The block parameter is the number of 188-byte Transport Stream packets
that the block will process at a time. It's set to 8 to match the next
block upstream/downstream which is the Convolutional
Interleaver/Deinterleaver.
Because the block is processing the input/output in specific chunks, the
input/output pins are vectors. For the encoder, it's 188 * 8 or 1504 on
the input and 204 * 8 or 1632 on the output. If you want to connect a
byte block to one of these inputs/outputs, you can use a stream to
vector or vector to stream block in between to do the conversion. Blocks
like the File Source have a programmable vector length.
When you connect two blocks with the wrong vector size, the "View flow
graph errors" button will light up. Click on the button to see what the
mismatch is. For example, of you connect the File Source block to the
Reed-Solomon encoder, the error message will be "Source IO size "1" does
not match sink IO size "1504"". It's telling you you're trying to
connect a byte output to a vector of size 1504 input. If you set the
"Vec Length" on the File Source block to 1504, the error will go away.
Ron
On 7/14/20 16:13, George Edwards wrote:
> Hello,
>
> I am trying to validate proper encoder/decoder operation for the DvBT
> Reed Solomon FEC blocks. I have run into two problems:
>
> 1. There is a parameter "Block" that needs to be set in both the
> encoder/decoder, and the Documentation states that it is the number of
> blocks to process at once. Please explain what this means.
>
> 2. I thought the data types for the encoder and decoder input/outputs
> were type byte, but when I connected a byte stream to the encoder
> input it gave an error and on closer examination, the data type color
> is a little deeper than that used for byte. What data type does the
> encoder takes?
>
> I will appreciate any help you can provide.
>
> Regards,
> George
No comments:
Post a Comment