know where you came up with $50 for it. Also i wasn't aware it had floating point. If you are paying $20 for a 7312 you are paying too much (or ordering in very very small quantities), you can get a strongarm for that price. Lastly i'm fairly certain that the cost on the 7312 includes the cost for the ARM integerized mp3 decoder, you should talk to your cirrus rep about that.
I tend to disagree, if ogg is still in development then why are they licensing it? Presumably they feel it is stable enough to ship in devices such as the zaurus.
where did you get this beta firmware? it was never released outside of a controlled audience. monty and jack (jack works on the vorbis project) were allowed to have a copy; iomega granted permission for both of them to have a copy of the firmware, but no rights were given to redistribute it. this was primarily an advocacy item.
i built and burned the images with vorbis support as part of this arrangement. there were also demo players furnished to iomega, and for a while they were discussing releasing a tested approved firmware update with vorbis support, but that fell through.
so i'll ask again, where did you get that firmware image from?
This is incorrect; the strongarm 206mhz processor in the zaurus is not fast enough to decode ogg real time using floating point emulation. It's close, but not quite.
Pocket MP3 players typically use a cirrus 7212 74mhz arm processor, which are much faster than you describe (decode at less than 60% of the cpu). Also, as a side note, decoding 192kbps is easier than decoding, say, 96kbps; you just spend more time pushing the data through.
This is incorrect; the xiph integerized codebase is pure C. There is another integerized vorbis decoder which only plays up to beta4 content available at http://ivdev.sourceforge.net, but this is not endorsed or supported by xiph.
Yes, there is code to enable your iomega hipzip player to play ogg vorbis content from beta4. Unfortunately iomega never authorized the release of this code, so end users have never been allowed to use this.
wow it has? because, to my knowledge, that software was never generally released, and the only parties that were allowed to have a copy of that image were jack moffit and chris montgomery.
not that i should know or anything, since i built that image and burned it into the devices. hello.
I hate to suggest it, but perhaps links to sites that can't sustain a/. load shouldn't be posted? Besides the fact that nobody can check this site out after 2 minutes, it's undoubtedly difficult for the webmaster of said site (especially if they have bandwidth limitations, etc).
For all intents and purposes, you need at least a p133 to decode mp3 real time. You can do it on slower processors, you just end up ditching some basic audio quality.
Yeah, apple included the "Don't Steal Music" copy protection sticker on every iPod. That's it.
Re:It cannot use Ogg Vorbis encoding, can it?
on
iPod on Windows
·
· Score: 1
If you really encode at that bitrate, dont expect any portable to fill your needs. The iPod would get you about 5-6 hours battery life for 320kbps ogg content, since it has to buffer more often.
The shuttle cases use flex-atx motherboards and mounts, it is a standard just not one that is widely used by other companies. I searched around and found about 3-4 flex-atx form factor mobos, most of which have less functionality than the shuttle boards.
I am intruiged by your ideas and would like to subscribe to your newsletter.
In all seriousness though, this looks like a pretty cool product, but i would have different needs for it. Specifically, you guys are using...cs8900a or something for ethernet? have you looked at putting a pcmcia slot on it and writing drivers for 802.11b cards? I dont have ethernet running through my house, but i would be willing to do wireless or HPNA to get something like this to work.
By the way, on your web page, it says you use a cirrus 4334 dac; isn't that a 43l43?;)
Please excuse my mistake. It wasn't the manufacturer junking inventory, it was evidentally the retailer; from what i can tell best buy no longer carries the audiotron.
The whole basis for my comment was the $150 price one person mentioned; the device itself probably costs $100-$120 to manufacture, there's no way they can sell it for $150 on the street and make money.
The reason people don't do decoder tests on embedded devices is that everybody knows decoders on embedded devices suck. There's no floating point, so everything is integerized, and there is limited processing power, so shortcuts have to be taken in order to decode in real time. For the most part, this doesn't matter, because the expected audience is using headphones or cheap speakers instead of a nice stereo system.
