29 Jun
2013
29 Jun
'13
5:22 p.m.
Ive been trying to capture some screencasts of talks mine with avconv.
many sites seem to recommend using
-vpre lossless_ultrafast
From the pov of the machine does this consume more CPU
or less?
If I set sample-rate to 11000 will ordinary speech be spoilt? What about
5500?
Basically I have an i5 laptop but it starts making a racket if its driven
too much.
So trying to run it cool
29 Jun
29 Jun
7:39 p.m.
On 2013-06-29 13:22, Rustom Mody wrote:
If I set sample-rate to 11000 will ordinary speech be
spoilt? What about 5500?
11000 ought to be fine, as most speech-specific applications I've seen
default to 8000.
5500? Not sure. If your voice is low, it should still be fine, maybe a bit
muffled. But in the late '90s, when I was at a company looking at an early
VOIP PBX, we turned them down when the CEO and his wife tried to have a
conversation on their demo equipment and it bandpassed her very high,
squeaky voice right out of existence. So it's definitely possible to go too
low, though I can't imagine her voice was entirely over 2750Hz.
Should be pretty easy to test before doing any real recordings.
Rob
30 Jun
30 Jun
4:44 a.m.
On Sun, Jun 30, 2013 at 1:09 AM, Rob <lau(a)kudla.org> wrote:
On 2013-06-29 13:22, Rustom Mody wrote:
Thanks this is the kind of data I was looking for!
I just assumed that if the number used is not a multiple of what would be a
later standarized rate like 44100, then some kind of 'jitter' (dunno the
correct technical word) would appear on resampling.
However since 48000 is a standard, 8000 for capture should be fine I guess??
If I set sample-rate to 11000 will ordinary
speech be spoilt? What about
5500?
11000 ought to be fine, as most speech-specific applications I've seen
default to 8000.
12:21 p.m.
On Sun, Jun 30, 2013 at 10:14:30AM +0530, Rustom Mody wrote:
I just assumed that if the number used is not a
multiple of what would be a
later standarized rate like 44100, then some kind of 'jitter' (dunno the
correct technical word) would appear on resampling.
No. Resampling doesn't in any way depend on having an
integer ratio.
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
3:54 p.m.
On Sun, Jun 30, 2013 at 5:51 PM, Fons Adriaensen <fons(a)linuxaudio.org>wrote;wrote:
On Sun, Jun 30, 2013 at 10:14:30AM +0530, Rustom Mody
wrote:
[OT] Trying to figure out what happens when I set the sample rate very low.
Put it to 500 and avconv crashed!
I just assumed that if the number used is not a
multiple of what would
be a
later standarized rate like 44100, then some kind
of 'jitter' (dunno the
correct technical word) would appear on resampling.
No. Resampling doesn't in any way depend on having an
integer ratio.
Ciao,
6:37 p.m.
On 29/06/13 19:22, Rustom Mody wrote:
Ive been trying to capture some screencasts of talks
mine with avconv.
many sites seem to recommend using
-vpre lossless_ultrafast
From the pov of the machine does this consume more CPU or less?
If I set sample-rate to 11000 will ordinary speech be spoilt? What about
5500?
I don't think the audio sample rate will have a great impact on CPU
consumption, actually using your sound card's 'native' sample rate might
save a bit of cpu (if it's a laptop I guess it might be 48k?)
I think what's going to impact most on the cpu is the video framerate...
Many tutorials suggest using 30 fps, but for a screencast 20 or even 15
might be enough.
My two cents.
Lorenzo.
Basically I have an i5 laptop but it starts making a racket if its
driven too much.
So trying to run it cool
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
3984
days inactive
3985
days old
linux-audio-user@lists.linuxaudio.org
5 comments
4 participants
participants (4)
-
Fons Adriaensen -
Lorenzo Sutton -
Rob -
Rustom Mody