[Ht Davis]

Is this what you’re saying:

Capture all clips in a single timeline (captured video) to a single location (xsan or storage location; set in the project prefs), and then archive all offline (using the offline transcoding) as a single sequence transcode (thereby making a BATCH VIDEO, in one file, of all of the clips)?

For strictly theoretical purposes, great work.

A single project can be used like a database with a single sequence or many, depending on how you want it to work. If you set a single sequence up and place every video inside it, render out a preview of the clips in PROXY size (will give you a large set of files that isn’t searchable or usable on their own), you can actually speed up the output of the proxies from the initial render of the work area. You can then use the exporter to quickly (using Previews) render a full proxy to a useable file for any other work. If you make them all sub clips and tag them, you can search, render, and even add more very easily. I’m going to look into this workflow. It has potential. The only caveat I can see is the settings that denote where the previews are kept. I’ve seen two settings that affect this. One in the preferences of the app, that allows you to select where previews are kept in general, and whether or not they are kept next to the originals; and also in the project setup itself. It’s difficult to tell which would take over under which circumstances. For instance, I could keep my project file in one place (it is small enough to be on my main drive) while the files are all kept together on XSAN, I can have the project on the xsan in the same folder as the files and have both backed up recursively (several different backups). With a Raid 5 type system, a rebuild could be done at any time (but would take a while), and with both situations, there would still be a backup otherwise. The only problem would be whether or not the rendered previews would be detected and if not, how long it would take to re-render them (compression takes a while). If you split the clips across sequences with date ranges, however, this becomes less of a problem, and you can still bin a bunch of sub clips with tags for easy search and export; you just check the source of the sub clip to see if you have any rendered previews, if not, you know it will take a while to render out.

A lot of possibilities there.

[Ht Davis]

What do you mean, “Not fun”? It’s the way this type of edit was meant to be done. It’s usually not normal to move markers in such a fashion with old-skool edits, it’s just coming into the workflow. By adding the markers to the layer, it can be added non-destructively, and it can be edited the same way, while being reverted with a simple recopy of the original layers. Perfect for this kind of edit, and you can have the new sequences carry different frame rates. Bravo on the fix…
I’ve got a DEFAULT project that I saved with several adjustments etc in bins, I’m going to add a MARKERS bin with a default transparent layer for placing markers on, and then just copy that to each original sequence’s bin, so I can always have one to work with in that fashion. Rendering out the project should still yield the proper output to encore, and to a video file that allows for markers. You’re awesome. THX again.

[Ht Davis]

Any change in frame rate should affect the markers, as they conform to the frame number. Sorry, I don’t think they will retain Second or millisecond count in premiere. However, if you can place in a program that will allow this change (in milliseconds perhaps), you should be able to make the transition just fine. I haven’t used AE much, but if you could simply move the sequence there, and then into a milliseconds comp that goes into a new-framerate comp, then have that import as a new sequence with all of it’s markers, it’s possible you could keep your markers timed to seconds of video rather than frames.

Process:
Place duped sequence in AE
Create new comp with milliseconds timebase
import AE comp into Project file
Create a new sequence with your frame rate, etc
replace the sequece with AE composition
if this last doesn’t conform the Compostition to the frame rate, try nesting the comp in a framerate comp.

You can do this even easier in FCP, as it will allow you to go to the subframe or milliseconds level.

I tend to “OVERUSE” my edit marks, so I rarely have to rely on workarounds to conform my output. If I want the same video in a different frame rate, I simply Dupe sequence, delete everything but markers, change frame rates etc in sequence settings, then re-add my data, but don’t allow sequence to change to match. Then I drag markers into proper placement with the edit marks, which will still conform to frames of the original source, but retimed to what the new Rendered frames will be, so the markers will match the rendered video just fine. With snap on, I even have the indicator where the markers need to align. It’s not a workaround. It’s a workflow.
If you are using different video entirely and it was shot in a different frame rate, you’ll want to use FCP or even AE to do this kind of edit. OR: You can paste your original video in, move the markers where they need to align using the edit marks, then delete all the video again, and replace it with your new video. It’s a bit roundabout, but it still works… Put in the time, get the output.

