CS5 Premier Pro Export to MPEG2

[Bob Dix]

Hi Tim,
Even an Australian can appreciate that explanation, thanks for your time.

Thanks,
Bob

Freelance Imaging & Video
AUSTRALIA

[Bob Dix]

Hi John,

I agree.

Bob

Freelance Imaging & Video
AUSTRALIA

[Tim Thompson]

TimK:

Thanks for the input.

Wisconsin? I grew up in Northern Minnesota where the men are men and the women are too! I live in Tn so the Titans are my adopted football team but I am a fair weather fan. Only in the playoffs. I played hockey from when I was a kid all through high school. I managed to keep all my teeth though. I am more likely to go to a predators game than the Titans.

Thanks again for the info on the video. I’ll keep you posted on how I do today.

Later

Tim

[Tim Thompson]

OK, I’m getting ready to export. I’m on CS5, Premier Pro. Does anyone know what GOP settings are? The options are M frames and N frames.

Thanks

Tim

[Tony Connoly]

How about using Encore. Would that help?

[Tim Kolb]

I don’t know that messing around with the frame designations will amount to much…

Have you tried leaving the stock settings alone and running it with “max render quality” checked?

TimK,
Director, Consultant
Kolb Productions,

[Kris Koster]

I note that one of the user members above stated NOT to use the Premiere Pro CS5 ‘use previews’ option when exporting to final output because the quality is not the same as using export from the master files.

I’ve just attended the BVE expo show in London this week. The Premiere Pro expert that was giving live demonstrations of the Adobe product said there was no reason not to use the ‘use preview’ option when exporting the final render. That the quality is the same as exporting without the previews.

I’m not getting it. Could someone please clarify which is the case? I cannot seem to find any info on this particular point from Adobe documentation.

Kris

[Tim Kolb]

Hmmm… Without knowing the context of the response of the “Adobe Expert”, it’s hard to know what they meant exactly…

Depending on what your previews are set to be and your camera material is, there would certainly be scenarios when the differences in quality would be relatively minor and the speed gain might be a worthy tradeoff…

However the statement that encoding from rendered previews vs using the original material in a direct encode to the output format have the same results is not technically defensible as a general statement…

The preview render format options that are utilized most frequently are intentionally lossy and relatively small data rates…MPEG I-frame, DVCProHD…these are great for improving playback performance if you are editing uncompressed or some heavy datarate footage, but obviously the quality of the image is not as good in the preview…if you are editing 10 bit uncompressed footage and you have a 8 bit MPEG I-frame preview set up for the sequence, it will make the edit process and the system response much faster and help you edit, but you’ve obviously taken a huge quality hit to create the preview. Using that as your “program master” would obviously cost you image quality. Particularly when you consider that the effects were created on the lossy preview clip, and that’s as good as they’ll get…

In the case of using DVCProHD as the preview format…it stores your uncompressed 1280×720 footage as 960×720 (or your 1920×1080 footage as 1440×1080 or 1280×1080), which helps speed of playback, but can hardly be classified as being a net wash in the image quality department).

However, if you are shooting DVCProHD (and the camera stores the 1280×720 as 960×720 to begin with), and preview-rendering DVCProHD, you could probably make the case that the loss is minimal, but you have still introduced a re-encode step in between the original footage with queued effects, etc and your final output.

On a Mac, if you were editing ProRes and you set the preview file format to the same ProRes profile you were editing (if you have FCS installed so you have the ability to write ProRes files), you would introduce a rather minor amount of re-encoding into the process…so you might be able to make the case that the speed up in the output encode is worth adding that minor encode pass depending on how many effects you have in the material.

Whether you, the “Adobe Expert”, your client or the guy down the block can visually tell the difference between the two encode methods is a different discussion entirely and will differ with the camera format, the content of the footage, the amount of effects work done, and of course the chosen preview format.

