Converting DASHhigh to FLAC with ffmpeg

[Jeremy Nicoll - ml gip]

On 2017-01-24 20:19, Tony Quinn wrote:
> On 24-Jan-17 8:16 PM, Tony Quinn wrote:
> 
>> On 24-Jan-17 8:11 PM, Jeremy Nicoll - ml gip wrote:

>>> I have some original recordings made in the last year or two at 
>>> choral concerts - mostly
>>> 24 bit, 44.1 khz, stereo uncompressed WAV files, made with ecent 
>>> quality condenser mics.
>>> I was surprised how hard it was to tell the originals apart from eg 
>>> 320kbps MP3 let alone
>>> eg 160kbps.  And I was listening extremely carefully, criticising my 
>>> own recordings.
>> 
>> I assume that you weren't doing a blind test, so one can expect some 
>> sort of conformation bias to play a part here too.
> Obviously that should have read "....confirmation bias"

Bias could have worked in either direction though.

I /expected/ to be able to discriminate easily, and couldn't.  I didn't 
listen to the whole
concert at each quality level, just snippets of audio which I thought 
might be more or less
blurred depending on quality - rapid note sequences, very quiet passages 
(but then audience
and venue noise was a problem there), very loud passages and so on.  I 
was listening for
detail I knew to be present on the original recording trying to assess 
whether it wasn't as
precise in the MP3 files.  It wasn't scientific and I know I could (and 
probably will) do
a better job on that sometime, but nevertheless I didn't get the 
impression that 360kbps MP3
was all that bad.

So am I keeping the 24 bit originals?  Yes.  Maybe younger ears can 
still hear the difference.

-- 
Jeremy Nicoll - my opinions are my own