The Rio Riot is just a Rio 800 with a hard drive. Same processor, same DAC, same decoders. The iPod is a different animal, i dont know about the dac, but i do know that they have two arm7 cores instead of just one. I'm guessing they use the same decoders as the riot though.
Dont get your hopes up with strongarm. It's faster than what they use in the riot (cirrus logic @ 73mhz vs strongarm @ 206mhz), but there still isn't an FPU. Also, dont get your hopes up with vorbis on that, unless someone writes an integerized version of the libraries (or updates the open source ones to rcX; search for 'integerized vorbis' on sourceforge.net). The soft FP stuff isn't fast enough to decode vorbis real time on the SA.
Really, honestly, you're not going to get 24 bit audio out of an embedded device. At least not now. And you're definitely not going to get a good listening experience plugging sennheisers or grados into a portable device. It's just too expensive to put high quality audio components into a portable device.
Slashdot Mirror
I'm guessing you haven't seen this:
. asp?RecNo=713&Language=English
http://club.aopen.com.tw/News/News_showAnswer_New
It's a motherboard with onboard tube amplified audio.
the same as this? http://slashdot.org/article.pl?sid=02/06/13/147213 &mode=nested&tid=126
I would agree that the initial effort was to recreate existing unix kernels; there's a lot of baseline functionality that is common to most any OS, so there will be nothing innovative about that. However, as the kernel grows and the complexity of various problems increases, the level of original thought (such as that put into the VM system) increases, which is what i was trying to get at with the 'innovative' comment.
True, i'm not about to mail timothy personally and tell him to change his ways, i just get annoyed and deterred when i see the GNU/Linux name. It's like it wasn't enough for RMS that the linux kernel was released under the GPL and was a major catalyst for open source in general.
I'm sorry, i can't stand it when people call Linux GNU/Linux. I fail to see why people feel the need to bow to RMS's ego; the GNU utilities are primarily rewrites of existing utilities, not innovative new technologies. All the comparisons of lines of code are pointless - the kernel is one single chunk, making it much much more complex to work with than any of the hundreds of GNU utilities packages in the standard distro.
Seriously, if you are going to start referring to it as GNU/Linux, you should change your website's name to Apache/Slashdot; maybe you should start telling people who use slashcode that they have to have 'Slash/' at the beginning of their website name. It's just as retarded.
The strongarm is 206mhz but other than that you are right on. The cirrus processor is used in rio portables, nomad portables, etc - it's basically _the_ chip right now.
As far as vorbis, there are companies willing to pay development time for an integerized decoder, but there are a limited number of people who can do it (i would guess 3-6 people, monty being one of them, and everybody else would be error prone due to the lack of a spec).
The 9312 isn't released yet, so i dont know where you came up with $50 for it. Also i wasn't aware it had floating point. If you are paying $20 for a 7312 you are paying too much (or ordering in very very small quantities), you can get a strongarm for that price. Lastly i'm fairly certain that the cost on the 7312 includes the cost for the ARM integerized mp3 decoder, you should talk to your cirrus rep about that.
I tend to disagree, if ogg is still in development then why are they licensing it? Presumably they feel it is stable enough to ship in devices such as the zaurus.
where did you get this beta firmware? it was never released outside of a controlled audience. monty and jack (jack works on the vorbis project) were allowed to have a copy; iomega granted permission for both of them to have a copy of the firmware, but no rights were given to redistribute it. this was primarily an advocacy item.
i built and burned the images with vorbis support as part of this arrangement. there were also demo players furnished to iomega, and for a while they were discussing releasing a tested approved firmware update with vorbis support, but that fell through.
so i'll ask again, where did you get that firmware image from?
This is incorrect; the strongarm 206mhz processor in the zaurus is not fast enough to decode ogg real time using floating point emulation. It's close, but not quite.
Pocket MP3 players typically use a cirrus 7212 74mhz arm processor, which are much faster than you describe (decode at less than 60% of the cpu). Also, as a side note, decoding 192kbps is easier than decoding, say, 96kbps; you just spend more time pushing the data through.
