Would it be possible to show EAC peaks at db.cuetools.net? The same peaks are also in the CueTools log.
Something I just noticed recently: CUETools and EAC don't calculate peak values the same way. All other things being equal, CUETools will return a peak value of 100.0 if a track contains a max sample value of +7FFF or a min sample value of -8000 -- whereas EAC returns a peak value of 100.0% if a track contains a max sample value of +7FFF or a min sample value of -8000 or -7FFF.
The most common disc I've found so far that shows this difference is the South Park Bigger, Longer and Uncut soundtrack:
EAC rip log (v1.0b3, but I've confirmed v0.99pb3 returns the same peaks):
Track 1
Filename G:\tmp\testrip2\01 The People of South Park - Mountain Town.wav
Peak level 100.0 %
Extraction speed 10.3 X
Test CRC EFE9C48B
Accurately ripped (confidence 24) [12AB2BF3] (AR v2)
Copy finished
Track 2
Filename G:\tmp\testrip2\02 Terrance & Phillip - Uncle Fucka.wav
Peak level 100.0 %
Extraction speed 17.7 X
Test CRC 4AAC8A07
Accurately ripped (confidence 24) [C2B6672D] (AR v2)
Copy OK
Track 3
Filename G:\tmp\testrip2\03 Mr Mackey - It's Easy, Mmmkay.wav
Peak level 100.0 %
Extraction speed 18.5 X
Test CRC 95CF979A
Accurately ripped (confidence 24) [A4228687] (AR v2)
Copy OK
CUETools log:
Track Peak [ CRC32 ] [W/O NULL] [ LOG ]
-- 99.9 [684C677E] [A4F04842] W/O NULL
01 99.9 [EFE9C48B] [D180B7B0]
02 99.9 [4AAC8A07] [8B7B00A1]
03 99.9 [95CF979A] [2F03A86F]
Actual sample values (returned by SoX):
Trk Length EACpk% LMin LMax RMin RMax LRMSdB RRMSdB Indicies
T01 04:27:38 100.0% -7FFF +7FFE -7FFF +7FFE -14.62 -14.86 00:00:00 00:00:32
T02 01:06:00 100.0% -7FFF +7FFE -7FFF +7FFE -12.36 -13.14 04:26:57 04:27:70
T03 01:54:37 100.0% -7FFF +7FFE -7FFF +7FFE -14.23 -14.40 05:33:17 05:33:70
-7FFF really isn't the maximum negative value so of the two apps I think CUETools is doing the correct thing technically (100% - 1 isn't 100%), but it does call into question any comparisons between EAC-generated peak values and CUETools-generated peak values. (From what I've been able to tell they seem to agree about the difference between 99.8% and 99.9% but I haven't dug that deeply. I really wish EAC just reported max/min sample values per channel instead of boiling to down to a much-less-accurate percentage.)