Why not get one of the Zotac IONITX boards and stick it in the M350 case like I did? Fanless, low power, and OS can be installed on SSD, usb flash stick, Microdrive, SD(HC), CF, you name it. They all have usb adapters so that way your system will remain totally silent.

On Mon, Aug 16, 2010 at 9:16 AM, Seppo Ingalsuo seppo.ingalsuo@iki.fi wrote:

...

Fanless, low power, and OS can be installed on SSD, usb flash stick, Microdrive, SD(HC), CF, you name it.  They all have usb adapters so that way your system will remain totally silent.

True, my old IDE HDD is not compatible with new boards, so some flash based storage would be nice.

I've been using flash-based on a few of my systems for a few years now and I've worked pretty much flawlessly. Though on those I have never ventured away from the ext2 filesystem since it has no journaling. However, as cheap as they are now (usb sticks that is), it wouldn't be a big deal to have a couple on stand-by in case on died.

would you like please show qvdpautest results for zotzc ionitx

This is already been posted on a few different forums on the net. Google will get that info quicker I think.

Also, I would think any GT220 has an own audio controller and therefore HDMI audio but this is an assumption. My GT220 does. However, if it doesn't then I'm sure that it has an SPDIF IN connector (whether it's advertised as having one or not). I have at least one other card where there is no mention of SPDIF IN anywhere in any documentation but the connector is in fact there. And I used it to get audio with my DVI out, then a DVI->HDMI cable to get HDMI AUDIO/VIDEO going into my tv.

...

I recommend you to use PCI-E GT220 Geforce card with motheboard which you will choice

Asus Bravo 220 silent seems to be a passive model. Do you know if these non-motherboard integrated cards support 7.1ch PCM audio over HDMI?

seems, doesn't support

I got as PM one tip that pointed to this page

http://en.wikipedia.org/wiki/Nvidia_PureVideo#Table_of_PureVideo_.28HD.29_GP...

I wonder if some smaller number cheaper model with similar VP4/C skills would do? E.g. Asus EN210 Silent? It could also produce less heat? The text in Wikipedia is not perfectly in line with the table so I wonder if GT2xx is needed for best VDPAU support.

It seems that the IONs are VP3/B so your prosal is certainly better (if there is HDMI audio support).

...

this series has very good performance for 1080i temporal_spatial deinterlacing and h264 decoding

have a look on qvdpautest http://linuxdvb.org.ru/wbb/index.php?page=Thread&postID=16142#post16142

Fortunately there's google translate - so there's some worry about passive cooled GT220s? At least the picture of Gigabyte GT220 shows a pretty large heatsink and fan ;^)

yes, I will clarify this worries :)

Goga

Hi,

On Mon, 2010-08-16 at 19:05 +0300, Seppo Ingalsuo wrote:

Asus Bravo 220 silent seems to be a passive model. Do you know if these non-motherboard integrated cards support 7.1ch PCM audio over HDMI?

Yes, as VDR User said, the latest generation VP4/VDPAU feature set C cards (GeForce 210, GT 220, ...) have an integrated audio controller. The ALSA driver should support 7.1 channel PCM too.

Some older cards have an S/PDIF input header but they don't support multichannel PCM, obviously.

I got as PM one tip that pointed to this page

http://en.wikipedia.org/wiki/Nvidia_PureVideo#Table_of_PureVideo_.28HD.29_GP...

I wonder if some smaller number cheaper model with similar VP4/C skills would do? E.g. Asus EN210 Silent? It could also produce less heat? The text in Wikipedia is not perfectly in line with the table so I wonder if GT2xx is needed for best VDPAU support.

IMO the only reason to go for a separate card over ION would be higher quality 1080i deinterlacing. You'll need GT 220 for that since GeForce 210 is only slightly faster than ION. They have the same video processor and therefore the same video decoding capabilities, but post-processing is done on the graphics cores.

It seems that the IONs are VP3/B so your prosal is certainly better (if there is HDMI audio support).

ION2 has VP4/C, I think.

If you don't need advanced 1080i deinterlacing, Asus AT5IONT-I is a very good option right now. It has ION2, latest dualcore Atom and USB3, and it's passively cooled.

--

Niko

On Tue, 2010-08-17 kello 19:36 +0300, Seppo Ingalsuo wrote:

On Tue, 2010-08-17 at 13:39 +0300, Niko Mikkilä wrote:

...

