#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
A collection of functions to dsplay songs and results.
"""
import numpy as np
import matplotlib as mpl
import matplotlib.pyplot as plt
from mpl_point_clicker import clicker
from wavesongs.utils.tools import (
klicker_time,
klicker_multiple,
_LABELS,
_COLORS,
_MARKERS
)
from wavesongs.models.bird import bifurcation_ode, _F1, _F2
from librosa.display import specshow as Specshow
from mpl_point_clicker import clicker
from matplotlib.gridspec import GridSpec
from matplotlib.ticker import FuncFormatter, NullFormatter
from typing import (
Optional,
Tuple,
Union,
AnyStr,
Any,
TypeVar,
List
)
Syllable = TypeVar('Syllable')
Song = TypeVar('Song')
## ------------- Bogdanov–Takens bifurcation ------------------
beta_bif, mu1_curves, f1, f2 = bifurcation_ode(_F1, _F2)
# --------------------------
_COLORES = {
"Argentina": ["Blues", "lightblue", "blue"],
"Bolivia": ["Purples", "plum", "purple"],
"Brazil": ["Greys", "lightgray", "black"],
"Chile": ["Oranges", "bisque", "chocolate"],
"Colombia": ["Reds", "lightsalmon", "red"],
"Costa Rica": ["cool", "paleturquoise", "teal"],
"Ecuador": ["GnBu", "lightsteelblue", "steelblue"],
"Peru": ["Greens", "darkseagreen", "darkgreen"],
"Uruguay": ["copper", "peachpuff", "orange"],
"Venezuela": ["RdPu", "lightpink", "mediumvioletred"],
}
_CMAP = "magma"
_TITLE_FONTSIZE = 18
#%%
[docs]
def _suptitle(obj) -> AnyStr:
"""
Parameters
----------
obj : Syllable | Song
_description_
Returns
-------
title : AnyStr
Title template
"""
format = obj.file_name[-3:]
file_name = obj.file_name[:-4].replace("synth_","")
title = f"{file_name}-{obj.no_syllable}-{obj.type}.{format}" \
if obj.type!="" else f"{file_name}-{obj.no_syllable}.{format}"
return title.replace(" ","")
#%%
def _save_name(obj) -> AnyStr:
file_name = obj.file_name[:-4]
img_text = f"{file_name}-{obj.no_syllable}-{obj.type}" \
if obj.type!="" else f"{file_name}-{obj.no_syllable}"
return img_text.replace(" ","")
#%%
[docs]
def alpha_beta(
obj: Any, # Union[Syllable,Song],
xlim: Tuple[float] = (-0.05, 0.2),
ylim: Tuple[float] = (-0.2, 0.9),
figsize: Tuple[float] = (10, 6),
save: bool = True,
show: bool = True,
) -> None:
"""
Parameters
----------
obj : Syllabe|Song
Song or Syllable to be displayed
xlim : tuple = (-0.05,.2)
Time range
ylim : tuple = (-0.2,0.9)
Frequency range
figsize : tuple = (10,6)
Fogure size (width, height)
save : bool = True
Enable save plot
show : bool = True
Enable display plot
Return
------
None
Example
-------
>>>
"""
if not "synth" in obj.id:
raise Exception("This is not a synthetic syllable, remember create"
+ " a synthetic file using the funcion bs.Solve().")
plt.close()
if obj.alpha.max() > 0.2: xlim = (-0.05, 1.1 * obj.alpha.max())
if obj.beta.max() > 0.9: ylim = (-0.2, 1.1 * obj.beta.max())
viridis = mpl.colormaps["Blues"]
c = viridis(np.linspace(0.3, 1, np.size(obj.time_s)))
fig = plt.figure(figsize=figsize)
gs = GridSpec(2, 2, figure=fig, wspace=0.1, hspace=0.35)
# fig.tight_layout(pad=3.0)
gs.update(top=0.85, bottom=0.1, left=0.075, right=0.935)
ax1 = fig.add_subplot(gs[0, 0])
ax1.scatter(obj.time_s, obj.alpha, c=c, label="alfa")
ax1.set_title("Air-Sac Pressure")
ax1.set_ylabel("α (a.u.)")
