First impressions of #6800, definite progress

[Mits]

Merry Xmas to all, and developers, do not read this immediately, you must not work during holidays!

I just wanted to report that the Classic Spectrum Analyzer in the new December build #6800 (clean portable installation, not upgrade) still exhibits the same problems as in my older post (https://getwacup.com/community/index.php/topic,830.msg5311.html)

Namely, when the output plugin is Not So YASAPI v1.2.3, the classic Spectrum Analyzer consistently does not start after the first track played after launching Wacup. It may or may not start after scrolling a bit on the track or jumping to the next one. Sometimes the plugin starts by itself when advancing to the next tracks of the playlist (but not always). In addition, the display is not synchronized to the music and leads (not lags) about 250-300 ms, very obvious when playing music with slow beats. My personal impression is that this is not a simple bug but an algorithm problem in the data flow during the process of decoding and buffering to the WASAPI output, so it needs a lot of work. No biggy, I Will just sit quietly in my corner and wait patiently :-)

The other output plugins, Not So Direct, Not So Wavey, Nullsoft Directsound, Nullsoft WaveOut do not exhibit these problems. However, the playback volume varies wildly across the different output plugins. This is not technically a bug, since nobody changes the output method after choosing what's best for them, but it may be useful info for the devs when troubleshooting volume problems.

Other minor bugs, like the non-sticking of the checkbox in [Preferences - General - Taskbar - Taskbar Text - Scroll the text in the taskbar] or the non-sticking of the toggle to display the remaining (and not the elapsed) time in the main window Time Display have been blissfully fixed.

Finally, playback seems now to have higher execution priority and stutters much less when the filesystem is busy. My impression is that the buffering priority or its default values could be a bit more aggressive, but we are certainly in a good track!

Thank you for all your hard work, the length of the bugfix list is record breaking and utterly impressive.

[dro]

you must not work during holidays!

I failed on that as I'm stuck at home & have been doing things to prepare for an update for #6800 that fixes the main stability issues it's having.

I just wanted to report that the Classic Spectrum Analyzer in the new December build #6800 (clean portable installation, not upgrade) still exhibits the same problems as in my older post (https://getwacup.com/community/index.php/topic,830.msg5311.html)

That's expected still as I've yet to look into it.

The other output plugins, Not So Direct, Not So Wavey, Nullsoft Directsound, Nullsoft WaveOut do not exhibit these problems. However, the playback volume varies wildly across the different output plugins. This is not technically a bug, since nobody changes the output method after choosing what's best for them, but it may be useful info for the devs when troubleshooting volume problems.

It shouldn't really vary when comparing maximum volume whatever the output type used though what is done with the levels between 0 & 100% can vary so that might be more down to what's causing a perceived difference at times.

Other minor bugs, like the non-sticking of the checkbox in [Preferences - General - Taskbar - Taskbar Text - Scroll the text in the taskbar] or the non-sticking of the toggle to display the remaining (and not the elapsed) time in the main window Time Display have been blissfully fixed.

That's good to know

Finally, playback seems now to have higher execution priority and stutters much less when the filesystem is busy. My impression is that the buffering priority or its default values could be a bit more aggressive, but we are certainly in a good track!

I've not intentionally made any such changes but it might be down to trying to reduce what's going on on general. Also the skin type being used can have an effect on how much possible contention might be going on.

Thank you for all your hard work, the length of the bugfix list is record breaking and utterly impressive.

As long as I'm moving things in the right direction then that's all I can ask for

-dro

[dro]

when the output plugin is Not So YASAPI v1.2.3, the classic Spectrum Analyzer consistently does not start after the first track played after launching Wacup. It may or may not start after scrolling a bit on the track or jumping to the next one. Sometimes the plugin starts by itself when advancing to the next tracks of the playlist (but not always). In addition, the display is not synchronized to the music and leads (not lags) about 250-300 ms, very obvious when playing music with slow beats. My personal impression is that this is not a simple bug but an algorithm problem in the data flow during the process of decoding and buffering to the WASAPI output, so it needs a lot of work. No biggy, I Will just sit quietly in my corner and wait patiently :-)

I finally had a look into this last night & I've got it working reliably in my test installs. The lag comment was helpful as that was the cause of the issue in that the plug-in was reporting a latency back to the core which is sort of correct but no other output plug-in tends to do it & instead just returns 0 as an effective success state when the output is being opened for use. So this is one of those fun less code fixes things solution

-dro

[Mits]

I am glad to report that build #7170 fixed the visualization problems when using the Not So Yasapi plugin after a fresh portable install, using default settings. It's a dream come true, my gratitude to all devs and testers!

[dro]

Thanks for the additional confirmation the fix worked.

-dro