ForumsSoftware ← anomalous playback pitch in Kaivo?

Hi Randy,

Hate to have my first post about Kaivo be a possible bug! I haven't been able to do any other work since purchasing it...so other than derailing my workday, it's been wonderful!

Anyway, I've been experimenting with bringing in my own samples and I've noticed that the pitch can shift at specific points in the waveform for no apparent reason during playback with long samples. I've prepared a patch that demonstrates the effect, although you'll need some kind of pitched long drone sample (a simple generated wave will work) to hear it. It's not evident in the included waveforms that I've noticed, with the exception of "sine cloud ji7"

Anyway, I suppose it could be something in the samples I'm creating or my system but I thought I'd share this with you in case it's something real.

The test patch:

<Kaivo pluginVersion="65536" presetName="Alchemical 9/simple wave" scaleName="12-equal" key_mod="2" key_bend="2" key_glide="0" seq_trig="1" seq_rate="0.26600000262260437012" seq_pw="62" seq_value_delay="8" seq_value0="0.5" seq_value1="0.31000000238418579102" seq_value2="0.14999999105930328369" seq_value3="0.039999999105930328369" seq_value5="0.039999999105930328369" seq_value6="0.14999999105930328369" seq_value7="0.31000000238418579102" seq_value8="0.5" seq_value9="0.68999999761581420898" seq_value10="0.84999996423721313477" seq_value11="0.95999997854232788086" seq_value12="1" seq_value13="0.95999997854232788086" seq_value14="0.84999996423721313477" seq_value15="0.68999999761581420898" seq_pulse0="1" seq_pulse8="1" lfo_offset_x="-0.040000021457672119141" lfo_offset_y="-2.2351741790771484375e-08" lfo_rate="0" lfo_level="0" lfo_rate_p="-0.2400000840425491333" lfo_level_p="0.93999993801116943359" noise_width="2" noise_falloff="0.0099999997764825820923" noise_offset="-0.010000010952353477478" noise_rate="10000" noise_level="0.45999997854232788086" noise_auto="1" noise_filter="1" noise_offset_p="0.99999994039535522461" env1_decay="0.0099999997764825820923" env1_sustain="0" env1_release="0.0099999997764825820923" env2_repeat="0" env2_sustain="1" gran_position_y="0.5" gran_rate="0" gran_density="4" gran_pitch_env="2.98023223876953125e-08" gran_pan="-2.2351741790771484375e-08" gran_follow="1" gran_smooth="1" gran_wrap="1" gran_position_x_p="0.27999997138977050781" gran_position_y_p="0.99999994039535522461" gran_rate_p="0.99999994039535522461" gran_pitch_p="0.99999994039535522461" gran_pitch_env_p="0.99999994039535522461" gate_level="0.45999997854232788086" gate_decay="0.87999999523162841797" gate_level_p="0.23999999463558197021" res_position="0.45999997854232788086" res_brightness="0" res_pitch="220" res_nonlin="0" res_sustain="0" res_pan="-2.2351741790771484375e-08" res_out_pos="0.059999998658895492554" res_lambda="0.93999999761581420898" res_length="277" res_bridge_height="1.6000000238418579102" res_bridge_len="0.05300000309944152832" res_bridge_amount="0.019999999552965164185" res_mode_num="1" res_position_p="0.55999994277954101562" res_brightness_p="0.79999995231628417969" res_pitch_p="0.99999988079071044922" res_nonlin_p="0.4999999701976776123" res_sustain_p="0.2199999690055847168" res_pan_p="0.1099999845027923584" res_mix_wet="0" body_position_x="0.48999997973442077637" body_nonlin="0.099999994039535522461" body_tone="0" body_pitch="-2.2351741790771484375e-08" body_sustain="0" body_pitch_mode="2" body_mode_num="4" body_height="4" body_position_x_p="0.5" body_position_y_p="0.5" body_pitch_p="0.99999994039535522461" body_mix_wet="0" body_mix_dry="0.48999997973442077637" output_tilt="4.470348358154296875e-08" patcher_input_2="00000000000000000000000010000000000000000000" editor_x="0" editor_y="0" editor_width="981" editor_height="531" editor_num="1" editor_anim="1" lfo_function="circle" res_mode="metal string" body_mode="wooden box large" sample_name="Synthetic/sine cloud, ji7.wav"/>

Hi Roland,

Always a pleasure! Thanks for the report. I can't hear anything wrong in this patch you posted. But maybe that's because the sample is a shifting cloud of different pitches.

I've gotten another report about something similar going on with the granulator. I'll make some of my own longer sounds to test with and see what I can figure out.

Sounds like the issues I'm having. Try sampling an 8 seconds sawwave with constanst pitch, and then play in low octaves, pitching down an octave or two from base pitch, and then move the playback position around. Also, it only seem to happen on certain notes. A low C is a good note to reproduce it for me, low E doesn't seem to change pitch.

(OSX VST in Live)

Well. It's been well over a year. Still no word on this. I've made a small video showing the phenomenon. The further to the right in the sample the granulator reads, the worse the pitch resolution. The sample used is a 8 seconds long sawtooth of constant pitch.

https://www.dropbox.com/s/gkm9b9fnfm234ge/KaivoStrangePitch.mov?dl=0

Sorry for the rather bad audio quality, recorded from speakers to laptop microphone...

This makes Kaivo pretty much useless for the kinds of sound I was planning to use it for (morphing wavetable kind of sounds using 4 channel wavefiles, with distinct pitch), so it's more or less sat unused since I bought it at full price. Not too happy about this...

Hi @rydan,

Sorry I dropped the ball on this. It looks like I missed your first message here entirely and I didn't have this issue on my list of Kaivo bugs. There's only one of me and I'm definitely capable of slipping up sometimes, so it doesn't hurt to remind me if you're waiting on an answer. I

With the help of your video I was able to reproduce what you are hearing. I should be able to fix this for the next update. Meanwhile if you use shorter files (

Sorry, I was in a slightly cranky mode already, and then remembered I hadn't used kaivo in quite a long time and ran into the bug again. =)

Anyway, good news if you're able to reproduce. A fix would be great!

Wow, looks like my post got mangled somehow.

I was trying to say, if you use shorter files, you should be able to make a 2D wavetable with accurate pitch. With anything shorter than a second I was not able to hear pitch errors, so a stack of 4x128 waveforms should be in tune just fine and offer lots of wavetable goodness.