Quote:
Originally Posted by Lord Dredd
(Post 1915175)
yes and I gotta say I found to be most vibrant , the first three sounded alright the was also great but the other one had beefed up audio volumes so to me that sounded well.
I am not an expert and and I listened to them on my desktop speaker which is okish
|
Two of the samples were just a straight downmix (no compression) and there's one with very mild compression via Avisynth. The main difference between the two DAN samples is the speed at which it reacts to volume changes. My settings (f=150 b=1) react quite quickly, whereas the other settings (f=2000 g=23 m=15 b=1) are much slower. Both have advantages and disadvantages.
The DAN uses a kind of "window" for determining and adjusting the volume. If you compare the two samples, my settings keep the speech at much the same level throughout and the loud parts shouldn't be any louder than for a normal downmix.
For the other sample, the speech is louder at the beginning, but there's a section just before the first explosion where the volume of the speech starts to fade away, then the first explosion is probably a little louder than it was originally. I'm pretty sure that's due to the larger "window" it uses to determine volume.
The disadvantage of it reacting quickly is it can cause "volume pumping" on occasion....
Imagine a section of audio that's mostly normal speech with some light traffic noise in the background. In between people talking, where there's mostly silence, the volume of the background traffic can increase noticeably, then drop back down to a quieter level while someone talks, then get louder again where there's silence. That's the type of "volume pumping" I'm referring to.
If you use a TV with it's audio set to "night mode" there's a good chance you'll hear volume pumping for days, or just as bad, it'll sound like someone turned the volume down during a loud part, and they're slowly turning it up again while it's quiet.
Anyway.... the settings I use are what I found to give good compression, without noticeable volume pumping 99% of the time, and without being able to hear the volume being slowly adjusted to prevent volume pumping.... but it's all personal preference.
Quote:
Originally Posted by Lord Dredd
(Post 1915175)
strangely similar avisynth for a x264 video works well but when used with x265 cools the color off..
any tick you might wanna suggest to keep the colors as close to trhe original as possible.
|
I can't think why x264 and x265 should look different when all else is equal, unless it's a playback issue. Someone else may have a better idea as I never use x265.
By the way, I know MeGUI automatically adds frame rate conversion to the script for ffms2, but don't use it unless the source is variable frame rate and you want to convert to a constant frame rate, in which case MeGUI will probably pick the wrong frame rate to convert to anyway.
Even if the specified frame rate is exactly the same as the source frame rate (in which case it should do nothing) it can still, on occasion, cause frames to be dropped and duplicated when they don't need to be and make the encode look "jittery". I'm talking about this:
fpsnum=24000, fpsden=1001
Unless you're converting the frame rate, delete it from the script.
