Re: [Lame-dev] VBR tag

[Rob L. <ro...@ma...>]

On Friday, July 25, 2003, at 12:44  AM, Gabriel Bouvigne wrote:
>> I think it is wrong to keep the Xing header at the same offset 
>> regardless of CRC protection. The offset is already dependent upon 
>> whether the stream has one or two channels; why violate the rule that 
>> the header can always be found in the ancillary data of the frame?
>
> Is it forbidden to switch crc protection on/off inside a bistream? If 
> not, perhaps we could just never use crc protection for the Xing 
> header?

It's not forbidden to switch CRC protection within the stream, but I 
think the motivation to use it in the Xing header frame was for the 
benefit of programs that examine the first frame to detect stream 
parameters. So you might lose the ability of some programs to correctly 
identify whether the balance of the stream has CRC protection.

-- 
Rob Leslie
ro...@ma...