ax1.set_ylim(xlim)
ax1.grid()
ax2 = fig.add_subplot(gs[1:, 0])
ax2.scatter(obj.time_s, obj.beta, c=c, label="beta")
ax2.set_title("Labial Tension")
ax2.set_xlabel("Time (s)")
ax2.set_ylabel("β (a.u.)")
ax2.set_ylim(ylim)
ax2.sharex(ax1)
ax2.grid()
# ------------- Bogdanov–Takens bifurcation ------------------
ax3 = fig.add_subplot(gs[:, 1])
ax3.plot(-1 / 27, 1 / 3, "ko") # , label="Cuspid Point");
ax3.axvline(0, color="red", lw=1) # , label="Hopf Bifurcation")
ax3.scatter(obj.alpha, obj.beta, c=c, marker="_", label="Motor Gesture")
# label="Saddle-Noddle\nBifurcation"
ax3.plot(mu1_curves[0], beta_bif, "-g", lw=1)
ax3.plot(mu1_curves[1], beta_bif, "-g", lw=1)
ax3.fill_between(
mu1_curves[1],
beta_bif,
10,
where=mu1_curves[1] > 0,
color="gray",
alpha=0.2,
)
ax3.text(-0.01, 0.6, "Hopf", rotation=90, color="r")
ax3.text(-0.0425, 0.37, "CP", rotation=0, color="k")
ax3.text(-0.0275, 0.15, "SN", rotation=0, color="g")
ax3.text(0.1, 0.005, "SN", rotation=0, color="g")
ax3.set_ylabel("Tension (a.u.)")
ax3.set_xlabel("Pressure (a.u.)")
ax3.set_title("Parameter Space")
ax3.yaxis.tick_right()
ax3.yaxis.set_label_position("right")
ax3.set_xlim(xlim)
ax3.set_ylim(ylim)
ax3.sharey(ax2)
ax3.legend()
if obj.type!="":
gs.update(top=0.8)
suptitle = f"Motor Gesture Curves\n{_suptitle(obj)}"
else:
suptitle = f"Motor Gesture Curves: {_suptitle(obj)}"
# fig.tight_layout()
plt.suptitle(
suptitle,
fontsize=_TITLE_FONTSIZE,
y=0.99,
fontweight="bold",
)
if save:
save_name = f"{_save_name(obj)}-mg_params.png"
fig.savefig(
obj.proj_dirs.IMAGES / save_name,
transparent=True,
bbox_inches="tight",
)
print(f"Image save at {save_name}")
if show:
plt.show()
else: plt.close()
# %%
[docs]
def phsyical_variables(
obj: Any, # Union[Syllable,Song],
xlim: Tuple[float] = (),
figsize: Tuple[float] = (10, 6),
save: bool = True,
show: bool = True,
) -> None:
"""
Parameters
----------
obj : Syllabe|Song
Song or Syllable to be displayed
xlim : tuple = (-0.05,.2)
Time range
figsize : tuple = (10,6)
Fogure size (width, height)
save : bool = True
Save plot
show : bool = True
Display plot
Return
------
files_names : list
List with the audios files names
Example
-------
>>>
"""
if not "synth" in obj.id:
raise Exception("This is not a synthetic syllable, remember create"
+ " a synthetic file using the funcion bs.Solve().")
