jar_project/notes
2020-10-19 12:29:04 +02:00

53 lines
2.6 KiB
Plaintext

machen:
- wenn benötigt sin response fit mit neuer phaseshift berechnung nochmal durchlaufen lassen, dann phasshift all
- abbildung erstellen mit custom cutoff frequencies über ganzen bereich (0.001Hz-1Hz) um hoffentlich zu zeigen dass dabei lineare
Gerade entsteht, vergleichen mit uniformen Bereich bei Daten bei dem es sich auch eher linear verhält um zu zeigen auf was wir hinaus wollen
- filter: zieht mean von einer amfreq periode ab wodurch alles was nicht damit in Verbindung steht herausfiltert,
auch JAR. problematisch wird dies eher wenn JAR-Anstieg schneller abläuft als eine amfreq periode
- wenn fit nicht funktioniert einfach weglassen, wenn sättigung vorhanden nochmal anschauen
- base_eod und q10 und temp für eigenmannia
+ figures:
apteronotus: fundament by tims bachelor thesis, important that apteronotus only shifts his frequency up (as eigenmannia doesnt --> natalies measurements)
+ spectogram
+ jar trace out of specgram
+ filtering of jar trace: mean noise cut --> subtracting jar response over whole stimulus
+ fit and jar trace --> gain and phaseshift
+ this for different am-frequencies and delta f (-15/-5Hz) --> compare gain for them
+ gain curve for one or more single fish
+ fit of gain curve for cutoff frequency and tau
+ gain curve for all fish taken together
+ single gain curves inside gain curve for all fish --> different cutoff frequencies --> comparison to metzen/chacron
--> fig_apt_specgram,
fig_apt_jar_filter_fit,
fig_apt_rms_gaincurve,
fig_apt_gaincurve_cutoff_tau,
sin_all_normal (without single gaincurves),
sin_all_uniform (with gaincurves for 5Hz)
!!! discard 2019lepto03 -5Hz --> RMS always over threshold
!!! discard 2020lepto04 gain fit doesnt work
!!! discard 2020lepto19 gain fit doesnt work
eigenmannia:
+ deltaf / response: -2Hz different, show it
+ spectogram
+ direct to fit and jar trace
+ gain curve for one or more single fish
+ gain curve for all fish taken together !!! gains without filtering by RMS
- (step response eigen)
--> fig_eig_specgram
fig_eig_jar_filter_fit,
fig_eig_rms_gaincurve,
fig_eig_gaincurve_cutoff_tau,
sin_eig_normal (without single gaincurves),
sin_eig_uniform (with gaincurves for 5Hz)
!!! discard 2015eigen8 --> RMS always over threshold
??? mean_noise_cut_eigen?
fish properties:
+ parameters
+ cutoff frequency - dominance score
offen: phaseshift, not working fit for some fish
(
- 2019lepto27/30: 27 - 0.05Hz (7-27-af, erste dat mit len(dat)=1), 30 - 0.001Hz (7-30-ah mit 0.005 anstatt gewollten 0.001Hz --> fehlt)
)