jlohl

I prefer to avoid adding more graphs. So following your suggestion, I've added a (Lrms-Rrms) in the timeline to check channel separation. See picture : Those graphs show very diffrent kind of recordings : - first is overly compressed and limited, less dynamics, with lots of bass added : poor job of recording/mastering - second track is spectrally balanced, not compressed nor limited, great dynamics : a very good work You can see that the difference between peaks and rms values on the timeline is a good indication of the real dynamic range of a track. I also did some other minor mods : - bug removed : the spectrum was not shown when track had real silences (samples of value 0, often seen on mp3 files) - rms histogram is normalised to highest value Release is available for download at same place : http://www.ohl.to/audio/downloads/trackalyzer.zip Have a nice day !

Look at the first post I sent. The L+R level histogram gives the number of samples at a given level, if you see vertical lines at -1 and +1, it means that there are some fullscale overloads.

just a few comments on Mani's analysis : - peak+rms/time is a timeline of levels, above curve are the peaks and under are the rms. The diffrence between both curves can give you an idea of the dynamics evolution of the track. In the mani example, there is a strange difference between vinyl and CD track from 60 to 120s. Why ? what was done at mastering ? - on the rms histogram, the values under -40dB are more frequent on the vinyl, I suspect it may be noise. - on the spectrum analysis, we see big diffrences : the vinyl has some low frequency noise under 30Hz (turntable noise ?), the Eq are different near 50Hz : is it rumble or did the vinyl mastering add some "punch" ? Above 3 kHz, the difference is huge, it can be some mastering EQ but maybe the pickup frequency response has a big peak. That may be nice here because this track is missing higher frequencies : above 1kHz, the CD curve is quickly decreasing. I also see a problem in the soft : the peaks in the L+R level histogram should not give 0dB in the peak values on the right plot. I have to check this.

Hi all, this is my first post here and I see that you are quite active on anaysis of audio masterings, tracks, dynamic, aso... For my personal use, I did a free software to ease this analysis task. It only works on .wav files (mono, stereo, most samplerates and bitrates) so mp3, aac, flac have to be converted before (ie with foobar). For traks with emphasis, de-emphasis is available. Just choose two tracks to be compared and the soft outputs a picture with most important values : here is a comparison of Beck's Morning Phase, HDtracks vs CD. You can see that despite a lower DR value, the HDtracks version has more dynamics (see histogram and timeline). Note that I don't get exact same DR values as TT-DR and I prefer to measure LRA (loudness range as per ITU 3342). Those who want to play with this tool can get it at http://www.ohl.to/audio/downloads/trackalyzer.zip Sorry for mac and linux users, it only works on windows now. The soft needs Octave to be installed, there is a link in the setup page. The trackalyzer algorithms are open and can be checked in the trackalyzer.m file. Have a nice day and let me know if it works fine for you. JLO