IMO the only reason to go for a separate card over ION would be higher quality 1080i deinterlacing. You'll need GT 220 for that since GeForce 210 is only slightly faster than ION. They have the same video processor and therefore the same video decoding capabilities, but post-processing is done on the graphics cores.

Good point. I think 1080i is rare content for me. If ION is as good as a GT220 with 576i, 720p and 1080p then it could be very suitable for my needs.

ION 2 should be about as good as GT220 for those formats. Older ION works too since you probably don't need all the new features in VDPAU feature set C.

...

If you don't need advanced 1080i deinterlacing, Asus AT5IONT-I is a very good option right now. It has ION2, latest dualcore Atom and USB3, and it's passively cooled.

It's an interesting new board but I wonder if there's some risk for problems. I tried to search but I could not find reports about success with Ubuntu or Linux generally.

Yep, that's always the question with new hardware. I'd expect any showstopping issues to be fixed pretty quickly though, at least on Nvidia's side. Their Linux support has been quite outstanding in my opinion.

Do Nvidia's binary graphics drivers support ION2?

It should. The README lists it as "Second Generation ION". See: http://us.download.nvidia.com/XFree86/Linux-x86/256.44/README/supportedchips...

--

Niko

On Tue, 2010-08-17 at 13:56 +0300, Rene wrote:

I'm also starting to check for a new setup, and i'm on the same line as you: Silent, low-power device that will runn 24/7. I have not yet looked too much around, but there is one thing i'm wondering. What kind of dvb-devices will you use?

I use the multiple vdr instance hack proposed in xineliboutput documentation. My 2xDVB-T(dual tuners) and DVB-S2 cards are PCI and they are in a separate "home server" that runs in non-living areas of the house so noise doesn't matter. There are also several hard disks.

I don't know what happens if that mobo dies. It could be difficult to find motherboards with enough legacy PCI slots. I need four because there's also one extra network card.

My current setup has two pci-cards, Full featured Technoternd dvb-s, and an other TT budget dvb-t. I have to replace the dvb-t card with dvb-c, because i will soon move to an area with cable-tv.. Is it possible to go with completely dvb-usb devices? Are thet reliable enough to not freeze/crash etc?

At least for dish I need PCI type of card because USB cards do not output enough power for a motorized positioner.

BR, Seppo

is there really no recommendation for a board not using Nvidia graphics components? It would really be great to not depend on proprietary drivers.

The VIA chipset VX855 was supposed to have support 1080p support build in. But those devices do not seem to be available in non Asian regions.

AMD/ATI or Intel should also over some products fitting your need. And I heard the drivers matured quite a bit (besides Intel Poulsbo).

Unfortunately I do not own such systems.

I installed vdr-sxfe to Poulsbo/GMA500 netbook but there is no Xv and VA API is not supported by xine-lib. There is mplayer support but I don't know if deintelacing is good. I haven't tried. An unscaled window works but fullscreen is horrible. The real Intel graphics stuff is likely better but I have no experience about that.

Seppo

On 18.08.2010 00:36, Paul Menzel wrote:

What driver did you use?

I used the ppa:gma500/ppa repository as described in https://wiki.ubuntu.com/HardwareSupportComponentsVideoCardsPoulsbo for Ubuntu Lucid.

There is lot of green in the table the real situation for this driver is not as excellent/good. Or if it is "good" then as comparison Nvidia binary driver quality is superior and high above the scale... I have had problems with Nvidia drivers but they have been solved/mitigated to usable level. Even if closed binary drivers are generally bad there are poor and usable examples.

Anyway, »besides Intel Poulsbo« meant this chip is using closed drivers and is not supported by xf86-video-intel and so it does not fit into my wish list.

I fully agree. If I knew this Intel chipset had only closed binary drivers I would have bought my kids another netbook. I bet there will be problems with new X.org and next Ubuntu problems.

BR, Seppo

...

is there really no recommendation for a board not using Nvidia graphics components? It would really be great to not depend on proprietary drivers.

Hardware decoding through VA-API is working on some Intel chipsets and CPUs, but I haven't seen any usable GPU deinterlacing implementations besides those in Nvidia's VDPAU.

I also would like to remind the framerate issues. Naturally you decide what is enough precision and quality for you.

Computer hardware usually cannot provide 50.000Hz, 59.940Hz or 23.976Hz outputs to your TV/Monitor. This will cause some judder on display output as MPEG/AVC input-stream is not synchronized to output framerate.