plt.close()
if xlim == ():
xlim = (obj.times_vs[0], obj.times_vs[-1])
fig, ax = plt.subplots(2, 2, figsize=figsize, sharex=True)
plt.subplots_adjust(
hspace=0.25, wspace=0.2, top=0.825, bottom=0.1, left=0.075, right=0.99
)
ax[1, 0].ticklabel_format(
axis="y", style="scientific", scilimits=(-1, 1)
)
ax[1, 0].plot(obj.times_vs, obj.vs[:, 1], color="g")
ax[1, 0].set_ylabel("$p_{in}$ (x $10^3$)")
ax[1, 0].set_title(r"Trachea Input Pressure")
ax[1, 0].set_xlabel("Time (s)")
ax[1, 0].set_xlim(xlim)
ax[1, 1].plot(obj.times_vs, obj.vs[:, 4], color="b")
ax[1, 1].set_ylabel("$p_{out}$")
ax[1, 1].set_title(r"Trachea Output Pressure")
ax[1, 1].set_xlim(xlim)
ax[1, 1].set_xlabel("Time (s)")
ax[0, 0].plot(obj.times_vs, obj.vs[:, 0], color="r")
ax[0, 0].set_title(r"Labial Walls Displacement")
ax[0, 0].set_ylabel("$x(t)$")
ax[0, 0].set_xlim(xlim)
ax[0, 1].plot(obj.times_vs, obj.vs[:, 0], color="m")
ax[0, 1].set_ylabel("$y(t)$")
ax[0, 1].set_title(r"Labial Walls Velocity")
ax[0, 1].set_xlim(xlim)
if obj.type!="":
plt.subplots_adjust(top=0.8)
suptitle = f"Physical Model Variables\n{_suptitle(obj)}"
else:
suptitle = f"Physical Model Variables: {_suptitle(obj)}"
fig.suptitle(
suptitle,
fontsize=_TITLE_FONTSIZE,
y=0.99,
fontweight="bold",
)
# fig.tight_layout()
if save:
image_text = f"{_save_name(obj)}-PhysicalVariables.png"
fig.savefig(
obj.proj_dirs.IMAGES / image_text,
transparent=True,
bbox_inches="tight",
)
print(f"Image save at {image_text}")
if show:
plt.show()
else:
plt.close()
# %%
[docs]
def spectrogram_data(
syllable: Optional[Syllable] = None,
ff_on: bool = False,
tlim: Optional[Tuple[float]] = None,
figsize: Tuple[float] = (10, 6),
ms: int = 7,
labels: List[AnyStr] =_LABELS,
colors: List[AnyStr] =_COLORS,
markers: List[AnyStr] =_MARKERS,
) -> clicker:
"""
Parameters
----------
obj : Syllabe|Song
Song or Syllable to be displayed
syllable: Syllable|None = None,
chunck: Any|None = None,
ff_on: bool =False,
select_time: bool = False,
tlim : tuple = (-0.05,.2)
Time range
figsize : tuple = (10,6)
Fogure size (width, height)
save : bool = True
Save plot
show : bool = True
Display plot
ms : int = 7
Marker size
Return
------
klicker : cliker
Clicker object with the points selected
Example
-------
>>>
"""
ticks = FuncFormatter(lambda x, pos: f"{x*1e-3:g}")
ticks_x = FuncFormatter(lambda x, pos: f"{x+syllable.t0_bs:.2f}")
if tlim is None:
tlim = (syllable.time[0], syllable.time[-1])
else:
tlim = (tlim[0] - syllable.t0_bs, tlim[1] - syllable.t0_bs)
plt.close()
fig, ax = plt.subplots(1, 1, figsize=figsize)
img = Specshow(
syllable.Sxx_dB,
x_axis="s",
y_axis="linear",
sr=syllable.sr,
hop_length=syllable.hop_length,
ax=ax,
cmap=_CMAP,
)
ax.yaxis.set_major_formatter(ticks)
ax.xaxis.set_major_formatter(ticks_x)
if ff_on:
if syllable.ff_method in ("yin", "pyin", "manual"):
ax.plot(
syllable.time, syllable.FF, "co", ms=ms, label=r"FF$_{yin}$"
)
elif syllable.ff_method == "both":
ax.plot(syllable.time, syllable.FF, "co", ms=ms, label=r"FF$_{pyin}$")
ax.plot(syllable.time, syllable.FF2, "b*", ms=ms, label=r"FF$_{yin}$")
ax.set_ylim(syllable.flim)
