Ok, like plain accidents are never caused by one error only, two hours of not understanding at first, brought this :
1. The fact that the Crack Detect only tripped at a second track with Cue Files was caused by *after* the Crack Detection code was applied for Invert. See 2.
2. The code for Invert appeared to be wrong (and it always has been so); this is about an easy to make logical mistake and specific code needed to capture for minus (voltage) values going to -32768 (in digital decimal for 16 bit files) and positive values going to 32767. At inverting something has to be done about this non-lineairity, and the code did this *without* being subjective to the Inverting itself. See 3.
Regarding 1., after the first Crack Detect analysis, the whole Cue File album is changed (wrongly because of this), and at the second track the wrongly applied changes are captured ...
3. The bad news ... (for Dave) : With the Cornershop album als the example, this pressing is clipping all over. This means that digital (decimal) values reach their maximum ... -32768 and +32767 respectively. The wrong code from 2. tripped on that and changed the minimum voltage into the maximum voltage. Crackle ...
The 13750 number of heavy cracks as this album shows at Crack Detection, shows that this number of times the data actually clips.
At solving this, I found that only with the Digital Volume at -0dB and the DAC set at being 16 bits this all happened. Since I always use the Digital Volume, and at some further checking, I found that normal 16 bit playback without Digital Volume had some wrong settings applied, and all played at Mickey Mouse speed. I thought "strange that no one noticed", and then found this 5 minutes ago :
Speed problem with 44/16 playback in U-8 version.