[Ht Davis]

I’ve found the above issue to be heavily related to cameras using image stabilization in camera. It adjusts frame rates to “steady” the image.

I have used adobe to fix it… …AME will fix this problem, but will slow the motion down a bit, almost slowmo in areas where there is heavier drop, but I’ve found that even the speech being watched matches up alright. You just have to select “USE FRAME BLENDING”. You can do the same in AE, but AME is faster. I outputted Prores files with and without. Without, I had the same issue, with the blending, it was fixed. I then dropped the output frame rate… …nearly destroyed the motion quality. I did the same with AE, but it took 12 times as long to process; I used twixtor afterward and had it render without the blend… …Twixtor will create the frames in-Comp, but to render them out will take even longer, as the quality is much higher. I stopped this render when it was taking too long. I checked the quality of my other tests, and they were acceptable for the purpose of my video, so I rendered out fully using AME to output a prores LT422 profile and frame blending to force a constant frame rate.

[Ht Davis]

It sounds like the camera has an image stabilization algorithm that alters the frame rate.

Process the video out to an intermediate first (doesn’t matter which one).
When you are setting that up, just remember the following:
You need to specify the exact frame rate–no Automatic from source, as this will allow frame rate changes, resulting in a VFR instead of CFR (CFR is constant frame rate).
You need to set the size, and the field order, as some algorithms actually change these as well–I shoot progressive, so I set progressive, then output to interlaced only when I cannot output to PSF (a fake interlace).
If you do this correctly, your AV will be synced in premiere. But there’s more adjustment you can make:

decide how much headroom you want to keep for your audio and how “real” you want the variation to be. Headroom will be your KHZ (sample rate) which will give you some background noise and fill to play with if you go with 96khz or higher, will give you warm and true audio in 48khz, and will squeeze into 44.1khz to limit file size. Your bit width will determine how “real” the differences can get (how deep the decimals go and how big or small the numbers get in general. 16bit is CD quality, but usually isn’t for any kind of professional video\audio combination, unless you can engineer the sound in an advanced manner.

I used iMovie, imported the clips, sent them to compressor, and transcoded out to prores, which decompressed and created a proxy for me to use (2 transcodes, but that way I can load the first, place it, and then relink to the proxy for work.

Differences in editors:
Premiere doesn’t go down to the subframe level. Most editors today do (especially final cut), but do so by creating an intermediate frame between them. Premiere doesn’t create intermediate frames unless you apply some sort of motion effect or cleaning, and retime the frames.
Premiere is a professional oriented program. Most pros work in one standard frame rate that doesn’t change. Some consumer devices, however, vary frame rates with motion, in order to make the output look better or keep the camera shake to a minimum. Most editors enable this to work fine, and will create the intermediate frames if you tell them to, or when you output your final product, but Premiere only edits a standard video stream, it doesn’t “Create” the video frames. Rather, it orders the effects and frames you put together in a timeline that builds to a single frame rate video output. Premiere is more a LINEAR EDITOR than a NONLINEAR EDITOR. After-effects should be able to do the same as compressor, and do a better job than MEDIA ENCODER, as you can set more strict output profiling in your COMP, so that it forces the frame rate and the size.
I chose compressor because I could run it on more than one machine and output a video directly (farming), and it doesn’t just output images from the farm.

[Ht Davis]

Streamclip is good. And the reason the file is bigger is because you’re using different compression method, resulting in more data being stored in the file… Just an FYI.

My workaround is to use AE, drop the video into a COMP with strict settings and render out to a MOV file or another intermediate file type (prores, intra, etc). Just make sure you match your sequence settings with the comp from AE.