ax.set_xlim(tlim)
ax.set_ylabel("Frequency (kHz)")
ax.set_xlabel("Time (s)")
fig.suptitle(
"Waveform and Spectrogram",
fontsize=_TITLE_FONTSIZE,
y=0.99,
fontweight="bold",
)
fig.tight_layout()
plt.subplots_adjust(
top=0.9, bottom=0.075, left=0.075, right=0.86
)
return klicker_multiple(fig, ax,
labels=labels,
colors=colors,
markers=markers)
# %%
# %%
[docs]
def syllables(
obj: Any, # Union[Syllable,Song],
obj_synth: Any, # Union[Syllable,Song],
ff_on: bool = False,
figsize: Tuple[float] = (11, 6),
save: bool = True,
show: bool = True,
) -> None:
"""
Parameters
----------
obj : Syllable | Song
obj : Syllable | Song
ff_on : bool = False
Falg to enable fundamental frequency visualization
save : bool = True
Flag to save plot
show : bool = True
Flag to display plot
Return
------
None
Example
-------
>>>
"""
plt.close()
ticks = FuncFormatter(lambda x, pos: f"{x*1e-3:g}")
fig, ax = plt.subplots(2, 2, figsize=figsize, sharex=True)
img = Specshow(
obj.Sxx_dB,
x_axis="s",
y_axis="linear",
sr=obj.sr,
hop_length=obj.hop_length,
ax=ax[0, 0],
cmap=_CMAP,
)
ax[0, 0].set_title("Real", fontweight="bold")
ax[0, 0].yaxis.set_major_formatter(ticks)
ax[0, 0].set_ylim(obj.flim)
ax[0, 0].set_ylabel("Frequency (kHz)")
ax[0, 0].set_xlabel("")
img = Specshow(
obj_synth.Sxx_dB,
x_axis="s",
y_axis="linear",
sr=obj_synth.sr,
hop_length=obj_synth.hop_length,
ax=ax[0, 1],
cmap=_CMAP,
)
cbar_ax = fig.add_axes([0.95, 0.445, 0.015, 0.36])
clb = fig.colorbar(img, cax=cbar_ax, format="%+2.f")
clb.set_label("Power\n(dB)", labelpad=-25, y=1.2, rotation=0)
ax[0, 1].yaxis.set_major_formatter(ticks)
ax[0, 1].set_title("Synthetic", fontweight="bold")
ax[0, 1].set_ylim(obj.flim)
ax[0, 1].set_ylabel("")
ax[0, 1].set_xlabel("")
if ff_on:
ax[0, 0].plot(obj.time, obj.FF, "bo-", lw=2, label="FF")
ax[0, 0].legend(bbox_to_anchor=(0.975, 0.975))
ax[0, 1].plot(obj_synth.time, obj_synth.FF, "go-", lw=2, label="FF")
ax[0, 1].legend(bbox_to_anchor=(0.975, 0.975))
t_end = obj.time[-1] + obj.time[-1] - obj.time[-2]
ax[1, 0].plot(obj.time_s, obj.s, label="waveform", c="b")
ax[1, 0].set_xlim((obj.time_s[0], t_end))
ax[1, 0].plot(obj.time_s, obj.envelope, label="envelope", c="darkblue")
ax[1, 0].set_ylabel("Amplitud (a.u.)")
ax[1, 0].set_xlabel("Time (s)")
ax[1, 0].legend()
ax[1, 1].plot(obj_synth.time_s, obj_synth.s, label="waveform", c="g")
ax[1, 1].plot(
obj_synth.time_s, obj_synth.envelope, label="envelope", c="darkgreen"
)
ax[1, 1].set_xlabel("Time (s)")
ax[1, 1].set_ylabel("")
ax[1, 1].legend()
# removing y ticks labels
ax[1, 1].yaxis.set_major_formatter(NullFormatter())
ax[0, 1].yaxis.set_major_formatter(NullFormatter())
plt.subplots_adjust(
wspace=0.05, hspace=0.075, left=0.07, top=0.85, right=0.93, bottom=0.07
)
if obj.type!="":
plt.subplots_adjust(top=0.8)
suptitle = f"Comparing Syllables\n{_suptitle(obj)}"
else:
suptitle = f"Comparing Syllables: {_suptitle(obj)}"
fig.suptitle(
suptitle,
y=0.99,
fontsize=_TITLE_FONTSIZE,
fontweight="bold",
)
if save:
img_name = f"{_save_name(obj)}-SoundAndSpectros.png"
fig.savefig(
obj.proj_dirs.IMAGES / img_name,
