HydrogenAudio

Hydrogenaudio Forum => Uploads => Topic started by: krabapple on 2013-04-02 17:49:41

Title: Pretzel logic comparison EQ difference graph
Post by: krabapple on 2013-04-02 17:49:41
data derived from Audition 1.0 'Frequency Analysis' tool using Blackmann-Harris FFT, FFT Size = 4096, graphed in Excel.  Full description here: http://www.hydrogenaudio.org/forums/index....st&p=829758 (http://www.hydrogenaudio.org/forums/index.php?showtopic=92856&view=findpost&p=829758)
Title: Pretzel logic comparison EQ difference graph
Post by: Propheticus on 2013-04-02 17:58:30
Quite obviously equalized/altered far beyond the small changes (if any!) you'd expect from passing input untouched and only tweaking things like systemload(killing services)/cpu priority and the such.

Edit: my guess is a soundcard (controlpanel/driver) DSP
Title: Pretzel logic comparison EQ difference graph
Post by: dhromed on 2013-04-02 21:29:15
And my own, with all three.
Title: Pretzel logic comparison EQ difference graph
Post by: LithosZA on 2013-04-02 21:36:48
Which one is correct?
Title: Pretzel logic comparison EQ difference graph
Post by: krabapple on 2013-04-02 22:22:41
dhromed's is plotting the raw values (possibly level-matched?) and presents the log scale differently.    Mine's a difference plot of the replaygain-adjusted raw values vs. reference -- but in any case my first plot was wrong -- I made a mistake transferring the data for the reference file.  Here's the corrected graph.  The description of the slightly altered input files and workflow is here (http://www.hydrogenaudio.org/forums/index.php?showtopic=92856&view=findpost&p=829832).  Basically in this version all three of the files have been adjusted according to replaygain values before comparison, and all three files are of the same segment of audio.
Title: Pretzel logic comparison EQ difference graph
Post by: krabapple on 2013-04-03 01:55:22
here's the same graph, with one more curve added:  the result of  Pretzel_Logic.wav(played in foobar, output = Primary Sound Driver)-->recorded with Total Recorder (48kHz/16bit,w hich is the native format of this setup)-->trim to match the audio segment (audition)--->convert to 44.1kHz(audition) --> replaygain scan with Foobar--->frequency profile scan with Audition-->graph (with RPG adjusted levels) 

This is basically a test to see how foobar2k behaves with my work system (which is not set to output digital audio, and which outputs audio at 48kHz) compared to what sabreghagen got.  The result is the blue line.