This is incorrect; the xiph integerized codebase is pure C. There is another integerized vorbis decoder which only plays up to beta4 content available at http://ivdev.sourceforge.net, but this is not endorsed or supported by xiph.
Yes, there is code to enable your iomega hipzip player to play ogg vorbis content from beta4. Unfortunately iomega never authorized the release of this code, so end users have never been allowed to use this.
wow it has? because, to my knowledge, that software was never generally released, and the only parties that were allowed to have a copy of that image were jack moffit and chris montgomery.
not that i should know or anything, since i built that image and burned it into the devices. hello.
I hate to suggest it, but perhaps links to sites that can't sustain a /. load shouldn't be posted? Besides the fact that nobody can check this site out after 2 minutes, it's undoubtedly difficult for the webmaster of said site (especially if they have bandwidth limitations, etc).
For all intents and purposes, you need at least a p133 to decode mp3 real time. You can do it on slower processors, you just end up ditching some basic audio quality.
The flipside of that coin is if we give $300,005 to hollings, will he drop support for this crap bill?
Money, indeed, is the ultimate corruptor.
That's so funny, i was just looking at contribution info for senator hollings (the carolina looney backing the senate version of the bill).
His top contributors are lawyers/law firms at $1,158,134; second are tv/movies/music at $264,534.
I was kicking around the same thought you had above; bad bills help lawyers the most. I wonder what other bills hollings supports.
Another tidbit; 2/3rds of hollings contributions come from out of state. Pretty sweet.
Yeah, apple included the "Don't Steal Music" copy protection sticker on every iPod. That's it.
If you really encode at that bitrate, dont expect any portable to fill your needs. The iPod would get you about 5-6 hours battery life for 320kbps ogg content, since it has to buffer more often.
Soundblaster audigy cards come with firewire on them. I've also heard that recent PCs come with firewire ports now.
The shuttle cases use flex-atx motherboards and mounts, it is a standard just not one that is widely used by other companies. I searched around and found about 3-4 flex-atx form factor mobos, most of which have less functionality than the shuttle boards.
I am intruiged by your ideas and would like to subscribe to your newsletter.
;)
In all seriousness though, this looks like a pretty cool product, but i would have different needs for it. Specifically, you guys are using...cs8900a or something for ethernet? have you looked at putting a pcmcia slot on it and writing drivers for 802.11b cards? I dont have ethernet running through my house, but i would be willing to do wireless or HPNA to get something like this to work.
By the way, on your web page, it says you use a cirrus 4334 dac; isn't that a 43l43?
http://ivdev.sourceforge.net
only plays up to b4 content, but hey, what do you expect for free.
Please excuse my mistake. It wasn't the manufacturer junking inventory, it was evidentally the retailer; from what i can tell best buy no longer carries the audiotron.
The whole basis for my comment was the $150 price one person mentioned; the device itself probably costs $100-$120 to manufacture, there's no way they can sell it for $150 on the street and make money.
The reason people don't do decoder tests on embedded devices is that everybody knows decoders on embedded devices suck. There's no floating point, so everything is integerized, and there is limited processing power, so shortcuts have to be taken in order to decode in real time. For the most part, this doesn't matter, because the expected audience is using headphones or cheap speakers instead of a nice stereo system.
The Rio Riot is just a Rio 800 with a hard drive. Same processor, same DAC, same decoders. The iPod is a different animal, i dont know about the dac, but i do know that they have two arm7 cores instead of just one. I'm guessing they use the same decoders as the riot though.
Dont get your hopes up with strongarm. It's faster than what they use in the riot (cirrus logic @ 73mhz vs strongarm @ 206mhz), but there still isn't an FPU. Also, dont get your hopes up with vorbis on that, unless someone writes an integerized version of the libraries (or updates the open source ones to rcX; search for 'integerized vorbis' on sourceforge.net). The soft FP stuff isn't fast enough to decode vorbis real time on the SA.
Really, honestly, you're not going to get 24 bit audio out of an embedded device. At least not now. And you're definitely not going to get a good listening experience plugging sennheisers or grados into a portable device. It's just too expensive to put high quality audio components into a portable device.