[Ht Davis]

First, most of the creative suite was bought from other devs, so take what you would think would be or wouldn’t be and put it away somewhere else. Sorry to say it, but they are very different programs.

Now…

Remember that replacing clips isn’t like switching on the kitchen light. You have to select things in the proper order. IF you have the clip in the BIN and it’s selected (the one you want to plant), you can select the clip you wish to replace in the timeline and use the replace with clip>from bin. If it puts it back at the first frame marker, your playhead is there, and it’s trying to paste with no clip directly selected. SELECT with Left click, THEN right click and use the replace. Also, make sure your playhead is where you are replacing.

You need to replace clips with clips of not only the same length, but same general profile; i.e., they need to have the same in\out, the same or similar reference for sound, and they need to have the same number of frames. If you use sub clips, you’re better off, as it will define the clip as a sub, and allow different length, frame speed, sound ref, in\out; all without making a fuss as the sub clip conforms the sub data to the timeline, but references a source clip’s data that may be different from the timeline or the clip you wish to replace. Occaisionally, you’ll get a clip with a large audio ref or a bad header read. Repair disks, reformat the clip, or use a sub and try again. HDD errors can cause the read problem, offline media can occasionally cause it, and of course, some unsupported formats will cause it as well.

[Ht Davis]

A full format is an uncompressed file (in comparison with other formats; Compression “Loses” or drops data and causes drop in quality in the image detail and color). Basically, it means that more of the image and color data are retained, which usually results in better quality, higher bit-rate. If color is your problem, there are some possibilities:

1. Conversion of Colorspace: When you go from compressed formats from your camera or a video app to a conversion in another program you can lose or shift the color around based on how the two programs differ in how they handle the data.
2. File to Monitor: Make sure you have a profiled monitor. It needs to be calibrated correctly. There will be a definitive difference in the playback in video editors and players due to separate methods of handling the color and data. Different players will play it back slightly differently, especially if the colorspace is too wide (48 bit is wide gamut and some monitors and players don’t fully support it or need to be adjusted for it using an effects panel in the player). Check your color gamut. Try using 24bit colorspace. Most players will be very close with that selection. Wide gamuts require more handling, and those methods are different in every player. Try 24bit. If you are in 24bit already, then try checking for a whitepaper on the player you are using and find out what gamut it uses, and if it has a video filter you can use to adjust the brightness and contrast. You’ll find that different players zero out at different GAMMA values to comply with different standards of playback.

[Ht Davis]

It sounds like the server you’re using renders new sizes for you, much like google and youtube. If that is the case, it may be that there is a problem with the server, possibly an update that ruined the low-res output. Get in touch with the hosting people. See if you can offer your own output files in low res so they don’t have to render it for you. There have been a lot of GFX pitfalls lately with rendering and output systems, mostly related to video card drivers and parallel processing of video data. Storage may also be the problem. If you upload a file at 20gb in 1080p, it then gets processed to 13-15gb for 720p, and so on down the resolution line, taking 50-80gb for video storage. Maintain the storage and you’ll get better results. Alternatively, if you play a high resolution video at a low resolution, same data but with a low end algorithm for resizing on the fly, it will pixelate, which is why youtube re-encodes several files of output for every input. You set the max resolution, youtube handles the lower res, and quickly. The reason for dropping resolution: Bandwidth. The more speed you have the higher the res you can play back. If bandwidth is a problem, they may be setting it too low for your video to display properly. In that case as well, offer to give them a file that displays properly at that bandwidth; you may have to drop resolution to get a desirable result, but once you find the balance, you can continue to use it.

Check with your hosting. See which method they use and offer to pass along low res files of your own that match their bandwidth, resolution and encoding settings.

[Ht Davis]

You can import the sequence itself into premiere. The media is all offline, that means that you need to relink each media file to a corresponding file. Each clip should have it’s own file that you link it too. While adobe doesn’t natively support the import of the lastest AAF, you can import older AAF files, and then relink all the media. The AAF only brings in the REFERENCES, so you can see the references in the project, but you still need to tell premiere what they refer to. A clip does not refer to an AAF. It needs to point to a particular file. If these clips are all the same file but different sections of it, you may be stuck for the moment. Ask for renders of the clips and you’ll be set. They don’t have to be high quality while editing, so you can get low res and make any adjustments, then get the hi-res later for rendering a final output. If the media files are offline, they need to be relinked to their mov, mp4, fla, mp2 or other type of media file. An AAF is a scripting file similar to XML that outlines the sequence and references the media files; if your system doesn’t have them in the same exact location (or doesn’t have them at all), you won’t be able to link the files in premiere. Get the clips separate from the AAF, use the AAF to have the sequence and references import, then relink each file. If you have hundreds of files, it will take a while. You could try adding a folder called MEDIA, import the actual clips into it (not the offline ones but import new so they are online) and then follow the sequence replacing the offline clips as you go; when a clip is trimmed for the sequence, trim the actual clip you imported, then replace using it or a subclip from it.

