Thread View: comp.multimedia
2 messages
2 total messages
Started by Andrzej Duda
Thu, 12 Nov 1992 15:05
frame grabber for Sparc
Author: Andrzej Duda
Date: Thu, 12 Nov 1992 15:05
Date: Thu, 12 Nov 1992 15:05
12 lines
855 bytes
855 bytes
We are looking for a frame grabber card for live TV input. We want to digitize it in real-time without compression with the resolution from 320x240x16 to 160x120x16 at 10fps and store to disk. The card should be for Sparc (or for PC eventually). According to my information neither VideoPix, RasterOps nor Parallax can do it. Any suggestions? Also, we want to decode close-captionning from the input signal. Do you know about any decoder to do it? Thanks for your help, Andrzej Duda
Re: frame grabber for Sparc
Author: Edgar Ostrowski
Date: Mon, 16 Nov 1992 19:13
Date: Mon, 16 Nov 1992 19:13
39 lines
2519 bytes
2519 bytes
In article <1992Nov12.150556.25771@mintaka.lcs.mit.edu> duda@brokaw.lcs.mit.edu (Andrzej Duda) writes: >We are looking for a frame grabber card for live TV input. We want to >digitize it in real-time without compression with the resolution from >320x240x16 to 160x120x16 at 10fps and store to disk. The card should >be for Sparc (or for PC eventually). According to my information neither >VideoPix, RasterOps nor Parallax can do it. Any suggestions? > For my experience both the Parallax XVideo and the RasterOps SparcCard TC can be used for your tasks. I've run a small program on our Sparc2 equipped with a Parallax XVideo board to see what's possible (to get good performance values the framebuffer was mapped into memory which allows direct read access). An image of 240x180 pixels (24 bit/pixel) can be read from the card, converted to 16 bit (5/5/5 RGB) and written to disk in 65-71 msec. This was measured with the tmp filesystem in the swap partition. With the normal filesystem one can't get similiar values, there's too much overhead from the flesystem (from six successive frames three were grabbed, converted, and written in 67-70 msec each; one needed 150-170 msec and the other two needed 200-240 msec). This results in a mean value of about 130 msec /frame. I'd call this worst case. We experienced similiar access rates for reading from the framebuffer of the RasterOps card. Therefore I think that the performance of both cards will not be significantly different. The greatest performance bottleneck for storing uncompressed video is be the filesystem - if you use it; you might try using a raw partition instead. The UNIX filesystem probably has it's advantages, but there is anything else but support for continuous media. -edgar- ________________________________________________________________ Edgar Ostrowski phone: +49 30 314-25985 e-mail: ostrowski@prz.tu-berlin.dbp.de X.400: s=ostrowski; ou=tubkom; ou=prz; p=tu-berlin; a=dbp; c=de ________________________________________________________________
Thread Navigation
This is a paginated view of messages in the thread with full content displayed inline.
Messages are displayed in chronological order, with the original post highlighted in green.
Use pagination controls to navigate through all messages in large threads.
Back to All Threads