The 10.08Mbps limit could be invalid

[DivXtreme]

When i used bitrate viewer to analyse the original main movie vobs of a DVD, the maximum bitrate was 9626kbps. The movie offers DTS sound and two other soundtracks. I know that DTS can either be 1.536Mbps or 768Kbps and the two other soundtacks were 384Kbps and 192Kbps. If the DTS sound had a bitrate of 768Kbps and not 1.535Mbps then

9626Kbps + 768Kbps + 576Kbps = 10970Kbps , which is over the 10.08 limit.
Can anyone explain this?

Note: 10.08 * 1024 = 10322Kbps . The movie is really close to the limit. Probably 10.08Mbps is not so accurate.

Something else to note. The Max bitrate of Movie+sound is 10970 like i mentioned above. Therefore the limit must be greater than
10970/1024 = 10.71Mbps . This shows that the limit mentioned is not innacurate because of a rounding or trancation error during the convertion into Mbps. Also the innaccuracy is not due to using 1000Kbps/Mbps instead of 1024 because that would just decrease the actual limit.

Does anyone know a more accurate value for the limit or prove that the limit is correct?

Another note. The greatest max bitrate i have ever reached with my re-encoded movies was 11006Kbps and i received an error from scenarist. Thus the limit should be between 10970Kbps and 11006Kbps if my assumptions are valid. If that is the case, why would the scenarist manual state that the limit is 10.08Mbps?

[mpucoder]

You're gonna love this. The maximum combined data rate is 11.08Mbps, but that includes the navigation stream (DSI packets), which does not count in the 10.08 maximum.
If you have DVD Demystified, 2nd ed look at page 250.

[jdobbs]

DivXTreme,

Why are you adding all soundtracks? Unless you can play more than one at a time, I would assume the bitrate only has to include the video track and a single sound track...

jdobbs

After the fact:

Wait -- I get it now. The actual transfer rate from the disc spinning boils down to 11.08Mbs after error correction, demodulation, etc.... so you have to include everything in the stream.

[slk001]

Why are you adding all soundtracks?
All elementry streams coming off the DVD have to be included in the final BR (even if they are not being played). This also includes all subtitle streams (but they are around 10kbps each).

DTS can have bitrates from 64kbps up to 1535kbps.

The 10.08 Mbps IS an absolute maximum. It is impossible to exceed this rate (the data simply isn't there).

[mpucoder]

Should we mention the effect interleaved VOBUs (ILVU) have? The total BR for any one angle is lower and varies with the number of angles, but if you were to sum up all the angles it would higher than 10.08 This can happen because the player skips the other angles before sending them to the track buffer.
The 10.08 limit can be exceeded for short periods, how long depends on the size of the buffer (and I don't think there is any standard on that size)

[DivXtreme]

Originally posted by slk001
DTS can have bitrates from 64kbps up to 1535kbps.Are you sure about that? I havent worked with DTS sound many times but when i wanted to encode DTS sound i could only use 1.536Mbps or 768Kbps (provided the sample rate was 48.0)

Originally posted by slk001
The 10.08 Mbps IS an absolute maximum. It is impossible to exceed this rate (the data simply isn't there).
I accept that 10.08Mbps is the maximum bitrate but this DVD seems to be over that limit. I am sure there is logical explanation behind it.
What mpucoder said sounds interesting. I dont have the book mentioned so i dont know any more details. I will look into the subject more.

[slk001]

You have to remember that multiple angle DVDs simply navigate the DVD differently depending on your desired angle. The player ISN'T reading all the different angles at the same time. The 10.08Mbps simply cannot be exceeded. The DVD is read at a constant rate of 22.16Mbps. Because of the 8/16 encoding, this results in a data rate of 11.08Mbps. There is 1Mbps of error correction overhead, resulting in the MAX data through-put of 10.08Mbps.

My source for the DTS bitrates is DVD Demystified (I don't have any personal experience with DTS encoding).

[mpucoder]

I don't like to get into pissing contests, but I also don't like to see inaccurate info posted here. So, here are the data rates
Raw EFM (16/8) is 26.16 Mbps
after demodulation, it is cut in half to 13.08 Mbps
after error correction the rate is down to 11.08 Mbps
after peeling off the DSI the data goes to the track buffer which can supply data at a rate of 10.08 or less.

And you are right, the rate of data leaving the track buffer for the decoders cannot exceed 10.08 Mbps. But that is spread out over several frames, as the frames are not in temporal order. So the rate can be fudged a little, but not for more than a few frames.

edit: Almost forgot the important point I was trying to make in the beginning. When calculating the bitrate of your creation, you must remove certain overhead that does not go to the track buffer. This includes DSI packets, some header data, and the padding stream.

[slk001]

You are correct on these numbers - my bad - I was recalling this data off the top of my head (must have a bad memory chip up there somewhere!). There is some question on how the modulation format is named. Is it 16/8 or 8/16? It is called BOTH in "the book". But, the jist of it is that it takes 16 recorded bits to achieve 8 bits of data (or 2 in, 1 out).

But all this is really not central to the original question, namely why is the 10.08Mbps seemingly exceeded? My quess is that we need to re-examine our original assumptions. In other words, if the numbers don't add up, then recheck the numbers.