SCR Calculation 2 (SCR=1.001*GOP)

[MarcioAB]

Hello, I'm looking at DVD data (MPEG header) generated by my DVD Desktop Record to understand how SCR is calculated.

I found here a "SCR Calculation" thread from jdobbs, back on 2003. It seems SCR calculation should be something complex.

But instead (I think) I see a very simple way to calculate it: SCR=GOP*1.001 ( NTSC ).

We know the simple way is generally the wrong way, but please let me know where I'm wrong.

Thank you
Marcio

Further explanation: GOP timestamps are easy to calculate: frames displayed divided by 30, for DVDs running @ 30 frames per second (they are stored on each 0xB8 MPEG GoP streams).

[mpucoder]

If you are talking about the SCR value for the NavPk only, it is usually very simple - NTSC is 3003 * frames, PAL is 3600 * frames.
But this is not true under many circumstances. Also the following packs SCR calculation is very complex, it depends on the mux_rate and all 41 stream buffer models.

[MarcioAB]

Thank you mpucoder.

I was talking about the first VideoPk after the NavPk (the one with the 0x01B8 start code stream), but looking again in my data, I see now, even for the NavPk, there are some 90KHz ticks more on SCR than just the 3003*frame.

For NavPk I found approx 580 ticks (90KHz) more and for the next VideoPk (0xB8) I found approx 780 ticks more:

SCR (NavPk) = 3003*frame + 580
SCR (0xB8) = 3003*frame + 780

So, if I need to redo the SCRs, can I use these simple values, or do I need an exact calculation ?

Thank you
Marcio

[mpucoder]

It's much too complicated, and somewhat of a "trade secret". Every stream of data has to meet very strict P_STD buffer constraints.