Pro Recorder first Test about bmd-tools HOT 8 CLOSED

On Sat, 29 Mar 2014 16:49:30 -0700
garybruckheimer [email protected] wrote:

i test your bmd-tools with h.264 recorder and when i set HDMI input
to 1080p i got the message:

Nice. According to my knowledge the h264 pro recorder has MUX that is
controlled manually by the driver. That is if there is 'no signal' it
keeps jumping between the input types. Thus I think only the default
input will corrently work.

display mode 0x0f) not supported

That is a 1080p mode. Sounds that HDMI mode is the default. So I think
it should work with modes that I have tuning data for.

So, i think i have to add this display-mode into bmd-streamer.c.

Correct.

Currently i do not understand yet how to add other display-modes. Can
you give me some hints or how i can send you more infos about the pro
recorder device?

The steps I've done are:

1. Enabled the '#if 0' block in bmd_fujitsu_write(). Recompile.
2. Record data using the Windows/MAC official software.
3. Swap USB connection to bmd-streamer; see what are the old register
   values set by the windows software.
4. Construct new display-mode block based on the previous.

I should probably add some LEARNING_MODE switch, and make it print out
the C-code blocks for unknown modes - assuming that it was previously
used from the official driver.

So yes - I have no idea what some of those register are for - I just
copied the init values based on how the windows driver has set them.
Some of the register values seem to be irrelevant (even random) in
certain modes. But some just need to be set.

Atem Studio is working fine, also the firm. extraction works on both.

Nice.

from bmd-tools.

thanks for your fast feedback and help.

i will help to support your app for all BMC devices.

i try to capture the displaymode, i start blackmagic express, capture
some frames over hdmi, exit, kill -9 the bmdserver, start bmd_streamer
and got this bytes:device connected
EP8: 79 bytes: 00 4f 05 0e b2 e2 6f 5d 05 0e b3 2e d9 c5 05 0e b3 7b
44 53 05 0e b3 c7 ae bc 05 0e b4 14 19 49 05 0e b4 60 83 d6 05 0e b4 ac
ee 41 05 0e b4 f9 58 ce 06 01 b5 14 f5 64 05 02 04 01 02 05 90 02 07 04
02 06 09 09 08 b5 14 f6 8e 00 00 00 00
[26/4 1edb:bd43]: FX2Status: Idle (5)
[26/4 1edb:bd43]: MAC address 00:00:00:00:00:00
[26/4 1edb:bd43]: Input Mode, 0x90 (display mode 0x0f) not supported
EP8: 8 bytes: 00 08 05 0e b5 61 59 c1
EP8: 8 bytes: 00 08 05 0e b5 ad c4 5d
EP8: 8 bytes: 00 08 05 0e b5 fa 2e d4

i enable also the if (0 block in:
static void bmd_fujitsu_write(struct blackmagic_device *bmd, uint32_t
reg, uint16_t value)
..
so that this code is activated:
uint16_t oldvalue = bmd_fujitsu_read(bmd, reg);
fprintf(stderr, "%s: fujitsu_write @%06x %04x != %04x\n",
bmd->name, reg, value, oldvalue);

 if (value != oldvalue)
     fprintf(stderr, "%s: fujitsu_write @%06x %04x != %04x\n", 

bmd->name, reg, value, oldvalue);

i set then a breakpoint to the static func. but it seems it will never
call this function.

only this for your update, i will investigate more time to see how we
can implement more display-modes.

when this works, i can use my custrom embedded device to capture the ts
stream and do not longer need my macbook or a windows PC :)

thx.

Am 30.03.14 07:24, schrieb Timo Teräs:

On Sat, 29 Mar 2014 16:49:30 -0700
garybruckheimer [email protected] wrote:

i test your bmd-tools with h.264 recorder and when i set HDMI input
to 1080p i got the message:

Nice. According to my knowledge the h264 pro recorder has MUX that is
controlled manually by the driver. That is if there is 'no signal' it
keeps jumping between the input types. Thus I think only the default
input will corrently work.

display mode 0x0f) not supported

That is a 1080p mode. Sounds that HDMI mode is the default. So I think
it should work with modes that I have tuning data for.

So, i think i have to add this display-mode into bmd-streamer.c.

Correct.

Currently i do not understand yet how to add other display-modes. Can
you give me some hints or how i can send you more infos about the pro
recorder device?

The steps I've done are:

1. Enabled the '#if 0' block in bmd_fujitsu_write(). Recompile.
2. Record data using the Windows/MAC official software.
3. Swap USB connection to bmd-streamer; see what are the old register
   values set by the windows software.
4. Construct new display-mode block based on the previous.

I should probably add some LEARNING_MODE switch, and make it print out
the C-code blocks for unknown modes - assuming that it was previously
used from the official driver.

So yes - I have no idea what some of those register are for - I just
copied the init values based on how the windows driver has set them.
Some of the register values seem to be irrelevant (even random) in
certain modes. But some just need to be set.

Atem Studio is working fine, also the firm. extraction works on both.

Nice.

—
Reply to this email directly or view it on GitHub
#3 (comment).

from bmd-tools.

I'm supposedly getting a pro recorder soon. Will fix this issue after that.

from bmd-tools.

thanks a lot for letting me know.

gary

Am 15.04.14 20:10, schrieb Timo Teräs:

I'm supposedly getting a pro recorder soon. Will fix this issue after
that.

—
Reply to this email directly or view it on GitHub
#3 (comment).

from bmd-tools.

any update on this?

from bmd-tools.

i got a pro recoder yesterday. planning to look at it tomorrow or early next week.

from bmd-tools.

cool, if you need more then one i can prob send you a pro recorder too to test with.

from bmd-tools.

i also can try it out if there is a new version...thanks for spending
more time on this.

Am 01.05.14 20:28, schrieb coral:

cool, if you need more then one i can prob send you a pro recorder too
to test with.

—
Reply to this email directly or view it on GitHub
#3 (comment).

from bmd-tools.