For example dedicated blu-ray player Philips BDP3000 had a 24.000Hz output (before firmware fix) which resulted a jumped frame every 42 seconds as real input stream is 23.976Hz. It was annoying after you noticed it. But luckily now it is fixed to real 23.976 output.

With this bridge we come to VGA-hardware which might have 50.01Hz closest to 50Hz signal. So every 100 frames we get a jump in picture synch. Ever seen jumping camera panning while watching a film and got annoyed by it?

For ATI cards there is a dynamic framerate fix, unfortunately there is not one for Nvidia cards. Nvidia has good HW acceleration but potentially bad output. With ATI vice versa.

This ATI fix fixes 50.01 by dynamically reprogramming VGA timers so real output is 50.000Hz. (General description, the author can describe more if needed).

So if you aim to HD with full HD quality I'd be carefull with computer outputs. My answer was to setup Popcorn hour to output 50.000Hz with TV-stuff. Unfortunately you lose some easy-of-use with VDR as you need to have separate layer to set up etc.

Also as original question came from Finland and if you use commercial HD channels they are (or at least should be) paired to your receiver.

Happy hunting for HW & SW..

Hi,

ke, 2010-08-18 kello 10:13 +0200, jori.hamalainen@teliasonera.com kirjoitti:

I also would like to remind the framerate issues. Naturally you decide what is enough precision and quality for you.

Computer hardware usually cannot provide 50.000Hz, 59.940Hz or 23.976Hz outputs to your TV/Monitor. This will cause some judder on display output as MPEG/AVC input-stream is not synchronized to output framerate.

If the difference is only few percents or less, this is not a problem for playing back recordings. You can simply sync the video playback to screen refresh and resample audio or drop/duplicate audio samples. That way there is no video judder and only a very slight reduction in audio quality.

Heck, even most of the movies broadcast on PAL TV have been converted from 23.976 fps or 24 fps to 25 fps in a similar fashion; by speeding up and resampling audio.

With this bridge we come to VGA-hardware which might have 50.01Hz closest to 50Hz signal. So every 100 frames we get a jump in picture synch.

You mean every 100 seconds.

Ever seen jumping camera panning while watching a film and got annoyed by it?

That's usually due to a much larger framerate mismatch or failure to configure the video player so that it doesn't drop/duplicate video frames to keep A/V sync.

For ATI cards there is a dynamic framerate fix, unfortunately there is not one for Nvidia cards. Nvidia has good HW acceleration but potentially bad output. With ATI vice versa.

This ATI fix fixes 50.01 by dynamically reprogramming VGA timers so real output is 50.000Hz. (General description, the author can describe more if needed).

The actual problem is that normally neither the video nor the audio outputs in a PC can be synchronized to _live_ broadcast. When the stream is live, one can't replay it faster. Slower playback would be possible with a buffer, but I don't know if anyone has implemented that.

The vga-sync-fields (http://frc.easy-vdr.de/) patch collection provides support for synchronizing the video output exactly to a live transport stream. However, it only works on old pre-Avivo Radeons and Intel i9xx.

I'm guessing that most general-purpose network media players such as Popcorn Hour and PS3 have the same problem with live playback? The only advantage they may have compared to standard PC hardware is that they can have a single clock source for both video and audio, which means that recordings can be played back in sync without dropping or duplicating audio frames.

--

Niko

Thu, 2010-08-19 at 20:54 +0400, Goga777 wrote:

...

Computer hardware usually cannot provide 50.000Hz, 59.940Hz or 23.976Hz outputs to your TV/Monitor. This will cause some judder on display output as MPEG/AVC input-stream is not synchronized to output framerate.

do you mean that all nvidia vdpau cards with existing drivers from Nvidia can't provide exact 50.000Hz, 59.940Hz or 23.976Hz ??

There is no graphics card, BD/DVD player or other standalone device that outputs those rates exactly. I don't know how much they deviate, but I'd guess it's usually something like 0.01 % (50.005 Hz instead of 50 Hz), as Jori said.

However, the rate doesn't need to match exactly because the display device is synchronized to the video signal. The rate could be 50.1 Hz or maybe even 51 Hz and the display wouldn't mind. 50 fps video files would play slightly faster, but there would be no need to drop video frames because of that.

Things are more problematic when receiving live broadcast. Then the display and the video source (graphics card and software) needs to be synchronized to the broadcast to avoid dropping or duplicating frames. Set-top digital television boxes and FF DVB cards do that, but most graphics cards/drivers can't because they aren't designed to follow an external time source.

