Skip to contentSkip to conversation

TT-USB2.0 and high bitrate packet loss (DVB-C/T)

DISCUSSION Hans Petter Selasky 2 participants 4 messages May 30, 2013 – Jun 5, 2013 Media (V4L/DVB)

lore
Share & tools
lore
#0 lore raw

Hi there,

I need to get in concat with someone that can handle, test and review a patch for TT-USB2.0. I've found that for certain high-bitrate streams, the TT-USB2.0 sends more ISOCHRONOUS MPEG data than is specified in the wMaxPacketSize fields. I have a USB analyzer capture which shows this clearly. This of course won't be received at the USB host and packet drops appear inside the stream. The solution is to use another alternate setting. The technotrend chip has many of these. I've now tested using alternate setting 7 instead of 3.

Alternate setting 7 allows transferring a maximum of 3 * 1024 bytes. Alternate setting 3 allows transferring a maximum of 1 * 940 bytes.

--HPS

#1 lore raw
Quoted 11 lines · Hans Petter Selasky

Hi there,

I need to get in concat with someone that can handle, test and review a patch for TT-USB2.0. I've found that for certain high-bitrate streams, the TT-USB2.0 sends more ISOCHRONOUS MPEG data than is specified in the wMaxPacketSize fields. I have a USB analyzer capture which shows this clearly. This of course won't be received at the USB host and packet drops appear inside the stream. The solution is to use another alternate setting. The technotrend chip has many of these. I've now tested using alternate setting 7 instead of 3.

Quoted 4 lines

Alternate setting 7 allows transferring a maximum of 3 * 1024 bytes. Alternate setting 3 allows transferring a maximum of 1 * 940 bytes.

--HPS

Hi,

It turns out that this device, at least the version I bought, does not work with the XHCI at all when using multi-packet transfers, alternate setting 7, because DATA2 PID is used when transfer is less than 1024 bytes. It should be DATA0 PID first, then 1 and 2 in the end. Probably a firmware update can fix this, but I'm not aware about if such exists. The PID issue was found using a USB analyzer.

Apparently this bug has gone undetected, because at least the EHCI high speed only controller I've got silently ignores this kind of errors and receives the data whilst the XHCI not. Conclusion:
Existing alternate setting must be used.

I think I will have to get another USB based receiver with CI slot. Any recommendations for DVB-T ?

--HPS

#2 lore raw
Quoted 26 lines · Hans Petter Selasky
Quoted 11 lines · Hans Petter Selasky

Hi there,

I need to get in concat with someone that can handle, test and review a patch for TT-USB2.0. I've found that for certain high-bitrate streams, the TT-USB2.0 sends more ISOCHRONOUS MPEG data than is specified in the wMaxPacketSize fields. I have a USB analyzer capture which shows this clearly. This of course won't be received at the USB host and packet drops appear inside the stream. The solution is to use another alternate setting. The technotrend chip has many of these. I've now tested using alternate setting 7 instead of 3.

Quoted 4 lines

Alternate setting 7 allows transferring a maximum of 3 * 1024 bytes. Alternate setting 3 allows transferring a maximum of 1 * 940 bytes.

--HPS

Hi,

It turns out that this device, at least the version I bought, does not work with the XHCI at all when using multi-packet transfers, alternate setting 7, because DATA2 PID is used when transfer is less than 1024 bytes. It should be DATA0 PID first, then 1 and 2 in the end. Probably a firmware update can fix this, but I'm not aware about if such exists. The PID issue was found using a USB analyzer.

Quoted 7 lines

Apparently this bug has gone undetected, because at least the EHCI high speed only controller I've got silently ignores this kind of errors and receives the data whilst the XHCI not. Conclusion:
Existing alternate setting must be used.

I think I will have to get another USB based receiver with CI slot. Any recommendations for DVB-T ?

There is no many alternatives available. I suspect Anysee E7 serie is the only one. And I am not even sure if its CI works anymore. Lastly when I tested it I didn't get scrambled channels working - but it could be due to card entitlements were not upgraded. Anyhow, if there is bug then it should be easy to fix.

regards
Antti

Signature · 1 line
http://palosaari.fi/
#3 lore raw
Quoted 10 lines · Antti Palosaari
Quoted 2 lines · Hans Petter Selasky

I think I will have to get another USB based receiver with CI slot. Any recommendations for DVB-T ?

There is no many alternatives available. I suspect Anysee E7 serie is the only one. And I am not even sure if its CI works anymore. Lastly when I tested it I didn't get scrambled channels working - but it could be due to card entitlements were not upgraded. Anyhow, if there is bug then it should be easy to fix.

I just tested, actually first bisecting back to Kernel 3.3 and VLC2 and it didn't work... That makes me suspect it was VLC which has gone broken. So I tried gnutv & mplayer and surprise CI was working! Both Anysee E7 TC and Anysee E7 T2C. Tested only DVB-C as I don't have DVB-T subscription card. CI/CAM worked earlier for VLC somehow too.

Start tuning on terminal:
gnutv -channels /path/to/channels.conf "scrambled channel" Start mplayer on the another terminal:
mplayer /dev/dvb/adapter0/dvr0

It is a little bit boring situation as there is no very good Gnome Desktop TV application. Have never been...

regards
Antti

Signature · 1 line
http://palosaari.fi/