On Sun, Nov 23, 2014 at 11:20 PM, Timothy Gu <[email protected]> wrote:
> On Sun, Nov 23, 2014 at 2:59 PM, Vittorio Giovara
> <[email protected]> wrote:
>> ---
>> Using the test the author provided. They are going to be tested on oracle too
>> before pushing.
>>
>> Vittorio
>>
>>  tests/fate/audio.mak  |   4 +
>>  tests/ref/fate/dss-lp | 737 
>> ++++++++++++++++++++++++++++++++++++++++++++++++++
>>  tests/ref/fate/dss-sp | 334 +++++++++++++++++++++++
>>  3 files changed, 1075 insertions(+)
>>  create mode 100644 tests/ref/fate/dss-lp
>>  create mode 100644 tests/ref/fate/dss-sp
>
> First, the samples are kinda big. The ref for dss-lp is bigger than
> all but two tests in fate, one of them doesn't use a sample
> (iirfilter) and the other one is an official sample.

yeah they could be limited, but i sent them full up for discussion on
purpose, so that one could spot something like below. Any good
significant value you might suggest?

> Second, do you have any idea why the end of dss-lp test ref looks like this?
>
>> +0,     143040,     143040,      240,      480, 0x00000000
>> +0,     143280,     143280,      240,      480, 0x00000000
[dutiful snip]
>> +0,     176160,     176160,      240,      480, 0x00000000
>> +0,     176400,     176400,      240,      480, 0x00000000
>
> Is this a bug in the decoder or is it possible to just cut it out?

I'll leave this up to Oleksij :)

-- 
Vittorio
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to