transparent=True,
bbox_inches="tight",
)
print(f"Image save at {img_name}")
# return fig, ax
if show:
plt.show()
else:
plt.close()
# %%
[docs]
def scores(
obj: Any, # Union[Syllable,Song],
obj_synth: Any, # Union[Syllable,Song],
figsize: Tuple[float] = (11, 8),
ylim: Tuple[float] = (),
save: bool = True,
show: bool = True,
) -> None:
"""
Parameters
----------
obj : Syllable | Song
obj_synth : Syllable | Song
figsize : tuple = (10,10)
Size of the figure (width, height)
ylim : tuple = ()
Frequnecy range
save : bool = True
Flag to save plot
show : bool = True
Flag to display plot
Return
------
None
Example
-------
>>>
"""
plt.close()
ticks = FuncFormatter(lambda x, pos: f"{x*1e-3:g}")
ticks_x = FuncFormatter(lambda x, pos: f"{x:.2g}")
fig = plt.figure(constrained_layout=False, figsize=figsize)
gs = fig.add_gridspec(
nrows=7,
ncols=5,
wspace=0.1,
hspace=0.9,
left=0.06,
top=0.925,
right=0.78,
bottom=0.075,
)
# vmin, vmax = obj.Sxx_dB.min(), obj.Sxx_dB.max()
# --------------- scores: FF and SCI ---------------------------------
labelFF = r"FF, $\overline{FF}$="
labelSCI = r"SCI, $\overline{SCI}$="
max_error = 100 * max(obj_synth.deltaFF.max(), obj_synth.deltaSCI.max())
ax1 = fig.add_subplot(gs[:2, :])
ax1.plot(
obj_synth.time,
100 * obj_synth.deltaFF,
"*-",
color="k",
ms=5,
lw=1,
alpha=0.8,
label=labelFF + str(round(100 * obj_synth.deltaFF_mean, 2)),
)
ax1.plot(
obj_synth.time,
100 * obj_synth.deltaSCI,
"*-",
color="purple",
ms=5,
lw=1,
alpha=0.8,
label=labelSCI + str(round(100 * obj_synth.deltaSCI_mean, 2)),
)
ax1.legend(bbox_to_anchor=(1.235, 1.05), borderpad=0.6, labelspacing=0.7)
ax1.xaxis.set_major_formatter(ticks_x)
ax1.set_xlim((obj_synth.time[0], obj_synth.time[-1]))
ax1.set_ylabel("Relative Error (%)")
ax1.set_xlabel("")
ax1.set_ylim((0, 1.25 * max_error))
# ------------------ spectrum ---------------
ax2 = fig.add_subplot(gs[2:5, :], sharex=ax1)
img = Specshow(
obj.Sxx_dB,
x_axis="s",
y_axis="linear",
sr=obj.sr,
hop_length=obj.hop_length,
ax=ax2,
cmap=_CMAP,
)
ax2.plot(obj.time, obj.FF, "b*-", label=r"real", ms=7)
ax2.plot(obj_synth.time, obj_synth.FF, "go-", label=r"synth", ms=3)
ax2.legend(borderpad=0.6, labelspacing=0.7)
ax2.yaxis.set_major_formatter(ticks)
ax2.xaxis.set_major_formatter(ticks_x)
ax2.set_xlim((obj.time[0], obj.time[-1]))
ax2.set_ylim(obj.flim)
ax2.set_ylabel("Frequency (kHz)")
ax2.set_xlabel("")
# ------------------ SCI -------------------------
ax31 = fig.add_subplot(gs[5:7, :], sharex=ax2)
ax31.set_ylabel(r"Similarity (dl)")
ax31.set_xlabel("Time (s)")
# --------------- acousitcal features ----------------------
ax32 = ax31.twinx()
label_lskl = r"SKL, $\overline{SKL}$="
label_lr = r"$SCI_{real}$, $\overline{SCI}$="
label_ls = r"$SCI_{synth}$, $\overline{SCI}$="
label_lh = r"DF, $\overline{DF}$="
label_lc = r"cor, $\overline{corr}$="
lr = ax32.plot(
obj.time,
obj.SCI,
"b*-",
ms=7,
label=label_lr + str(round(obj.SCI.mean(), 2)),
)
ls = ax32.plot(
obj.time,
obj_synth.SCI,
"go-",
ms=5,
alpha=0.8,
label=label_ls + str(round(obj_synth.SCI.mean(), 2)),
)
lh = ax32.plot(
obj.time,
obj_synth.Df,
"H",
