USB VIDEO ON THE DENEB?

[alexh]

Chris, if you are interested in developing a class for USB->VGA adapter then gimme a PM.

Several of my closest friends work as hardware designers for DisplayLink I am pretty sure I could get you a developer kit including driver source for linux.

[alexh]

Most DVB-T PCI cards and (all?) DVB-T USB dongles require MPEG2 software decode. Not really something the Amiga could do well eh?

[alexh]

wawrzon wrote:
there are a lot pci tv_cards that are advertised as equipped with hardware mpeg2 decoder for a better performance

I wouldn't say there was "a lot" of DVB-T PCI TV cards in the first place let alone "a lot" with hardware MPEG decoders.

The majority of DVB cards nowadays use software decoders. (Although quite a few have hardware ENCODERS)

Hauppauge (Pronounced Hop-pog) are the biggest seller of TV cards and in 2005 (when I left the TV hardware industry) all but one of their cards used software decoders.

It's just economics. The cards are much cheaper without one and the PC's today are more than capable of software SDTV decode.

[alexh]

platon42 wrote:
A 50Hz true color stream would be at a sustained rate of 260MBit/sec.

A RAW decoded digital TV stream would be 601 video (16-bit 4:4:4 Y'CbCr) at PAL SD resolution/refresh would be 720x288x16x50 = 158MBit which would be about manageable with USB2.0 HS

Soon as you go to Desktop size and modern refresh rates then uncompressed RAW goes out of the window for almost any link 1024x768x24x75 = 1.3GBit/s

But I am pretty sure these USB gfx cards do not just get streamed raw RGB data. I would imagine they are very similar in architecture to PCI gfx cards with their own framebuffer RAM and hardware acceleration such as a blitter and line drawing etc. To cope with the much lower bandwidth vs 32-bit 33MHz PCI I would imagine the host tries to buffer and compress the data prior to transfer.

Just gotta hope the compression they perform on the host isn't too CPU hungry for a high end Amiga :-(

[alexh]

I've discovered the server software for all DisplayLink products is proprietory. They are not going to release the code in any form other than binaries for windows :-(

[alexh]

arnljot wrote:
Could your friends indicate how they operate. how much work is on the device and how much in software.

I got a tidbit of info. From the small conversation we had, theirs appears to be about 80% software and 20% hardware.

arnljot wrote:
Also, could some linux sweet talk work, or that it's amiga so no threat?

They said no. But I have just heard from one of them that they have already ported their work to MacOSX
 

arnljot wrote:
Could they at least indicate which chips are used so manuals could be sourced?

Erm, they use THEIR chips. They do not sell products direct (AFAIK) they sell a HW & SW solution which companies integrate into their products. They do not distribute anything but high-level datasheets.

http://techgage.com/article/displaylink_-_multi-display_setups_with_usb/1

[alexh]

Jose wrote:
Isn't a VGA usb card sending almost raw data too ?

If they work like I think they work, then some composition and all the refresh is done by USB devices themselves reducing the amount of bus bandwidth. Once the gfx data is transferred to it's RAM unlike RAW video it does not need 100% screen updates every 1/25th of a second if the image does not change.

For VGA usb cards I fear that the host has to compress this data on the fly. Perhaps too much for all but the most powerful Amiga. But we'll see.

VNC works on the Amiga (just) and that uses an 11Mbit ethernet connection.

[alexh]

platon42 wrote:
maybe the registers of the DVI encoder chip are directly accessible over control transfers. If that's the case, luckily the docs of the chip are available on the net.

The docs of the DVI chip?

platon42 wrote:
It uses the LibUSB, which is LGPL & GPL, so at least here we could demand the source code for. I doubt, it will be any use though.

That is not a bad idea. In our office some managers are dreading a phone call asking about this subject. GPL software has been used in some products and legally no-engineer knows where we stand when it comes to releasing part/all of our source code.

[alexh]

platon42 wrote:
BTW: x86 assembler sucks big time.

You looking at the Intel MacOS drivers?

DisplayLinkGA
DisplayLinkDriver

I thought with them being beta they might have accidentally left in source level debugging but no such luck!

[alexh]

The MacOSX intel beta drivers would probably be a better place to start? They have no hardware acceleration etc.

Use TransMac to open the .dmg stuffit file and 7Zip to open the .pak file

[alexh]

Is there any way for this kind of driver for Subway usb?

Unlikely to be usable even if there were. I suspect there is not enough bandwidth on the clockport?