Normalized value of MIDI sources (before unipolar or bipolar mapping) is always in the range [0..+1], regardless of MIDI event (even for FLUID_MOD_PITCHWHEEL).
Actually some basic modulators check are done at noteon time (in fluid_voice_add_mod()). That means that we know if a modulator (modx) is invalid only when a MIDI noteon is received and only for a preset modx belongs to. This is not appropriate.
This moves the modulator checking at soundfont loading time. Enhancements are:
1) A better verbose modulator integrity check, for any soundfont loaded at appropriate time.
1.1) All modulators are checked (preset zone (local/global), instrument zone (local/global).
1.2.1) Modulators check are enforced to source src1 and src2 (for non-CC and CC sources) (following SF specs (except for CC LSB) ( see comment in fluid_synth_cc_LOCAL()).
Modulators CC sources checking is coherent with the actual behaviour in fluid_synth_cc_LOCAL() in regard of modulation triggering.
1.2.2) Also, identical modulator in the same zone are detected.
1.2.3) Any invalid modulator(sources invalid, or modulator identical) is removed at loading time with a warning message displaying the cause and name of the modulators.
2) This fix a bug in noteon, in the case of identical modulators in global preset zone.
Assuming 2 identical modulator (m1 and m2) in a preset global zone, the actual noteon doesn't check this case (the actual code detect identical modulator in all others zones (instrument (local or global), preset(local)) but not preset global).
3) NoteOn is faster.
3.1)There is no more modulators checks at noteon making this more efficient.
3.2) As there are no identical modulator in the same zone, there is no more identity modulator check (i.e local zone against local zone), (i.e global zone against global zone). This result in a faster code and the bug described in (2) is gone.
4) Modulators sources checking as been added in API functions fluid_synth_add_default_mod() and fluid_voice_add_mod(). Please
- When primary source input (src1) is set to General Controller
'No Controller', output will be forced to 0.0 at synthesis time
(see fluid_mod_get_value()).
That means that the minimum value of the modulator will be always 0.0.
We need to force amount value to 0 to ensure a correct evaluation of the
minimum value later (see fluid_voice_get_lower_boundary_for_attenuation()).
- When secondary source input (src2) is set to General
Controller 'No Controller', output will be forced to +1.0 at synthesis time
(see fluid_mod_get_value()).
That means that this source will behave unipolar only. We need to force the
unipolar flags to ensure the modulator will behave correctly later in
fluid_voice_get_lower_boundary_for_attenuation().
This change removes the warning message if retrieving the soundfont file
modification time fails in the sample cache loader. The warning made sense
while we didn't have mtime support on all platforms, but after switching
to GLibs g_stat, it's no longer needed.
To reduce the number of calls to fluid_get_file_modification_time and also
get rid of a possible race condition, getting the mtime has been moved to
fluid_samplecache_load.
This change also fixes#483 because it removes the bogus warning messages
if a soundfont was loaded from memory by abusing the filename to store a
pointer.
- v variable is renamed min_val for clarity.
- This fix a bug when calculating the possible minimun
value (min_val) of a modulator. min_val should only
be dependant of:
- signs of mapped sources (ie bipolar [-1..+1] or unipolar [0..+1])
- sign of amount.
- absolute value of amount.
If we move UNKN_ID to the bottom of the enum, in the for() cycle inside chunkid() if the value is not found then the "i" variable will be already UNKN_ID.
I tried to check the generated ASM code when WITH_FLOAT macro is defined and when it is undefined. In both cases I noticed that there are some points where the math number are converted from float to double and viceversa. Actually this happens whether the FPU is present or not, perhaps for granting the right precision in digits. I also noticed few points that could be simplified a bit by using integer math, so I also included them.
* Since fluid_voice_get_actual_key() returns an integer value, the value 60 could be subtracted directly as integer, the code generated looks a bit better.
* FLUID_NOISE_FLOOR needs to be cast to fluid_real_t to generate a bit more efficient code.
* DC_OFFSET needs a cast to fluid_real_t to generate a bit more efficient code.
* "last_fres" and "output_rate" are already fluid_real_t, so they do not need type cast.
* "chan_add", "par1_add" and "par2_add" are integer types, so they can be added without FPU. "chan_mul", "par1_mul" and "par2_mul" are already fluid_real_t, so they do not need type cast. Instead, the constant 0.5 needs cast to fluid_real_t to be more efficient.
Libreadline is already used into fluid_istream_readline().
In my opinion, there is no need to have this duplicated code, so I would suggest to move the call to add_history() into fluid_sys.h
Closes#460.