ms=3,
label=label_lh + str(round(obj_synth.Df.mean(), 2)),
)
lskl = ax32.plot(
obj.time,
obj_synth.SKL,
"s",
color="purple",
ms=3,
label=label_lskl + str(round(obj_synth.SKL.mean(), 2)),
)
correlation_synth = obj_synth.correlation.mean()
lc = ax32.plot(
obj.time,
obj_synth.correlation,
"p",
ms=3,
label=label_lc + str(round(correlation_synth, 2)),
)
lns = lr + ls + lh + lskl + lc
labs = [l.get_label() for l in lns]
ax32.legend(
lns,
labs,
bbox_to_anchor=(1.075, 1.1),
title="Acoustical Features",
title_fontproperties={"weight": "bold"},
)
ax32.legend(bbox_to_anchor=(1.3, 1))
ax32.set_ylabel("SCI (dl)")
ax32.set_ylim((0, 5))
cbar_ax = fig.add_axes([0.8, 0.33, 0.02, 0.315])
clb = fig.colorbar(img, cax=cbar_ax)
clb.set_label("Power (dB)", labelpad=12, y=0.5, rotation=90)
if obj.type!="":
suptitle = f"Scoring Variables\n{_suptitle(obj)}"
gs.update(top=0.9)
else:
suptitle = f"Scoring Variables: {_suptitle(obj)}"
fig.suptitle(
suptitle,
fontsize=_TITLE_FONTSIZE,
y=0.99,
fontweight="bold",
)
if save:
img_name = f"{_save_name(obj)}-ScoringVariables.png"
fig.savefig(
obj.proj_dirs.IMAGES / img_name,
transparent=True,
bbox_inches="tight",
)
print(f"Image save at {img_name}")
if show:
plt.show()
else:
plt.close()
# %%
[docs]
def spectrum_comparison(
obj: Any, # Union[Syllable,Song],
obj_synth: Any, # Union[Syllable,Song],
cmap: Union[AnyStr] = "afmhot_r",
figsize: Tuple[float] = (11, 6),
sharey: bool = True,
save: bool = True,
show: bool = True,
) -> None:
"""
Parameters
----------
obj : Syllable | Song
obj_synth : Syllable | Song
figsize : tuple = (10,10)
Size of the figure (width, height)
sharey: bool = True,
Enable share y axis
save : bool = True
Flag to save plot
show : bool = True
Flag to display plot
Return
------
None
Example
-------
>>>
"""
if cmap is not None:
_CMAP = cmap
labelrotation = 90 if obj.time[-1] < 1 else 0
_fontproperties = {"size": 12, "weight": "bold"}
plt.close()
ticks = FuncFormatter(lambda x, pos: f"{x*1e-3:g}")
fig = plt.figure(figsize=figsize)
gs = fig.add_gridspec(
nrows=2,
ncols=3,
hspace=0.35,
wspace=0.2,
top=0.825,
bottom=0.15,
left=0.05,
right=0.95,
)
vmin = obj.Sxx_dB.min()
vmax = obj.Sxx_dB.max()
# ------------------ spectrogams ----------------------------
ax1 = fig.add_subplot(gs[0, 0])
img = Specshow(
obj.Sxx_dB,
x_axis="s",
y_axis="linear",
sr=obj.sr,
hop_length=obj.hop_length,
ax=ax1,
cmap=_CMAP,
)
clb = fig.colorbar(img, ax=ax1)
clb.set_label("Power\n(dB)", labelpad=-16, y=1.25, rotation=0)
ax1.tick_params(axis="x", which="both", labelrotation=labelrotation)
ax1.set_title("Real", fontproperties=_fontproperties)
ax1.set_xlim((obj.time[0], obj.time[-1]))
ax1.yaxis.set_major_formatter(ticks)
ax1.set_ylim(obj.flim)
ax1.set_ylabel("")
ax1.set_xlabel("")
if sharey:
ax2 = fig.add_subplot(gs[0, 1], sharex=ax1, sharey=ax1)
else:
ax2 = fig.add_subplot(gs[0, 1], sharex=ax1)
img = Specshow(
obj_synth.Sxx_dB,
x_axis="s",
y_axis="linear",
sr=obj_synth.sr,
hop_length=obj_synth.hop_length,
ax=ax2,
cmap=_CMAP,
)
clb = fig.colorbar(img, ax=ax2)
clb.set_label("Power\n(dB)", labelpad=-16, y=1.25, rotation=0)
