Want to resurrect your D/D2 MAC kit?

[orizatriznyak]

When you say you see cut-points, is it the small black dots running around?

Yes, exactly. And also the picture is grainy if the scrambling is on. Is this problem also there in videocrypt mode?

How the decoder knows the cutpoints? Is there a connection between ecm data and the cutpoints? Or the free mode ecm just instructs the decoder to guess and descramble?

 

[homercartman]

@orizatriznyak
Videocrypt runs fine without any visible cutpoints on a SVA1 decoder.

 

[Captain Jack]

Yes, exactly. And also the picture is grainy if the scrambling is on. Is this problem also there in videocrypt mode?

How the decoder knows the cutpoints? Is there a connection between ecm data and the cutpoints? Or the free mode ecm just instructs the decoder to guess and descramble?

The cut point calculation is very well documented by official D/D2MAC specification documents, so it's generated on the fly by HackTV using the default seed in free access mode (60-bit - all 1s).

 

[fsphil]

The dots are why I haven't exposed the scramble option yet. Some of my decoders show them, some don't. I don't believe I have the timing correct, and the edges should be smoothed/blended which I'm not doing yet.

 

[orizatriznyak]

Thanks for the info Phil.
I can see there is also a mac teletext related part in the source.

I know it is not the place for this, but I have some ideas what improvements would be great to this project:
-Wegener Panda (stereo) preemphasis support
-Alternative audio source file support(not just for mac) to simulate a radio station as a secondary audio track
-Support for A2(zweikanalton) stereo
-J17 audio preemphasis support

 

[Captain Jack]

I think Phil looked into Wegener before but couldn't find detailed enough spec to implement it. Secondary audio tracks might be possible but the main issue currently is they ffmpeg does not support decoding more than a single video/audio stream (unless there were several instances of them). This would be especially useful in MAC where you can shove 4 audio tracks in there.

 

[fsphil]

J17 pre-emphasis is already there for NICAM (and MAC).
I don't think any of my devices support A2 stereo. It looks quite simple though, it might be simple to add.

 

[fsphil]

Also missing is the FM audio 50us pre-emphasis filter. It's a bit annoying as the high frequencies are boosted and often clip - I need to play about with the levels a bit (reduce the audio level, and compensate with a higher FM deviation).

 

[Channel Hopper]

Something in the back of my mind thinks that one of the cards in the Comppanion III decoder box here is a Wegener stereo board

 

[orizatriznyak]

J17 pre-emphasis is already there for NICAM (and MAC)..

I meant J17 for an analog subcarrier. This was used on the French Telecom satellites mostly mono at 5.8MHz with secam.
So also a secam-fm and secam d,k option would be great. This secam d,k was used here in Hungary for many years as well (SECAM colour, 25 fps, 625 lines, AM (complex), 6.5 MHz FM.) Same with ntsc-fm but this was rarely used in Europe as I remember.

 

[Captain Jack]

Russia also used SECAM D,K back in the day. I think some minor transmitters in Russia and Ukraine are still broadcasting in this.

 

[fsphil]

I've pushed SECAM D, K and FM modes, as well as NTSC FM. None have been tested, and SECAM support is still very poor :-)

Now that you mention it, I think my Pace decoder has J17 as an audio option. It's not connected at the moment but I'll have a look tonight.

 

[steeviebops]

Yes, exactly. And also the picture is grainy if the scrambling is on. Is this problem also there in videocrypt mode?

I have two VideoCrypt decoders - the original SVA1 and an Amstrad SRD2000 IRD. The SVA1 is perfect and doesn't show the cutpoints. The Amstrad does show them.

 

[fsphil]

The Amstrad does show them

Is this with PAL FM? If so this might be due to the lack of a proper pre-emphasis video filter in hacktv.

 

[steeviebops]

Is this with PAL FM? If so this might be due to the lack of a proper pre-emphasis video filter in hacktv.

Yes, PAL-FM alright. I did use the --filter option but got a warning that it's only designed to run at 16 MHz. I was running it at 14 MHz for the encryption to work. So it's probably this mismatch that causes it.

 

[fsphil]

There will also be some distortion due to the limited bandwidth at 14MHz, FM video needs a fair bit of space. The next ideal sample rate for VC is 28Mhz but that's too much for the hackrf.

So two things on the TODO list are:
* Sub-sample shift so cut-and-rotate can be done at any sample rate (may be slow)
* Function to generate CCIR-405 pre-emphasis filter taps at any sample rate

 

[Captain Jack]

Which Amstrad is it? On SDR510, you can potentially mitigate this:

(Q.30) What are these decoder spots?

>Further to my e-mail yesterday I have now purchased the Amstrad SRD510.
>Generally, all is well. However, on a couple of channels 'UK Horizons, UK
>Gold', the picture is literally swarming with minute black and white dots.
>These are *not* sparklies but something quite different.
>I've had a good look through your FAQs but I can't find any relevant info on
>this problem. Would these be so-called 'decoder spots'? If so, is there
>anything I can do to reduce them?
>Thanks in anticipation,

It sounds like decoder spots. This is to do with the "seed code" that the decoder uses to determine the precise cut and rotate point for each line. If it's a couple of pixels out, you'll see these "buzzing bees". The c & r seed value is determined by four jumper links inside the decoder box near the front edge.

From SatCure satellite Amstrad SRD510 receiver FAQ

 

[steeviebops]

I'm conscious that I'm going off-topic here but have a few points to address.

I've pushed SECAM D, K and FM modes, as well as NTSC FM. None have been tested, and SECAM support is still very poor :-)

Thanks, I played around with it this morning. I've just tested NTSC-FM and I think I found a bug. The code is using a 625-line raster and it doesn't work properly as a result - no colour and looks somewhat garbled. I manually edited it to use the 525-line parameters and it looks a lot better but the colour is slightly off. Then again I just copied and pasted the values from the AM section so it's probably wrong.

Which Amstrad is it? On SDR510, you can potentially mitigate this:

SRD2000. It looks like new, I'm guessing it was never used before now. Was still in its box with instruction book, batteries and cables!

 

[fsphil]

The code is using a 625-line raster and it doesn't work properly as a result

Whoops. Indeed, fixed now. This will teach me not to push code out without testing it :-)

There is a bug in how the colour carrier is generated, at certain sample rates the NTSC colour is wrong.

 

[Captain Jack]

SRD2000. It looks like new, I'm guessing it was never used before now. Was still in its box with instruction book, batteries and cables!

Ah that was on eBay doing rounds recently. I also have one in the loft but way too much for our purpose with its surround sound malarkey built in.