[Ht Davis]

If you want to use the proxy preset you downloaded…
It has to be a preset Premiere can see as a preview preset.
If you don’t want to drop resolution of the main file, or play with “Fit TO FRAME”, you could use the previews.
Close premiere.
On mac, open the .app with a right click and “Open package contents”, you’ll see a window open with the folder “Contents”, in this search around until you see the Sequences or Presets (or both) folder, and find the folder inside that one that’s called “Previews”, and drop the .epr file you downloaded (the preset file) into the previews folder.

On windows:
Open explorer (the folder icon at the bottom of the screen or My Computer), and double click your main drive (Most often the C:\ drive). Go to “Programs” folder. Not sure at this point what the exact path is… …But navigate through, trying Adobe, Adobe CC, Adobe CS# folders until you find one that has your Premiere install. In the Premiere folder, you should find a folder that says “Sequences” or the like, and inside that folder there’s one named similar to “Preview”. Drop your .epr files that you downloaded here.

Open premiere, and create a new sequence. Set up your output target, then look to the preview area. Click the arrow for your codec, select the codec you wish to use (the r3d codec?), and select a preset for it. Now you can select your output size if you want, but some presets already have one set for you.

Place your clip in the sequence. Render the entire work area. This will create your preview files. You won’t be rendering effects in full target size until you output. This should keep you sane. Adobe has Built in Proxy functions.
If you still wish to add those .epr lists to your AME, you can import them. Then dump your large files into AME and render out a proxy with those. Open premiere. After you have dropped your clips onto a sequence already: Right click your clips, offline them. Now right click and select relink. When the box opens, choose your newly rendered proxy. IF you have edits already, check to make sure you have rendered the effects in the work area. If it’s still choppy, your drives may not be fast enough. Externals can be faster sometimes, as they use a BUFFERED transfer. Internal drives have only the drive buffer, not a transfer buffer. This means they can only get up to 3gbps for the bandwidth, but spinning drives can only move 300-600mbps max, and that’s used by several read operations for your OS, your editor, your files etc. SO the reads drop in bandwidth and speed very quick.

[Ht Davis]

Start by Creating your sequence, importing files etc…
Drop your clips into sequence… Save project.

I know you “Downloaded the PRESETS”, but did you add them to anything? If not,
Open AME, open any preset by double clicking on it. Now see the top right of the dialogue that has opened up… …it has a text box, and several buttons next to it. You can import those presets one at a time. You can also get all of them by simply opening AME and selecting the IMPORT from the top right, just above the preset lists, and then finding the .epr file you downloaded.
In media encoder, you go to file, and select open source. Now you can open a file from your system. Once you open it, it shows in the QUEUE, but it’s just sitting there. Drag the preset to the file marker in the queue, watch it change… OR …the marker in the queue has three major sections with small down arrows. You can click the text next to the arrows and it will open the dialogue to change settings. This is less preferable if you are unfamiliar with AME. You can edit the settings in the dialogue, and adjust to taste, selecting the preset list your preset falls under, then the actual setting for the preset, and then adjusting from there.
Now…
You output that file to the same folder as your big file by selecting the output location in AME either in the queue where you see a filename, or in the dialogue where a file with an extension is shown in text (click the text and it will allow you to select an output location). When you’re ready, click the PLAY arrow at the top of the AME window.
It will build a small file from your large one with faster playback.
When it finishes…
Move the large file somewhere else. Open your project in premiere. It will ask you to select the file to use for your clips, select your proxy.

Alternative:
Once you have the .epr in AME, you should be able to copy it to the “Previews” folder in your premiere program’s application folder under sequence presets. Make sure to send it to the PREVIEWS folder. This will allow you to generate preview files of your proxy type.
Now, instead of creating the proxy in AME, you can simply alter the sequence settings to make a preview with your proxy settings, drop the large clip onto the sequence, render the work area of the sequence. This will create a segmented proxy to use, and if you only render effects and audio afterward, you’ll be golden. You should be able to use your previews to speed up rendering of your final output from premiere.