ax2.tick_params(axis="x", which="both", labelrotation=labelrotation)
ax2.set_title("Synthetic", fontproperties=_fontproperties)
ax2.yaxis.set_major_formatter(ticks)
ax2.set_ylim(obj.flim)
ax2.set_ylabel("")
ax2.set_xlabel("")
# ------------------ Mel spectgrograms ------------------
if sharey:
ax3 = fig.add_subplot(gs[1, 0], sharex=ax1, sharey=ax1)
else:
ax3 = fig.add_subplot(gs[1, 0], sharex=ax1)
img = Specshow(
obj.FF_coef,
x_axis="s",
y_axis="linear",
sr=obj.sr,
hop_length=obj.hop_length,
ax=ax3,
cmap=_CMAP,
vmin=0,
vmax=100,
)
clb = fig.colorbar(img, ax=ax3)
ax3.tick_params(axis="x", which="both", labelrotation=labelrotation)
ax3.yaxis.set_major_formatter(ticks)
ax3.set_ylabel(" " * 65 + "Frequency (kHz)", loc="center", labelpad=10)
ax3.set_xlabel("")
ax3.set_ylim(obj.flim)
if sharey:
ax4 = fig.add_subplot(gs[1, 1], sharex=ax1, sharey=ax1)
else:
ax4 = fig.add_subplot(gs[1, 1], sharex=ax1)
img = Specshow(
obj_synth.FF_coef,
x_axis="s",
y_axis="linear",
sr=obj.sr,
cmap=_CMAP,
vmin=0,
vmax=100,
hop_length=obj_synth.hop_length,
ax=ax4,
)
clb = fig.colorbar(img, ax=ax4)
ax4.tick_params(axis="x", which="both", labelrotation=labelrotation)
ax4.yaxis.set_major_formatter(ticks)
ax4.set_xlabel("Time (s)", labelpad=10)
ax4.set_ylim(obj.flim)
ax4.set_ylabel("")
# ------------------ Delta Sxx - Mel ------------------------
if sharey:
ax5 = fig.add_subplot(gs[0, 2], sharex=ax1, sharey=ax1)
else:
ax5 = fig.add_subplot(gs[0, 2], sharex=ax1)
img = Specshow(
obj_synth.deltaSxx,
x_axis="s",
y_axis="linear",
sr=obj.sr,
vmin=0,
vmax=1,
ax=ax5,
cmap=_CMAP,
hop_length=obj_synth.hop_length,
)
ax5.set_title(r"Difference ($\Delta$)", fontproperties=_fontproperties)
ax5.tick_params(axis="x", which="both", labelrotation=labelrotation)
ax5.yaxis.set_major_formatter(ticks)
ax5.set_ylim(obj.flim)
ax5.set_ylabel("")
ax5.set_xlabel("")
clb = fig.colorbar(img, ax=ax5)
clb.set_label("Power\n(dB)", labelpad=-16, y=1.25, rotation=0)
if sharey:
ax6 = fig.add_subplot(gs[1, 2], sharex=ax1, sharey=ax1)
else:
ax6 = fig.add_subplot(gs[1, 2], sharex=ax1)
img = Specshow(
obj_synth.deltaMel,
x_axis="s",
y_axis="linear",
sr=obj.sr,
hop_length=obj_synth.hop_length,
ax=ax6,
cmap=_CMAP,
)
ax6.tick_params(axis="x", which="both", labelrotation=labelrotation)
ax6.yaxis.set_major_formatter(ticks)
ax6.set_ylim(obj.flim)
ax6.set_ylabel("")
ax6.set_xlabel("")
# ax6.yaxis.set_major_formatter(NullFormatter())
plt.text(0.04, 6e3, "MEL", rotation=90, fontproperties=_fontproperties)
plt.text(
0.04, 3.5e4, "Linear", rotation=90, fontproperties=_fontproperties
)
fig.colorbar(img, ax=ax6)
if obj.type!="":
plt.subplots_adjust(top=0.8)
suptitle = f"Comparing Spectral Content\n{_suptitle(obj)}"
else:
suptitle = f"Comparing Spectral Content: {_suptitle(obj)}"
fig.suptitle(
suptitle,
fontsize=_TITLE_FONTSIZE,
y=0.99,
fontweight="bold",
)
if save:
img_name = f"{_save_name(obj)}-ComparingSpectros.png"
fig.savefig(
obj.proj_dirs.IMAGES / img_name,
transparent=True,
bbox_inches="tight",
)
print(f"Image save at {img_name}")
# return fig, gs
if show:
plt.show()
else:
plt.close()