Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Frequency bleed? #39

Open
unfa opened this issue Feb 6, 2019 · 1 comment
Open

Frequency bleed? #39

unfa opened this issue Feb 6, 2019 · 1 comment
Labels
enhancement 💎 Refinement of existing feature needs research 🔬 Further research is needed to fix this issue

Comments

@unfa
Copy link

unfa commented Feb 6, 2019

Here's comparison of how Log scale looks with different window sizes:

Using 2048:
image

Using 4096:
image

Using 8192:
image

Using 16384:
image

It's a bit strange, but only the highest window gives me clear view on the low frequencies.

I wonder if this could be improved somehow? I don't remember something like this form using say - Sonic Visualiser. I need to do a strict comparison though.
This makes me want to ask for even bigger windows sizes, since the time resolution is still pretty ok at 16k.

@pdesaulniers
Copy link
Member

Perhaps the rendering could be improved so that the low end would look clearer. I think the issue is mostly with the current interpolation code, which makes the whole spectrogram look a bit "spread-out" overall.

Also, @fundamental mentioned that a log-spaced DTFT might help for getting a smoother visualization on the low freq end, and pointed me to this implementation: https://github.com/original-male/non/blob/5ae43bb27c42387052a73e5ffc5d33efb9d946a9/mixer/src/SpectrumView.C#L93-L176 .

Maybe something like an adaptive spectrogram could be used too. See vamp plugin as reference https://vamp-plugins.org/plugin-doc/qm-vamp-plugins.html#qm-adaptivespectrogram

@pdesaulniers pdesaulniers added new feature ➕ New feature or request enhancement 💎 Refinement of existing feature needs research 🔬 Further research is needed to fix this issue and removed new feature ➕ New feature or request labels Apr 18, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement 💎 Refinement of existing feature needs research 🔬 Further research is needed to fix this issue
Projects
None yet
Development

No branches or pull requests

2 participants