Vincent,
When you generate your profileX, you should use the "spread" option
instead of the default option "error on the mean"
Replace the line
TProfile * prof = h2->ProfileX("profile");
by
TProfile * prof = h2->ProfileX("profile",-1,-1,"s");
Rene Brun
Vincent Régnard wrote:
>
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hello all,
> I am playing with the profile representation of 2Dhisto but encountering some
> strange behaviour:
> I fill a 2Dhisto from a tree then call for the ProfileX method to get the
> profile and plot it.
> Means are ok, but variance (on the graphical representation) seems to be a
> lot under estimated, it seems ther is a wrong multiplication factor
> somewhere. Meanwhile a 1D plot of a slice along X gives me the right answer
> with correct mean and width which is not reproduced after in the profile.
> Hope you can help me.
>
> You find a superposition of a counter plot and associated profile here:
> http://vincentregnard.free.fr/tmp/profile.png
> and the code responsible for that koind of result there:
> http://vincentregnard.free.fr/tmp/alpha.C
>
> I am using root version 3.10.01 on GNU/Debian gcc 3.0.4
>
> - --
> \|/ V Régnard
> / ~~ GPG and ID: http://vincentregnard.free.fr/html/persopages/pgp.html
> @--OO Key fingerprint = F350 E39E 8C1B 4E8E D078 6D77 81D8 B38A E8F3 ED1E
> \ ? Key ID: E8F3ED1E
> || Please note that all mail containing html code will land in my junk box.
> / \ and all mail containing M$formated document will be bounced
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.3 (GNU/Linux)
>
> iD8DBQE/3cOdgdiziujz7R4RAqezAKCMZpomMa/+bblmqjcHXeabn1JIGQCfenz9
> OpIAN6FFLVUN3SNGvCJ9RU8=
> =DZUX
> -----END PGP SIGNATURE-----
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:17 MET