Audio playback synchronation is another issue, and somewhat difficult to handle properly on a PC where the audio chip's clock is almost always separate from the graphics card's clock. By default, many media players time everything according to the audio clock, and therefore they need to drop/duplicate video frames every now and then. The other alternative is to drop/duplicate audio frames or resample the audio completely.

--

Niko

Fri, 2010-08-20 at 10:08 +0200, jori.hamalainen@teliasonera.com wrote:

...

There is no graphics card, BD/DVD player or other standalone device that

outputs

...

those rates exactly. I don't know how much they deviate, but I'd guess

it's usually

...

something like 0.01 % (50.005 Hz instead of 50 Hz), as Jori said.

If you can find a modeline what your output is currently using you can use online services to check framerates it provides.

"xvidtune -show" prints the current modeline, but the refresh rate calculated from it will not be the same as the actual rate. The clock chrystals have deviations in order of 0.01 % and the only way to fix that is by adjusting the timing slightly at frequent intervals as vga-sync-fields does.

Having a more accurate time source might not help either because the broadcast may not be accurately timed.

It does not help using direct Toslink-output from VDR which mostly prohibits resampling of audio.

Audio can be resampled with any PCM output. The idea is to adjust the playback speed slightly by resampling for example 48 kHz audio to 48.01 kHz and then playing the result back at 48 kHz.

This is more difficult if you'd like to play multi-channel AC-3 with S/PDIF passthrough; then the audio would need to be decoded, resampled and re-encoded. An easier alternative is to drop/duplicate AC-3 packets, but that may be more noticeable on playback.

Both Xine and XBMC support resampling audio. XBMC also supports dropping/duplicating audio packets.

And why would you like to have audio decompressed, speeded up 1% and then recompressed.

Just to avoid your output software to duplicate or drop frames. Synch perfectly..

Yep, that's the best way to get fluid video playback with synchronized audio on current PC hardware.

Re-encoding is unnecessary with stereo streams or HDMI1.3 and analog multichannel outputs. High-quality resampling is totally unnoticeable.

--

Niko

...

If you can find a modeline what your output is currently using you can use online services to check framerates it provides.

"xvidtune -show" prints the current modeline, but the refresh rate calculated from it will not be the same as the actual rate.

My used modeline is (1080p60) $ xvidtune -show -display :0.0 "1920x1080" 148.50 1920 2008 2052 2200 1080 1084 1089 1125 +hsync +vsync

Which compiles to 60Hz output on theory level at epanorama calculator. If it is 60.001Hz because of clock Inaccuracies the TV will synch to it, no problem. As I use my VDR's X-output only for TV-based web browsing with randon youtube/vimeo.

However I am now at office, so I cannot use "get-edid" to read all modelines TV is announcing for PAL-modes/timings. And how Nvidia is altering the proposed timigs as horizontal synchs must be at multitude of 8 (so you cannot have f.ex resolution 1924x1080).

- Jori

On Fri, Aug 20, 2010 at 01:37:10AM +0300, Niko Mikkilä wrote:

Thu, 2010-08-19 at 20:54 +0400, Goga777 wrote:

...

...

Computer hardware usually cannot provide 50.000Hz, 59.940Hz or 23.976Hz outputs to your TV/Monitor. This will cause some judder on display output as MPEG/AVC input-stream is not synchronized to output framerate.

do you mean that all nvidia vdpau cards with existing drivers from Nvidia can't provide exact 50.000Hz, 59.940Hz or 23.976Hz ??

There is no graphics card, BD/DVD player or other standalone device that outputs those rates exactly. I don't know how much they deviate, but I'd guess it's usually something like 0.01 % (50.005 Hz instead of 50 Hz), as Jori said.

However, the rate doesn't need to match exactly because the display device is synchronized to the video signal. The rate could be 50.1 Hz or maybe even 51 Hz and the display wouldn't mind. 50 fps video files would play slightly faster, but there would be no need to drop video frames because of that.

Things are more problematic when receiving live broadcast. Then the display and the video source (graphics card and software) needs to be synchronized to the broadcast to avoid dropping or duplicating frames. Set-top digital television boxes and FF DVB cards do that, but most graphics cards/drivers can't because they aren't designed to follow an external time source.

Audio playback synchronation is another issue, and somewhat difficult to handle properly on a PC where the audio chip's clock is almost always separate from the graphics card's clock. By default, many media players time everything according to the audio clock, and therefore they need to drop/duplicate video frames every now and then. The other alternative is to drop/duplicate audio frames or resample the audio completely.