In general, encoding from the preview format is faster because there is only one, or a handful of contiguous files to access, and the effects do not have to be re-executed as a component of the output encode.

However, those preview files are not necessarily at the same image quality level as your footage. Particularly when you are using the preview files to “pre-render” effects, understanding how all these factors relate to each other is how you can best decide if the loss inherent in using the preview files as the input for your encode is worth the speed gain you achieve…

But in a general, theoretical sense…using the previews is not the same as using the original footage.

In a visual sense the differences may be slight and acceptable… Or, I know there are certain third party effects plugins that simply don’t operate correctly when you try to re-execute them in a normal encode, so “pre-rendering” those effects into a preview file is the only way to use them in the final video.

I advocate that people test it for themselves as often the preview will appear to be as good as the original, but launching a very lossy output encode for DVD, Blu Ray, or the web on that preview file vs the original footage will often reveal just how much of a compromise was made…and it can be a little surprising.

Each user needs to determine their own best course…

TimK,
Director, Consultant
Kolb Productions,

[Tim Kohn]

This has been an interesting read (the entire thread), so coming into the discussion late in the game. I had posted a related question on the Adobe CS4 forum regarding encoding quality. My specific question was regarding an SD encoding project, but have experienced the same issues with HD encoding to SD. The issue/problem I am having is encoding a high quality DVD video from and source SD video (that is also high quality).

As a bit of background, I’ve got two XHA1’s, and recently added a 7D dslr into the mix, so shooting HD footage and downsampling to SD has always been a challenge with the adobe products. I’m currently using cs4, but have also been a long user of premiere iterations… 6.x, 1.5x, 2.x, 3.x, and now 4.x.

I have noticed that with the cs4.x products using the maximum render quality has a BIG impact on quality when downsampling, at the expense of significantly longer render times, 10x-20x, depending on other timeline effects. However, I am still battling the quality issue even with Maxiumn Render Quality setting enabled (and using high bit rate settings, generally around 7).

To add to my general confusion/frustration, I recently shot an SD event with the XHA1 in SD mode. The resulting AVI (NTSC 16:9) looks great, both viewed in Windows Media Player, and also in the Premiere Timeline. When I export to MPEG2-DVD however the quality is poor, not nearly as good as the source. This really surprised me as I was not expecting to have any quality issues, as no scaling was involved, and also no video effects, just a simple timeline cut project, and then export to MPEG2-DVD for encore to authoring. Note that I have also tried to encode in Encore with the same results.

So this got me thinking a bit, recently I have been experimenting with workflows based on h264 (for HD authoring, not SD), so wondered what the video would look like using a encoding workflow that was not based on an Adobe’s solutions (AME, Premiere/Encore). To verify that my source material was “good”, I first exported a new AVI file from Premiere so that I would have an edited version to work with (I verified after exporting there was no visible difference between my source and new AVI).

I then used Handbrake (free open source h264 encoder) which I had recently discovered (found out about on a different forum), and exported a new h264 mpeg4 file (same resolution, SD 16:9 854×480) and the video looks great (smaller file size than the mpeg2 file as well)! I then compared the encoded DVD version (created using Premiere/Encore) to the h264 version, created using Handbrake (I was able to view the h264 file using Windows Media Center with the XBOX360 being the extender to view the H264 file on my PC) on the same TV (60″-sony HD). The DVD version looked fuzzy, with ghosting, while the h264 version looked extremely good (like HD quality good!), it was very impressive!

So this whole test gets my back to my original question…. why does the adobe MPEG2 version look so bad (relatively speaking), compared to the handbrake H264 version. The H264 workflow does not solve my immediate problem, and that is getting a HIGH quality MPEG2 based DVD to the client. What it did tell me is that the Adobe solution of creating and MPEG2 video file has problems. I like using PPRO for editing, and Encore for authoring, so would like to learn if there are other encoders (free, or commerically available) that others are using that are proven to be better than the Adobe MPEG2 encoding solution.
-tim