Subject: Re: Re[2]: [linux-audio-dev] peakfiles and EDL's
From: Richard C. Burnett (burnett_AT_tality.com)
Date: Tue Feb 27 2001 - 20:32:31 EET
Hehe, I didn't mean that what you were saying was not true for the context
you were describing, only that the peakfile method is a proven method for
displaying sound data. It also does not matter what you hear if you are
overcoming a limitation in hardware. If one signal is at 90% max and the
other 90% and you are mixing these with a gain of 1, there is a real good
chance for clipping. These peak files are in no way for precise editing,
just a visual aid. I mean, I use it also to make sure I am getting the
most of my range of audio with my recording equipment. I record as loud
as I can without clipping so that I have to apply negative gain to the
signals. This allows me to push noise further down if I am getting any,
and using these peak views allows me to see how close any sample data is
from clipping.
I think one issue is that you are looking at this from a different
viewpoint altogether. Phase/Frequency aren't key issues all the
time. Downsampling is going to remove frequency data, so thats no good to
display, which is why the peak file display method is very useful.
Now, as a new note, the new tool from Samplitude has a new display
method. They call it compusonics I think, and along with the peakfile,
they change the color of the line that is drawn for each peak to some
color to represent some statistic of the data. They say you can see more
about the general 'feeling' of the data. Its probably some frequency
response color.
NEW IDEA FOR ARDOUR
I was thinking that another cool Idea would be something along these
lines. This will be computationally expensive, but would be a neat
option. Behind the peak view, generate a thermal FFT of those peak bins
and draw that behind. I forget what this methiod of looking at sound data
is, but then as you go across the recorded data, you can see the makeup of
the data.
Let me know what you think.
Rick
On Tue, 27 Feb 2001, Robert Schrem wrote:
> > No, I think you are missing the point entirely.
>
> I think I'm not ENTIRELY wrong ;)
>
> In the end there are different ways how peak data can
> be computed, just as stated. Min/Max values is one
> way, but (non digital) peak instruments in studio
> equipment don't work this way and they are often
> referenced as useful. I just wanted to state this...
>
> > And when you are mixing a bunch of waves, this lets you
> > know how large the values are within ranges.
>
> I think this is not true in many cases because you would
> still just don't care about the phase of the mixed
> signals.
>
> Example: If you mix two signals of white noise with
> min/max = +50%/-50% you woun't get a resulting signal
> of +100%/-100%. This is espacially true for the case
> of mixing two identical sine waves with an phase
> shift between them (signals with slight phase shifts
> are not too seldom: e.g. stereo recording, mixing
> raw and effect signals...)
>
> robby
>
+------------------------+-----------------------+
| T a l i t y | +------+ |
+------------------------+ +----+-+ | |
| Richard Burnett | +-+ | |
| Senior Design Engineer +---+ +----+ |
| burnett_AT_tality.com | | |
| | | |
| Phone: 919.380.3014 | |
| Fax: 919.380.3903 | | |
+------------------------------------------------+
This archive was generated by hypermail 2b28 : Tue Feb 27 2001 - 21:03:55 EET