[Ht Davis]

1. 10.8 isn’t officially supported. Go back to CS6, it runs clean. If you have those specs, you should be running early CC (late 2013 to mid 2014) for compatibility in 10.8, as the graphics situation is much different. There were fixes to retina problems that were compatible with 10.9 and above, but do not function properly in 10.8 and lower.

2. Even if you are running OPENCL, you may still have CUDA installed. It isn’t that RUNNING it causes a problem, it’s having it installed at all causes issues. You must wipe it off your system completely, and then run your adobe software. IF you still have a problem, make sure CUDA is still gone, and WIPE\REINSTALL the adobe package.

3. There has been little effective solution to this problem within the CUDA firmware by any of the parties involved–NVIDIA, APPLE, ADOBE–and as of now, I urge you to consider this software DEAD. It’s time to ABANDON SHIP. CUDA was a novel idea. It works on Windows in many cases, but it fails the MAC test. If you wish to rely on it, get a Windows PC. I’m a Mac Man. I’ll live with OPENCL. I’d also urge you to learn to renderfarm with AE and Compressor. Any edge you can gain in multi-machine-encodes makes up for the lack of CUDA. Since Transcodes take the bulk of the time for my work, tied only with archival, I set one to go, while I prepare the other. This means I only spend that time once, finish the project, and then archive for a better bottom line. CUDA is DEAD. Find another way. RETINA in 10.8 is not compatible with the latest CC. Try an older version.

good luck.

[Ht Davis]

DO NOT USE CUDA!!!! Lately it has been having a multitude of FAILURES. I’ve been watching this for some time, and I believe that, while it is great when used on Windows, it WILL NOT WORK ON MAC.

The story on CUDA: APPLE has been pushing their small device market, and abandoning their pro-computing model. Most of their resources have been leveraged toward small devices, especially with the market in china blowing up; they are a status symbol almost everywhere, though they are a bit restricted compared to their competing counterparts (we are treated like drones; lets face it).
This latest problem with computing gfx is no surprise, and there hasn’t been any real action toward finding a solution; they blame nvidia, which reports they are attempting to track down the culprit, but they have seen very little help from Apple in the way of allowing access to the frameworks. Because it is designed as an engine to allow PARALLEL graphics processing (basically both cpu and GFX card processing at the same time in real time), it apparently has an issue where the system architecture forces it to access random memory locations (IRIS locks down the memory it uses?) along with a few proper ones, or, the two fight each other for the control of the screen in such a way that scrambles playback, at least in ADOBE software (mac software doesn’t necessarily use CUDA, I think it sticks to OpenCL and IRIS load balancing). IRIS is the intel architecture that checks the workload, and only sends it out to the GFX card when it needs to, otherwise processing the data using the main CPU and internal ram only. In more extreme situations, it actually utilizes both, but all data has to pass through the main CPU in either direction. CUDA skips the CPU passthrough. Something in how this is handled on the mac causes playback problems in ADOBE. Skip the CUDA drivers. They aren’t working.

My setup:
I use a core2duo MBP with 4gb ram, 2x 1tb internal hdd’s,256mb gfx, esata express card, and firewire based RAID to work on 1080 and smaller video while editing and occasionally rendering on it as well. I do this for a non-profit performance group, and allow them to upload to youtube, or send discs out to benefactors. I don’t have a budget at the moment, they are just getting established; there is talk of allowing me to build a small render farm with mac minis. I’ve outlined the specs as i5 16gb models from 2011=2012 (non soldered memory that I can upgrade later), with hg gfx cards of 1gb, 512 SDDs for programs but cheap external raids on thunderbolt or usb3 with 2-3 drives at 1-4tb each, for storage (makes a fairly decent farm for shorter projects at 1080 or 2k).

My Suggestion: 4k, 5k, 6k… …If you have the money to go with a faster machine, it will last a little longer, sure, but not by much, especially when working with the higher end of resolution (it’s always going up). Since it is an iMac, it is an all in one, and not very adept at upgrading much other than a hard drive. The Processor, IDK for sure, but some have attested that it is soldered to the board (YUCK!!!). As for Ram, I’ve only seen that soldered in the latest mac mini (ugh! I just threw up in my mouth a little there). The GFX card should be your main target, but these other considerations are actually going to be your main limiters. Since it is not very upgradeable, you’re looking at 3-5yrs feasibility no matter which option you choose.
4gb is wonderful if you work large projects (2-4hrs +) and render on a regular basis in 2k or 4k. If you are only occasionally uploading to youtube, you can get away with 2gb and stick with that until you’re ready to upgrade to higher resolution. IF you are at 2k, this will be very fast indeed; at 4k it will make the cut. Beyond that, it will begin to seem slow. If you are planning on doing any kind of heavier work, you could go with the lighter model for direct editing, and save up for some minis to use as render farms for output. Even older minis would be helpful when outputting your files. Just use AE to render out a folder of images, and a sound file, roll that into a movie, and compress that with AME or compressor (compressor will farm out the compression step).
Plus, with CC and the new ANYWHERE function, i’ve heard it is possible to render the sources on outside machines but playback and edit on your primary in premiere (similar to farming a render, but for previews only; set previews to higher quality and it will function like a full render farm, but that will slow down the edits more than likely). The key here is to define your workflow and match it to a setup.