I assume you guys are aware of projects like: http://frc.easy-vdr.de/

It was originally started to get perfectly synced RGB output from a VGA card (to PAL TV), just like from FF DVB card.

I haven't really used that myself, but afaik they've been working on making that exact synchronization (variable framerate) possible with new HD/VGA/DVI outputs aswell.

-- Pasi

On Wed, Aug 18, 2010 at 10:13:56AM +0200, jori.hamalainen@teliasonera.com wrote:

Computer hardware usually cannot provide 50.000Hz, 59.940Hz or 23.976Hz outputs to your TV/Monitor. This will cause some judder on display output as MPEG/AVC input-stream is not synchronized to output framerate.

it's correct that computer hardware normally cannot provide 50.000Hz. Nor can it adapt the framerate dynamically as I did for some handpicked hardware in my sync-fields-project [1].

But even in the case of non synchronized output framerate (VDPAU) it does not necessarily mean that judder will result.

Nvidia graphics use excellent deinterlacers. These produce a stream of progressive frames with field frequency (e.g. 50Hz) out of an interlaced source.

Single frame doubler/losses at 50Hz are not visible to the human eye. To put that in other words: VDPAU uses some brute force (deinterlacing) to work around their design flaw (missing synchronization). In practice this works very well.

Surely it's not an optimal solution: the price for that is excess graphics power requirement if you think in terms of 5W units. But nonetheless the smoothness of picture flow is excellent.

If some people report they observe judder though this mostly is a result of a wrong setup (xorg.conf) or use of temporarily broken beta software (xine derivates). With VDPAU design there exists nothing like inherent judder.

For the future I wish there once will become available some graphics with native dynamic framerate support. Never seen such thing until today.

cheers Thomas

[1] http://lowbyte.de/vga-sync-fields/vga-sync-fields/

On Wed, Aug 25, 2010 at 11:16:37PM +0300, Prelude wrote:

I have suffered this annoying judder in both live and recording viewing. My config_xineliboutput: http://pastebin.com/kxe7RvX0 xorg.xonf: http://pastebin.com/6fCvWvxU

sorry, from that configuration alone I can't tell you what's going wrong.

When I started to setup my VDPAU machines (end of last year) I first made sure that the Xserver really is running a 50Hz modeline. I mostly play 25/50Hz streams. This way I found an Xserver bug/misbehavior. Only if I explicitly set the modeline params manually in xorg.conf I get real 50Hz output (instead of 60Hz). Please see attached my xorg.conf. Always try:

DISPLAY=:0 nvidia-settings --query RefreshRate

to verify. Even the Xorg.0.log lies about the fact that 50Hz are active but in reality it uses 60Hz. Maybe nVidia has fixed that in the meantime.

After this tuning to a channel with live ticker gives me perfect results. There is not any judder at all.

The problem with all this juddering is that the symptom 'judder' alone gives you no hint at all about the root cause of the problem.

Maybe you first verify your current nvidia-settings result.

cheers Thomas

Hi,

Am 26.08.10 12:16, schrieb Laz:

Then my old TV blew up and I now have a full-HD LCD TV. I've been using the same setup with the new TV and it still works pretty well for the SD signals I can currently receive.

if you new LCD with HDMI use it to connect your HDMI output of the ION board with your TV. Works perfectly here. Xorg works perfectly with it (even sound over hdmi had worked till I upgrade the firware of my Denon receiver :( ).

Here the settings from my xorg.conf: Section "Monitor" Identifier "Monitor0" VendorName "Monitor Vendor" ModelName "Monitor Model" EndSection Section "Device" Identifier "Card0" Driver "nvidia" VendorName "nVidia Corporation" BoardName "Unknown Board" BusID "PCI:3:0:0" EndSection Section "Screen" Identifier "Screen0" Device "Card0" Monitor "Monitor0" SubSection "Display" Viewport 0 0 Depth 1 EndSubSection SubSection "Display" Viewport 0 0 Depth 4 EndSubSection SubSection "Display" Viewport 0 0 Depth 8 EndSubSection SubSection "Display" Viewport 0 0 Depth 15 EndSubSection SubSection "Display" Viewport 0 0 Depth 16 EndSubSection SubSection "Display" Viewport 0 0 Depth 24 EndSubSection EndSection

I use the xineliboutput plugin.

