diff --git a/doc/FAQ.html b/doc/FAQ.html
index 16973a7e..ff49567d 100644
--- a/doc/FAQ.html
+++ b/doc/FAQ.html
@@ -44,6 +44,7 @@ a:visited { color: #0000EE; }
Can LameXP be used to convert/extract tracks from an Audio CD?
Why is the maximum normalization level limited to -0.5 dB?
What do the different 'Equalization' modes do?
+Why does my encoded file come out slightly longer than the original?
Why is the Hibernation option disabled (greyed out) on my system?
Why do I get the error 'Executable doesn't support Windows compatibility mode' on startup?
Why do I get the error 'Executable requires Windows XP or later' on startup?
@@ -644,6 +645,42 @@ WARNING: The "None" mode appears to be broken with more than two channels! This
+Why does my encoded file come out slightly longer than the original?
+
+Because of the way most (lossy) audio compressors work, a number of "silent" samples have to be prepended to
+the first input sample. Also a number of silent samples may be appended to the last input sample. Obviously
+these additional "padding" samples will make the encoded file longer than the original input. If the decoder
+does not compensate for the padding samples that have been added by the encoder, the padding samples at the
+beginning of the encoded file will cause a certain delay/shift, compared to the (uncompressed) original file.
+Another effect is that there will be a "gap" between the files, when playing multiple files in sequence.
+
+It has to be noted that the padding is an inherent property of the audio compressor that can not be avoided.
+But, as mentioned before, the padding samples can be removed by the decoder. In that case, the decoded output
+will have the exactly same length (in samples) as the original file. However this can only be done, if the
+decoder is able to determine the exact number of padding samples that had been added. While the Vorbis audio
+format stores the number of padding samples in the stream and thus any compliant decoder will remove them,
+the MP3 and AAC audio formats do NOT define an "official" way to store the the number of padding samples!
+
+Nonetheless, the LAME MP3 encoder, which is used by LameXP, uses its own "non-standard" way to indicate the
+exact number of padding samples. This is done by adding a so-called "LAME header" to the MP3 file. Hence MP3
+decoders, which recognize the LAME header, will be able to remove the padding samples. This feature often is
+referred to as "gapless" decoding/playback. The 'mpg123' decoder, which is used by LameXP, does support the
+gapless decoding. This means that if you encode MP3 files with LameXP/LAME, then gapless decoding will be
+possible. And if you decode these files with LameXP/mpg123 again, they will come out at the original length.
+
+Still you have to be aware that NOT all MP3 decoders support the LAME header. So if you encode MP3 files with
+LameXP and then decode them with an arbitrary MP3 decoder, they may come out at a "wrong" length. And there
+is absoloutely nothing that LameXP could do about that! Furthermore, if you decode a MP3 file in LameXP that
+has been created with some MP3 decoder which did not add a LAME header, then it is impossible to remove the
+padding samples. That's because in this situation there is NO way to know the number of padding samples!
+
+Some more info on this topic can be found here:
+
+
Why is the Hibernation option disabled (greyed out) on my system?
In order to use "Hibernation", your system must support the S4 power saving mode (aka "Suspend-to-Disk") and