Why this opinion:
I’ve done a few jobs with a pro at 2k, and they had a lot of machines at 1gb gfx or no gfx but 32gb ram across the board and i5’s. We also used several PC’s for another project with similar specs. Both had roughly the same render times, very fast compared to my laptop setup. Here’s the kick: I did all the editing on my laptop, and only used the other for rendering. I had to relink on pc, but it allowed me to make 2 of everything, a directors’ cut and one that “went to print”, or that was sent out to the masses. The compression steps were done on mac with compressor and qmaster. I couldn’t believe how fast it went. That was just last year. A 2 hour video took us a week and a half to produce, with several of those days spent waiting or reviewing with the client (a non profit client, but a client all the same).

I know that explanation was long and tedious. But maybe it’ll be of more help.

Brass TAX: if you’re a hobbyist, you’ll do fine for a while with 2gb gfx, just up your ram to 32gb; but forget about cuda. It’s a dead horse, shoot it and put us all out of the misery. If you can work with windows, you can get a cheaper computer and OS, with better specs, and still have access to CUDA. IF you play your cards just right, you can get a MAC COMPATIBLE system (mackintosh) that will allow you to make small upgrades as you go, putting you in the 5-10 year range for feasibility and upgradeability, but you may have to customize it on your own, and you’ll have little support from apple, adobe or any software manufacturer, but better support from the community that goes with this type of setup (go figure… …a bunch of non-paid garage computing enthusiasts give better support to the masses than some big company who could care less if the damned thing works as long as they put billion$ in their own pockets; sound like anybody we know? Or should I say, “Knew”… I’m sorry for the way these imbeciles are destroying the house that STEVE built… …We miss him…)

[Ht Davis]

Does anybody here know what a disk image is? You can make one and have it on a shared location (network drive), and actually just copy necessary files to it. I have 2 situations to expound on this methodology.

MACS
Mac users have access to Compressor (for a price of course), and can even have it leverage across several machines for any operation. That said, this question is about adobe software, lets get to that. Since you cannot go from PPRO to compressor directly, let’s talk about how adobe software handles the files, so you can utilize AE for this operation.
Lets say you have your source video in full format on one drive with a NAME (lets call it “SOURCE”). You can create a Disk Image with the same volume name, that will show up with that same name on OTHER MACS as long as they have access to where the disk image is stored (the image file needs to be in a shared location). The reason for giving the disk the same name is that you want to copy the necessary files over to the image, so that they can be accessed from another machine, but use the same volume (disk) name when finding the files. Now lets make an image with the same name as the disk with your project files\cache files, previews. You’ll need a disk image for every separate disk you span your files across (note that this “disk” term refers to PARTITIONS not physical hard drives unless each drive has only 1 partition; raid setups use several physical disks for one big partition), and you’ll need to use the same names for the image volume (partition on the image) as the actual drive you are trying to copy data from. Copy your files onto the images, and store the images in a shared location. Open the images from the alternate location, and share the volumes to any other computers you wish to farm out to, render it out from AE. Now the data is not on the drive you’re working on, but is still being rendered without needing to relink, re-conform, or recreate any information.
Now back to Compressor: With Compressor and the queue program it attaches with, there is an automatic sharing function that will work if the files are on the main drive. Turning this off does not prevent sharing. You just need to make sure you share the drives. By using an image in a shared location, your network speed will be your main limitation, and it is best to use wired net. However, I’ve done compression farming with wireless. I render out using adobe software for a full format, but afterward, I use compressor. It doubles and triples your power in a mac environment.

