[linux-dvb] Question about Frontend performance (TDA10046 + TD1316)

Chun Chung Lo cclo at astri.org
Thu Aug 17 02:55:43 CEST 2006


Thanks for your quick reply.

As the STB is operating under linux kernel 2.4 (powerpc), so my driver
is composed by several versions of DVB-T codes:

For SAA7130/4 driver, I used the source codes from
http://www.linuxtv.org/download/kfir/mpex.tar.gz

For the driver of TDA10046 + TD1316, I used the codes from linux kernel
2.6.16.18.

I know the SAA7130/4 driver is old, but the main problem is I do not
know how to back port the video-buf-dvb.c (SAA7130/4 DMA buffer handling
functions)from linux kernel 2.6 to linux kernel 2.4 (mainly due to the
"kthread_run()", "kthread_create()" stuff) as I am a newbie on driver
porting.

As the TS capture of the SAA7130/4 driver is not done by kernel thread,
it consumes lots of resources and the performance is not so good during
I capture TS.

Are there any suggestion to me?

Thanks for your kindly help.

Best regards,
Lo Chun Chung

-----Original Message-----
From: Hartmut Hackmann [mailto:hartmut.hackmann at t-online.de]
Sent: Thursday, 17 August, 2006 5:50 AM
To: Chun Chung Lo
Cc: linux-dvb at linuxtv.org
Subject: Re: [linux-dvb] Question about Frontend performance (TDA10046 +
TD1316)


Hi,

Chun Chung Lo wrote:
> Hi all,
>
> I am now working on a DVB-T STB projects. SAA7130HL + TDA10046A +
> TD1316 is used.
>
> Previously I have made some tests on our STB. The TDA10046 + TD1316
> status are monitored during all the tests performed, by using the
> 'tzap' utilities (bundled in linuxtv-dvb-app-1.1.1)
>
> Different degrees of corrupted TS is captured (a TS-to-PES software is

> used to runtime convert the catpured TS to a normal mpeg2 movie).
>
> There are different fields can be monitored by using 'tzap', such as
> Signal Strength, SNR, BER, UNC, etc ... And I would like to know the
> following things:
>
> 1. All tests has a zero UNC, but a little BER is recorded (about 120
> out of 128K). Will this is the main reason of the corrupted TS?
>
I can't tell right now how these values are scaled. UNC stands for
uncorrected blocks while BER stands for the raw bit error rate.

> 2. When BER > 0, is this means there exists some corrupted TS packets?

> If no, what does BER > 0 means (under this circumstance)?
>
It depends on the DVB configuration, but to my experience in Germany,
you are fine as long as BER is less than ~4500 (and UNC is 0)

> 3. When UNC or BER (or both) > 0, will this means the discontinuity of

> the TS stream?

yes.
>
> 4. Or the corruption of the captured TS is due to insufficient
> processing speed of the driver? (DMA? Buffers processing?)
>
There was a bug in the transport stream capture (saa7134-ts.c) i fixed
11 months ago. If your driver is older, you need to update it. Processor
load is low if you just capture the stream. But a problem might be that
IDE DMA is not working properly on your system. This might cause high
latencies in the kernel and thus data loss.

> 5. Will the TS stream BW (or other settings) affect the capture
> performance? (I do not know but I have tried Test A and B several
> times and the result are quite coherent, as stated above).
>
If you use wrong parameters for the channel decoder, it will not lock in
most cases. But what about the code for the TD1316? Where does this come
from? do you have a TD3116A or do you use some seperate IF amplifier?

> Please help and thank you for your attention.
>
> Best regards,
> Lo Chun Chung
>
Hope this helps

Hartmut

This message (including any attachments) is for the named addressee(s)'s use only. It may contain
sensitive, confidential, private proprietary or legally privileged information intended for a
specific individual and purpose, and is protected by law. If you are not the intended recipient,
please immediately delete it and all copies of it from your system, destroy any hard copies of it
and notify the sender. Any use, disclosure, copying, or distribution of this message and/or any
attachments is strictly prohibited.





More information about the linux-dvb mailing list