17 Sep
2013
17 Sep
'13
4:37 a.m.
Hi, I'd like to compare recording quality of portable recorders.
I only care about the line-in input of these portable recorders.
I'm thinking:
* play audio through a good DAC and feed the line-in of each recorder
* copy audio files off the SD cards from the recorders
* level match the inputs. Probably based on "sox FILE -n stats"
output change input levels on each recorder to match levels.
* play audio again through the DAC and feed the line-ins
* copy audio files off the SD cards from the recorders again
* run something like diff(1) for audio
I'm willing to look at a spectrogram/graph image with feh(1), but I
_really_ don't want to touch a mouse or tab through form
buttons/boxes/inputs on a GUI.
Thanks for reading.
17 Sep
17 Sep
9:04 a.m.
A recorder that might do the better mic recording, might be less good in
copying an analog line in signal. I would record the same audio source
in the same location, from the same position, by different recorders
using the microphones and then compare the signal from the read cards or
at least copy it back using a digital output, resp. if you will using it
to record line in, at least don't copy it back by using the analog way.
I've got doubts that all recorders provide the same audio quality for
mic in and line in.
18 Sep
18 Sep
12:27 a.m.
Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote:
On Tue, 2013-09-17 at 02:37 +0000, Eric Wong wrote:
I use external mics + battery box with the portable recorders for
recording live music.
Occasionally, I may also feed the tape-out of my stereo receiver to
digitize vinyl since it's the only usable line-in+ADC I have.
I only care about the line-in input
My apologies :D, didn't read carefully enough. However, why do you want
to use analog line out?
1:19 a.m.
On Tue, 2013-09-17 at 22:26 +0000, Eric Wong wrote:
Ralf Mardorf <ralf.mardorf(a)alice-dsl.net>
wrote:
Sorry again, I didn't read carefully. I meant the recorder's line out,
but you didn't wrote about using it. Later I noticed that you copied
from the SD card and then feed the line ins again. Instead of DA I read
AD and I missed to read some sentences. :D
On Tue, 2013-09-17 at 02:37 +0000, Eric Wong
wrote:
I use external mics + battery box with the portable recorders for
recording live music.
Occasionally, I may also feed the tape-out of my stereo receiver to
digitize vinyl since it's the only usable line-in+ADC I have. I only care about the line-in input
My apologies :D, didn't read carefully enough. However, why do you want
to use analog line out?
17 Sep
17 Sep
9:47 a.m.
On Tue, 2013-09-17 at 02:37 +0000, Eric Wong wrote:
Hi, I'd like to compare recording quality of
portable recorders.
I only care about the line-in input of these portable recorders.
I'm thinking:
* play audio through a good DAC and feed the line-in of each recorder
* copy audio files off the SD cards from the recorders
Sorry, I guess I don't understand ...
* level match the inputs. Probably based on "sox
FILE -n stats"
output change input levels on each recorder to match levels.
* play audio again through the DAC and feed the line-ins
... why you want to do this?
Forget what I've written. You seemingly want to test a copy of a copy,
while doing some processing. IMO that won't give a valid result.
12:46 p.m.
On 09/17/2013 04:37 AM, Eric Wong wrote:
Hi, I'd like to compare recording quality of
portable recorders.
I only care about the line-in input of these portable recorders.
if you have a "known good" signal source (that surpasses the quality of
the unit under test in all aspects you care about), you can measure the
line-in of your recorder from that.
your best option would be to measure the impulse response of the
recorder by feeding it a sine sweep from, say, 20 hz to 22 khz (or more,
if you care about extended frequency response and the device is running
at above 48khz sampling rate), recording that, and generating an impulse
response from it. fons adriaensen's "aliki" tool is useful for that, if
somewhat tricky to use.
a look at the impulse reponse will reveal distortion, and a look at the
raw sine sweep will give you a pretty good idea of the amplitude
response over frequency.
then there's "jnoisemeter", also by fons, which can give you a pretty
good measurement of the noise floor.
but all these measurements either rely on reference gear to produce the
test signals, or _very_ careful thinking if the test signals might are
already laden with their own problems.
there's an old saying in german which unfortunately doesn't translate
well: "wer viel misst, misst mist.", literally, "s/he who measures a
lot, will measure dung" (as in, will obtain false or misleading results).
there is no such thing as an audio diff, nor would it be meaningful. the
only practical application of something resembling an audio diff is a
dual fft that constantly compares the input of a P.A. system with its
output as picked up by a measurement microphone in the room. you could
do the same with your recorder, but it's compromised, because the
program material will not be a good source signal for all aspects of
audio quality. you can obtain better results with special test signals
such as pink noise (which is obviously not possible during a show,
except maybe for some fringe genres, hence the dual fft :)
--
Jörn Nettingsmeier
Lortzingstr. 11, 45128 Essen, Tel. +49 177 7937487
Meister für Veranstaltungstechnik (Bühne/Studio)
Tonmeister VDT
http://stackingdwarves.net
12:56 p.m.
On Tue, 2013-09-17 at 12:46 +0200, Jörn Nettingsmeier wrote:
but all these measurements either rely on reference
gear to produce
the test signals, or _very_ careful thinking if the test signals might
are already laden with their own problems.
Good hints :)! Your hints anyway will provide a more valid result, than
checking the copy of a copy.
2:08 p.m.
Am Dienstag, 17. September 2013, 12:46:06 schrieb Jörn Nettingsmeier:
there's an old saying in german which
unfortunately doesn't translate
well: "wer viel misst, misst mist.", literally, "s/he who measures a
lot, will measure dung" (as in, will obtain false or misleading results).
"who measures much, measures mess", may sort of transport the idea, if it's
not obvious from reading the original already... :D
Edgar
3:43 p.m.
On Tue, Sep 17, 2013 at 12:46:06PM +0200, Jörn Nettingsmeier wrote:
On 09/17/2013 04:37 AM, Eric Wong wrote:
The primary tools to measure an A/D or D/A (that's what it amounts
to, the rest just be pretty transparent in any recorder) are jaaa
and jnoisemeter. Both were explicitly designed for the job.
Hi, I'd like to compare recording quality of
portable recorders.
I only care about the line-in input of these portable recorders.
a look at the impulse reponse will reveal distortion, and a look at
the raw sine sweep will give you a pretty good idea of the amplitude
response over frequency.
then there's "jnoisemeter", also by fons, which can give you a
pretty good measurement of the noise floor. but all these measurements either rely on reference
gear to produce
the test signals, or _very_ careful thinking if the test signals
might are already laden with their own problems.
I you have a good quality audio card it is perfectly possible to
use this as a signal source. First measure it separately, so you
have an idea of the limits of what you can do using it. For some
measurements you also need to calibrate it using a true RMS audio
voltmeter which you can beg, steal or borrow.
You may need some simple HW: passive attenuators (three resistors),
and plugs or test cables to short-circuit inputs. For the rest you
need to perfectly understand *what* you try to measure, and have
a very systematic approach to doing it.
there's an old saying in german which
unfortunately doesn't
translate well: "wer viel misst, misst mist.", literally, "s/he who
measures a lot, will measure dung" (as in, will obtain false or
misleading results).
99% percent of false results are due tp lack of understanding, and
failing to verify basic things such as levels.
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)
4138
days inactive
4138
days old
linux-audio-user@lists.linuxaudio.org
9 comments
5 participants
participants (5)
-
Edgar Aichinger -
Eric Wong -
Fons Adriaensen -
Jörn Nettingsmeier -
Ralf Mardorf