WMV encoding quality has changed?

[JOHN MONDO]

Has anybody else noticed that the WMV encoding quality has fallen off badly in the 2015 version of CC? I'm not sure exactly what version

I can go back to encodes done last year using an identical template and we go from "reasonable" quality to banding and pixelation. Nothing changed on my end. Yes its a fairly low bit rate 720P @ 1Mbs 2 pass VBR but i was getting comparable quality to H264 based MP4s.

Whats more if I do the same encode in Squeeze it's also markedly better. I don't want to lodge a bug until I can get confirmation from other users but it is like this on all machines.

I've included a couple of screen grabs - the actual data rate is around 1.2Mbs including audio. The quality of the source video has always been quite noisy - that hasn't changed either.

[thnord]

What previous version are you comparing your results against?

[JOHN MONDO]

Sorry - all I can tell you is it seemed to happen between 27th June and 18th July this year. I don't track exact version updates anywhere. I also can't tell you if it corresponds between moving from W7Pro to W10 pro..

It also seems the files are coming out slightly smaller, despite the fact that the templates are identical.

Audio on both these files was 44.1 Stereo @ 128Kbs

Video in July 63min 32 sec and 448MB which according to my calculations makes it about 963Kb/s including audio

Video in June 60min 42 sec and 493MB makes it about 111Kb/s including audio

So neither files are 100% correct at 1128Kbs (1000 for video + 128 for audio) but the earlier file seems to be a lot closer.

Still we are only talking a difference of 48Kb/s but i suppose it is feasible that it made that it made that much difference?

I've always had it on 2 pass vbr, bitrate is set 1000Kbs with a maximum of 5000Kbs. I never use maximum render quality as I find the results to take longer and generally introduce additional artefacts (especially when rendering MPEG2 - I actually get frame problems).

[thnord]

Well there's going to be a certain amount of variation to the actual bitrate due to it being variable so I'm not too worried that it isn't coming in at exactly 1128kbps. What does concern me is that you are clearly seeing a drop in the quality. When I go back and encode the same source across multiple versions I get basically the same quality which leads me to a possible source of the issue: I've seen cases in the past where the source file was the determining factor in whether a certain quality bug appeared or not. Would it be possible for you to re-encode the example above that is good using a more recent version of AME? If that one comes out clean then my suspicion may be correct (but it would still be a bug). If it comes out bad then I'd like you to re-encode it in an older version (maybe CC 2014) to see if it is any better.

Something goofy is going on here. It will just take a bit of trial and error to figure out what.