WINDOWS:
Granted mac has an advantage with the compression and sharing that last task, but that task is easier, more accurate, and better quality after the initial rendering to a full format, no matter what system you are on.
You can farm in windows the same as mac with adobe after effects. You just create a disk image using your favorite utility, or the command line. You will have to take your work machine out of the farm or relink your files to the new disk images, but only once on that work machine; the rest will access the same files and folders the same way.
Windows uses drive letters as volume names. It saves on memory space in the OS, although the recent GFX of the interface have made that moot… …Sorry for the social commentary. Nothing against Windows. –getting back to the topic– You’ll need to mount (connect) the disk images to the same drive LETTER across all your machines. In a WinServer environment, this gets even easier, as you can add the location at the server level, and have it propagate to every computer connected, while you only need to attach the images to the server. In some of the newer Server setups, you can use ISCSI to access the images as if they were actual drives, and provide even faster access, as it allows every computer to bypass the authentication and flagging of the server, even while not logged in. Just make sure the drive letter is the same, and you’ll have an easy time of setting the farm from one machine and running it out to the others, without overtaxing your working drives.
Once the primary file is rendered, you can compress the file using any compression program you like. You can send it to AME for compression; or, if you have access to a few macs, you can send it to compressor and qmaster it out.

In order to render farm from either case, you’ll need to install after effects on more than one machine, and you may need to install premiere or any other apps so that their XML engines are installed. Rendering out from AE in a farm may also require any third party plugins. A test of each of them in a single short clip both farmed and single computer rendered should provide a proper comparison that will tell you right away.

Render farming is the standard for longer projects, higher resolution, and\or lower cost machines. There are many linux editing environments that are now being enabled for farming over third party server-based applications that use endpoint connections to leverage all machines connected, or a subset of those from a profile created at the server app level, to actively process sections of a project, and automatically re-utilizing machines that finish faster for more sections. These are provided with Freeware basic licensing for small renders, but require licensing for larger farms. The more awesome part of this is that, with the right plugins, you can mark your cuts, edits, effects etc, and then move all of that to a new environment, and redo any incompatibles quickly, moving to the render step shortly after. Some of these third party Linux based farms even have plugins for adobe software that link them and allow them to utilize all adobe’s power of editing, but all the power of every processor linked to the farm to quickly and efficiently render out any project. I only wish I had the funding to put one of these together… A guy can dream can’t he?

[Ht Davis]

I have a multicam shoot, multiple angles, multiple clips. Speedgrade isn’t an option because I haven’t been able to get it to open.

Depending on what you’re actually trying to match, the fix changes. Speedgrade was great before, but it was extremely limited. I prefer AE. It has a multitude of methods for matching, whether it be luma\chroma, or masking and swapping. And, because photoshop can also handle video, you can use photoshop for single color replacement the same way you would for any photo, applying it to the entirety of a clip or set of video clips.

If luma\chroma is your problem, maybe you’re like me, and you shoot with many different models\brands of cameras on low budget projects and have to grade multiple angles. This method works when the two have some crossover (i.e. an object or area of high, low, and mid that matches). It works even better if there’s a scene where you have a balance card in the shot (you can shoot the card to balance a shot and use premiere, but if not, you can use similar objects and areas in AE).
——Fix:
Correct a main angle and render it. We will call this Color A. Now… …Line up your multi cam if you want, some wait until later and just use proxies with corrected color. The process is very much the same.
At this point, I’ll try to follow both, so… …For Premiere Line up your multcam, then select the clips you wish to correct, each in turn, and replace it with a comp. It will open in AE. You can make it into a new comp. You will do this for each clip you want to correct, but don’t worry it gets faster afterward. I suggest you just save the AE project, and do it for each one.
If you’re going the second route, open each full media in an AE project (this method allows you to use a single project file). For each video you want to correct, create a comp by right clicking and creating a comp from it. Save the project.
Now we get into the correction. With the first method, you’ll need to open each project, with the second, you will be able to just repeat the changes for each comp.
Import the already corrected video into the project, but leave it outside comps. Open the video in a panel, and drag it outside the other panels ( you want two panels open) On the video layer in a chosen comp, a video you wish to correct, add the Composition Color Matcher from the Composition wizard box, and open the video in the second panel (again, you want to see the corrected video and the uncorrected). Now look at the effect controls on the uncorrected video, and look at the Composite color matcher. It has two Luma\chroma settings. In the first one, you want the whites, grays, blacks (highlight, mids, and shadows) from your uncorrected video, so use the dropper to select those. In the second set of values, you want to select the same or similar areas from the corrected shot. One will be adjusted to match the other. You can render a proxy to use in Premiere, and afterward, you can render the full video for use in final production (I’ve gotten better quality from using full size files, but with better equipment, you’ll get similar results either way).

[Ht Davis]

So you’re showing a PIP style effect… Why not do a PIP style edit?

Render a copy of whatever you’re putting in that box, at that exact size. Now you can place it over the top of itself, the same as your box really, but now it’s another video layer all it’s own, and will move on it’s own. Move the lower video, and the upper video doesn’t go anywhere.

[Ht Davis]

I don’t know if you got this to work, but as I recall, EDL conforms to:

1 video with 1key video tracks (only 2 video tracks)

and 1 audio track, stereo or mono.

If you have more tracks than that, you may need to choose a different export format. You might try exporting to final cut xml, then importing that into premiere.