r4329 was a minor change to how video capturing is handled (duplicate frames are not written out which reduces cpu usage), it would make no difference if you're not recording the gameplay.

Hi jmarsh,

I edited my last sentence, since my message was unclear.

All releases until 4329 (including 4329) were fine. From 4329 I updated to 4356, and 4356 was not fine.

So, the problem appeared in a to be determined release between 4329 and 4356.

It may be an issue specific to the particular game, so which one is it?

Hi ripsaw8080,

I'm sorry I've started this new post unnecessarily. When I downloaded r4356, I rewrote dosbox.conf to my liking, from scratch. When I did it, I left an inconsistency: I changed [sblaster] oplrate to 49716, but inadvertently left [mixer] rate at 44100. When I changed [mixer] rate to 49716 as well, the problem was gone.

Answering your question, I was noticing the problem in two games I was playing: Day of the Tentacle, and Indiana Jones: The Fate of Atlantis.

Differing rates causes internal resampling, but the freezes and buzzing sounds you mentioned seem more severe than expected results. What cycles setting are you using for the games? What OS and processor are you using to run DOSBox SVN?

I am using "cycles = fixed 5000". Processor Intel Core i7-7700HQ @ 2.80 GHz. Microsoft Windows 10 Pro Version 2004.

I am really at a loss—I tried today to recreate the problem by recreating the mismatch between oplrate and rate, but alas the problem did not occur. Your answer, regarding the problem severeness, convinced me that the problem was not DOSBox related. It was probably some other issue in my computer that I can't figure out yet. Presumably some other process was using too much of the CPU in the background, without my awareness. In any case, if the problem occurs again, I will report it here.