[Ht Davis]

HMMMMM…. If you are outputting 59.94, you’re not going to get a 24p look… UNLESS>>>>>>
Use After effects or another editor of similar gen, and find twixtor or red giant plugs that will recode the frame rate to a 24p if you want that look.

OR

If you want a more 24p motion, you could apply a motion effect that blurs the edges, smooths out tones, and drops some noise. Plugins can do this, but you’re better off creating an effect by hand, and applying it to every frame.
For this type of filter, I’m envisioning the adjustment in photoshop, but I haven’t worked out the transition to AE just yet…
open the frame in photoshop, dupe background to layer 1, invert layer 1>convert to smart object>set to color dodge or linear dodge, whichever gets you most white>gaussian blur until edges come out, select all layers and use snapshot to get a compilation of them as layer 2, turn off layer 1, black and white layer 2, then flatten, and keep as layer 2, burn edges to a stronger black, gently, but doge mid tones and highlights lightly in other areas (5-10 percent) and set layer 2 to Multiply blend while making sure that background turns on. Duplicate layer 2 to layer 3, apply a blur to layer 3, set to luminance, and adjust opacities to tastes. Flatten, and repeat for each frame.
This creates more transition area around major edges and in shadows, but blurs the luminance so that you get a smoother grain. This drops the pixelation down, and has a smoother “Feel” like film. If you apply a grain filter to this, first drop noise a little with the photoshop noise reduction. This will give your farm a “Film” look. Do this to every frame and you’re all set. Your motion should have a film look, your color should too.

If I could apply an edge detection in AE and dial it in, to darken then blur the transition areas, then have AE smooth the luminance and color noise, and apply grain per frame, I’d walk you through it. But my AE knowledge is a lacking along these lines. I know how to create comps and apply plugs\filters, but I’m not as knowledgeable on which filters are available to AE out of the box. I’m more familiar with photoshop, using TWIXTOR to adjust time\frame rates, and outputting\editing through premiere and AME.

If you can find a filter for a Film look, you should be able to tweak the settings to taste and just run with that.

[Ht Davis]

I’m slower than that and my video is fine…

My workflow:
Use Prelude to select clips and output to ProRes of my chosen level (for longer shoots, use proxy for file size drop). DO NOT RESCALE. We’ll do this later in a more workflow centric manner. You can use AVC intra that matches your size, but you may have to play with the settings.

Create a disk image with about 60-100g of space, and a compatible filesystem (exfat, fat, ext2\3). Now open premiere, create a project and place it all on that disk image (except scratch space, leave that alone; but previews should be with the project and they will get big depending on your chosen settings). I render both audio and video together (in preferences of premiere), but to each their own.

Import your AVC intra or PRORES to adobe (leave them outside the disk image, and it doesn’t matter where–external or internal; most prefer external esata or usb\thunderbolt with plenty of space to save internal drive wear and tear). Once imported, you can build the project file methodically.

Start with your main sequence settings:
set up your sequence to match the output you want to get, equal or lower than resolution of your largest resolution clip (I usually target a blu-ray output and my input is set to match that frame rate, so I resize down to what adobe will transcode anyway). Set the previews to an MPEG or compressed output size, and you can set the resolution to be small enough for your main monitor (small enough to view clearly). Now you should be able to work, but the video will still be choppy… …What to do… …What to do…
Remember this montra and you’ll be fine:
oOOOHHH… …Render in the morning,
Render in the evening,
Render round supper time…
Rounder out the lot at the start, at night, and when the edits are done.

Basically:
Do an initial render. This will hash out a preview for you to use while you work. It will take a while.
As you go, you’ll find the video gets choppy about every 4 or 5 edit marks (when you apply effects or transitions), so render only the effects in the work area constantly, and it will be pretty fast. When you are done for the day, set it to render it all out overnight. When you are done editing, you can pipe your preview straight out and view the whole thing pretty easily. This will give you some idea of how it will look when you output your finished product (as most are compressed products anyway), simply tell the first encoding operation to USE PREVIEWS.

