Forum Replies Created
-
If you are using Warp Stabilizer in Premiere Pro, create a 1080p sequence as per the clip’s properties. Insert the footage into the timeline and apply Warp Stabilizer without scaling.
Insert the 1080p sequence with the stablized footage into your 720p timeline as though it were a clip. For best quality, adjust the (x,y) position and scale% in the 1080p sequence (using the Warp scale control) until it fits properly in the 720p timeline.
-
Ivan Myles
October 11, 2013 at 10:58 am in reply to: from 1-2 hour export times to 10+ hour export, 15 min file.First verify whether GPU acceleration is enabled. Second, did you enable Maximum Bit Depth or Maximum Render Quality in the export settings?
-
[Steven Sall] “I tried those settings – I believe it turned out to be the same quality which is ok. Just not great.”
Were you able to import the encoded file back into Premiere Pro and compare it to the source clips in the edited sequence? The degradation in quality when playing on the TV might be caused by the hardware, not the file. (For example, one of my displays is prone to banding and blocking even though the same videos look fine on other monitors.)
[Steven Sall] “Is it normal to fully render the sequence first before exporting? Does that yield a better quality end result?”
The question could be interpreted a few ways:
– If you are referring to the encoding process, the only times a sequence is fully rendered prior to encoding is when preview files are used (i.e. smart rendering), or an intermediate file is created. Otherwise, frames are rendered and encoded one after another as part of a single process.
– If you are referring to two-pass VBR encoding, then, yes, two-pass typically generates better quality than single pass encoding at a given bitrate. (Alternatively, two-pass generates the same quality as single pass but at a lower bitrate.)
– If you are referring to running test files, then no, it is not always necessary. That said, it probably should be done more often, especially when one is learning about the different encoding parameters.
-
Happy to help.
-
-
Thanks for the info.
Do you only see the artifacts in the encoded file, or also in the Program Monitor prior to export?
What media player(s) did you use to view the encoded file? Do the artifacts appear if you import the encoded file into Premiere pro and view it in the Program Monitor?
Was Maximum Bit Depth enabled in the Export panel? If so, try encoding again with it disabled.
If GPU acceleration is enabled, try again with CPU only.
If the issue still exists apply the Broadcast Colors effect to one or more of the problem clips and set the IRE limits to 0-100 or smaller (e.g. 7.5 to 93). Enable either the “Key Out Safe” or “Key Out Unsafe” options to verify whether there are any illegal colors in the same locations as the artifacts.
-
Glad to hear you are making progress. Try 20 Mbps Target bitrate with 30 Mbps Maximum using High Profile at Level 4.1. There is no advantage to using a higher level in this case (but not much penalty if the TV supports L5.1). Also, set the Level first and then check the other settings; Premiere Pro resets some of the parameters whenever the Level is changed.
Scroll down to the Advanced section in the Video tab and set Key Frame Distance to 20 frames. Click on the Multiplexer tab and check that Multiplexer is set to MP4 and Stream Compatibility is set to Standard.
You could potentially go up to 25 Mbps target bitrate with 40 Mbps maximum. Beyond that should not make much difference with 2-Pass VBR unless there is a lot of motion throughout the video. The video might start stuttering if the bitrate is too high for the TV.
To check the colors, import the encoded file back into Premiere Pro and insert it in a sequence with the edited footage. Set either the Reference Monitor or Program Monitor to All Scopes or RGB Parade. In the timeline panel, toggle the track with the encoded footage on and off. Watch the scopes to see if there is a significant difference in the waveforms between the edited footage and the encoded file. It is normal to see a slight shift, but it should not be more than a few IRE difference on the scale. If you see a major difference please post screen captures of the scopes for a frame of the edited sequence and the same frame of the encoded H.264 file.
-
Thank you for posting your settings. It is very helpful when troubleshooting. The lines you saw were probably because the file was exported as interlaced but the source footage is progressive. For clarification, why is the sequence 119fps if the source material is 120fps?
What are the TV’s requirements for video files? Things to look for include supported codecs (H.264/MP4 vs H.264/MOV vs other), maximum Profile and Level, maximum bitrate, and maximum frame rate at 720p and 1080p.
Pending further info regarding the TV specs, I would recommend using 1280×720-60p and letting the TV up-convert on the fly rather than exporting at 1080p. Also, a higher bitrate will yield better quality. Try making a short test file at different rates from 10-20 Mbps to see what the TV/USB can play.
-
– What effects were applied to the problem clips?
– Have you checked the YC Waveform scope to see if there are any values greater than 100 IRE? -
[David Edwards] ” In that case what would be your preferred workflow to add audio, do you re-import the image sequence, layer audio and then recompress to delivery codec?”
Essentially, yes. Use ProRes instead of DPX if storage space is constrained. When the video is not being re-scaled I send jobs to the AME queue.
[David Edwards] “I guess even with an image sequence, this seems to me like the same issue as I’d have with any codec, the need to import, edit and re-render a timeline? (short of the “smart render” and “use previews”.)”
DPX is already expanded so you only have the encoding time. ProRes needs to be expanded and then compressed. Both options will still be faster than rendering and encoding together if you have more than one file to encode.
