Currently Being ModeratedApr 20, 2012 7:02 AM (in response to Russ H)
the original footage is AVCHD which was ingested and 'Optimized' in FCPX (ProRes). Below is the Inspector window contents of Compressor's SD DVD preset Summary that I was using when the issue arised.
I exported the Timeline (Current Settings, basically unchanged from the original timeline), using the 'Export Media' option in FCPX share menu, then I bring in the Exported .mov file to Compressor.
Name: MPEG-2 for DVD
Description: MPEG-2 elementary stream for DVD Authoring
File Extension: m2v
Estimated size: 3.36 GB
Type: MPEG-2 video elementary stream
Width and Height: Automatic
Selected: 720 x 576
Pixel aspect ratio: PAL CCIR 601 (16:9)
Frame rate: (100% of source)
Frame Controls Automatically selected:
Retiming: (Fast) Nearest Frame
Resize Filter: Linear Filter
Deinterlace Filter: Fast (Line Averaging)
Adaptive Details: On
Detail Level: 0
Field Output: Same as Source
Start timecode from source
Aspect ratio: Automatic
Field dominance: Automatic:
Selected Top first
Average bit rate automatic,
selected: 7.826087 (Mbps)
1 Pass VBR enabled
Maximum bit rate: 9 (Mbps)
Best motion estimation
Closed GOP Size: 12, Structure: IBBP
DVD Studio Pro meta-data enabled
Currently Being ModeratedApr 20, 2012 9:29 AM (in response to StingRay67)
I wouldn't be surprised if the difference in what you're seeing from v4 compared to v3 has more to do with your material (lighting, background, movementt) than it does with Compressor. However, it would be most interesting for you to make that comparison yourself – doing the job in whatever earlier Compressor version you still have access to.
BTW, whart flavor ProRes did you use?
But a couple of things: Was the footage shot progressive or interlaced? Definitely don't de-interlace if progressive and probably don't even if it were shot interlaced. If you feel you must de-interlace, use Frame Controls and select Motion Compensated; don't use the de-interlace option in the filters tab.
I'm not a fan of the new DVD video quality settings (and I use 3.5 for DVD encodes)…and I recommend changing to two pass VBR – unless your movie is fairly static, like an interview. Furthermore, the new bit rate settings are very aggressive, and I'd lower them so the average is around 7.2 Mbs.
Try encoding a small but representative section with a variety of adjustments and bring them into DVD Studio pro to test…waste a disk or two. If you don't get satisfactory results, break it into two jobs…in the first, do a re-size to SD – while staying in Pro Res – with Frame Controls on and resize filter set to Best. Then take that outut file and make your MPEG-2.
Currently Being ModeratedApr 21, 2012 5:04 AM (in response to Russ H)
Hi Russ, thanks for your input here. As far as lighting goes, yes its just my sons school play and lighting wasnt great, but also not that poor, a home project not a commercial one. However it's not the set lighting that is the issue here. There is way too much artefacting on horizontals and aspects like that, which points to a compression-at-post issue.
I was sure i used ProRes 422, but now you've have me thinking as this was a big project, I may have lowered the quality. How can one check if its Proxy or Standard ProRes used in the Timeline?
Also, if I have ingested using proxy, is there a way to Media-Manage (like FCP7) whereby I can just reencode the footage on the timeline?
Currently Being ModeratedApr 21, 2012 5:23 AM (in response to StingRay67)
So try the steps with small test files that I suggested in my earlier post.
You don't have to use the preset. Use it as a starting point. Experiment with adjusting the quality controls and frame controls. When you find something you like, save it as a custom preset.
Let us know how much improvement you are are able to achieve.
Currently Being ModeratedApr 21, 2012 2:39 PM (in response to Russ H)
To my amazement I noticed the average bit rate was around 5.0 with a max set at 9 at a 1 pass VBR! In my experience the average bit rate is the critical value here. I'm absolutely sure this was set much higher in previous versions of Compressor. Anyhow, I'd set the values to average around 7.7 with a max of 9.0 (highest) and used 2-Pass VBR.
Result: it worked great!
Thanks Russ for pointing out what should have been obvious to me in the first place!