If you’re like me, Quality and Speed are not mutually exclusive. I set my previews to an uncompressed format (or lightly compressed proxy style file) that renders quickly when done right. I skip the overnight rendering completely, and render each transition. I do an initial render to hold the previews, and then I work until I have about 10-12 short areas of yellow or red for render quality (small bar area above video tracks but below the work area bar is green when rendered, yellow when may need a render, and red when needing a render). Since my previews are very standard and not very compressed, I get short runs of them, just fine, but they do get choppy, unless I render out a compressed preview.
I get around this by creating a second sequence, then fiddling with the preview settings, and nesting the first sequence into it. I render that one overnight. Doesn’t hurt my work a bit, and when I want to see my work smoothly, a render of the offending areas will do. Yes this implements a double render for smooth playback. But in the end…

My last step is to render out my projects. I have to get my output work flowing. Since I use a mac, I can use Compressor for a lot of the work. If I want to have a blu-ray video, I just output the bluray video; and I can actually FARM with compressor if I have it on 2 or more machines. This will cut down render time by concatenating the renders from separate machines. I have the file output to my disk image. Why? Later…
I output my premiere sequence using my proxy format and select USE PREVIEWS. It won’t take long to tie those together. I then move to compressor, and set it for my desired use. I’ve been dying to try and use the X264 components in this manner (farming), but there isn’t much info on it. Doesn’t mean I won’t try it soon… But I output with a blu-ray standard, and set it to work on the video only, then do a separate setting for the audio. While that’s working, I can work my sequence in premiere, making all of my markers for encore, and sending my audio to an audition project. I can have audition output my audio or wait for compressor, but either way, it will probably be recompressed in encore.

In encore I send my premiere sequence, and my audio to a timeline, then build anything else I need. Then output to a separate disc image, stored on the original disc image.
I then burn the video disc.

I then burn my Disk IMAGE as a single DISK IMAGE FILE (not in image to disc mode, but in data to disk mode, I want to drag a single file someplace and have it be compatible instantly) to a BDXL and dump it from my system.

A disk image can be opened on any system and almost instantly be compatible. If you have a script to edit the xml in your project files that changes the notation for moving between systems, perfect.
I always name my drives on mac by the project. This gives you a leading name of /Volumes/[disk name]. On a PC, the lead is simply: {disk letter}:/ and conforms to dos naming. Using an XML properties editor, or a text editor, you could change all occurrences of one to the other. I prefer starting with a fresh project and importing the other. This will allow you to simply relink some data, and poof, your file’s are up and running again. Why does this matter? I like to output to AVCHD for alternate blu-ray disc burning encodings, and I usually have to do it on windows. When I do it, I output an AVCHD file into the disk image. Because it has a compatible file system, I can move that image anywhere, by any means. Once on windows, I can open the image, and output my final video straight back to it, then go back to encore and poof. The workflow bounces around, but I can cut my work to weeks instead of months, and it’s all done on a slow 2.16ghz 4gb ram 256mb video card 2xhdd (2tb) macbook pro, with an occasional borrowing of a buddy’s avhc encoding program on windows.

PLUS:
Some Disk Image formats can be compressed, saving a little space. When you are finished you can burn the image right to a disc in image-to-disc mode, or, (more preferable) you can burn the entire image as a file to a disc. Now you can duplicate that disc any way you want, but you’ll always have a backup, and you can archive your work easily.

My final note to you:
I always burn disc images of my flash cards to disc so I have the original video; I make 3 copies. I dump the uncompressed version, usually (as they are hundreds of gb), but occasionally, I use a disk image, and have a special burning algorithm split the file; I make 3 copies. I archive my project on a disk image, and I make 3 copies. This results in a 12-30 disc archive, with one disc for output and my drives are cleaned afterward.