Bye, Matthias

On Thu, Aug 26, 2010 at 03:06:51PM +0100, Tony Houghton wrote:

Are you aware of the interaction between DynamicTwinView and XRandR? The

no. But if Xorg.0.log tells me (in the bad case):

(II) NVIDIA(0): Validating Mode "1920x1080": (II) NVIDIA(0): 1920 x 1080 @ 50 Hz (II) NVIDIA(0): For use as DFP backend. (II) NVIDIA(0): Mode Source: EDID (II) NVIDIA(0): Pixel Clock : 148.50 MHz (II) NVIDIA(0): HRes, HSyncStart : 1920, 2448 (II) NVIDIA(0): HSyncEnd, HTotal : 2492, 2640 (II) NVIDIA(0): VRes, VSyncStart : 1080, 1084 (II) NVIDIA(0): VSyncEnd, VTotal : 1089, 1125 (II) NVIDIA(0): H/V Polarity : +/+ (II) NVIDIA(0): Mode is valid. [...] (II) NVIDIA(0): "1920x1080_50" : 1920 x 1080 @ 50.0 Hz (from: EDID) [...] (II) NVIDIA(0): Config Options in the README. (II) NVIDIA(0): Setting mode "1920x1080_50"

and it uses a 60Hz modeline though this is clearly a Xserver bug for me.

- Thomas

On Thu, Aug 26, 2010 at 04:31:28PM +0200, Paul Menzel wrote:

Do you know if it has been reported to Nvidia? I think the X.org people would not look at it, since you use the proprietary driver.

I dont't know if especially this problem has been reported to nVidia. There are heaps of error reports of that kind reported to nVidia that all sound quite similar.

- Thomas

On Thu, Aug 26, 2010 at 11:23:22PM +0200, Paul Menzel wrote:

Thomas, having just

VertRefresh 50.0

in your `xorg.conf`, did you measure the 60 Hz directly at the output or were you just saying the tools reported 60 Hz?

without specifying the modeline directly in xorg.conf (the bad case) the tool reported 60Hz, Xorg.0.log reports 50Hz and at the same time it was juddering like mad. This is clearly a bug for me because the Modes "1920x1080_50" setting is ignored and on top of that Xorg.0.log lies about this.

After specifying the modeline directly in xorg.conf (the workaround) the tool and Xorg.0.log reported 50Hz. And live ticker scrolled smoothly without any jumping.

For me this is evidence enough the tool reported the right thing in both cases.

cheers Thomas

...

to verify. Even the Xorg.0.log lies about the fact that 50Hz are active but in reality it uses 60Hz. Maybe nVidia has fixed that in the meantime.

To watch HD broadcasts I use the hdmi output on my laptop. It isn't configured in xorg but I just configure it on the fly with nvidia-settings in twin-view mode (the laptop panel at 60Hz and the hdmi port at 50Hz). Provided that I disable "Force Full GPU Scaling", the tv reports a 50Hz signal.

if you disabled full gpu scaling , how upscaling will work really ? what and how will upscale SD video to 1080p ?

Goga

On Tue, Aug 17, 2010 at 8:13 AM, Paul Menzel paulepanter@users.sourceforge.net wrote:

is there really no recommendation for a board not using Nvidia graphics components? It would really be great to not depend on proprietary drivers.

Just wondering why you even care about this... The Nvidia drivers work well for me and honestly that's all I care about.

On Tue, Aug 17, 2010 at 2:45 PM, Paul Menzel paulepanter@users.sourceforge.net wrote:

...

Just wondering why you even care about this...  The Nvidia drivers work well for me and honestly that's all I care about.

Good for you. As I wrote I do not want to start a flame war. But you asked. ;-)

Everyone should decide for himself what he wants. I am on the side of Free Software and believe that it has more advantages than just »it works«.

For example I believe supporting hardware requiring non-free software puts more work to the distributions and people working on those because they have to answer a lot of support request when people contact them and not Nvidia and they cannot fix problems because they do not have access to the code. Secondly I guess you will find a lot of requests for fixes the Nvidia driver which have not been accomplished.

I think you will find a lot of information about this whole issue on the Internet and others can lay down the points better than I can. But as I said, it is just an axiom, so you cannot argue what is right or wrong.

I just wondered what your reasoning is, nothing more. There's absolutely no reason you should assume it's attempt to "start a flame war" or "argue". There is no crime in people having different points of view. I've now lost interest in this thread but thanks for sharing your opinion. Good luck in your qwest.

Cheers, Derek