Lessons

Lesson Info

Video Compression Summary and Q & A

As I was testing this software over the last month, I was struck by how much variation there was in compression times, though much less variation. An image quality or file sizes, especially at high bit rights. Now you're results may very depending upon the source video Kodak that you're using the image size and the computer hardware because different Codex, as we've discovered, compress at different times. And it may make sense because you want to save time and compression toe trans code from shooting camera native to programs civil, because progress is much more efficient to compress. What this really means is that testing to see what works fastest for you is always a good idea. Brian's asking. What about the simple export? Right out of final cut 10 to YouTube, all the work is done. Is there a hitch? The answer is maybe compressor and final cut used the exact same compression engine. It's what's called the demon Ah ah, UNIX application that runs without any user interface Control and ...

compressor is simply a front end for that back in process. Final cut feeds that that background process the same was compressor. Thus, you get the exact same results. Eso Let's just fire a final cut here while we're looking at it. When you go to YouTube export, notice the compression preset. You have better quality or faster. Encode better quality means to pass faster. Encode means one pass as long as you select faster encode, then you'll be doing the one pass setting, so that's gonna be the fastest out. I would not, however, select better quality when you're going to YouTube because you're not going to notice the difference of to pass over one pass when your data rate is as high as it is. So this is working with the default YouTube setting based upon here. Notice your resolution. It's using the YouTube 7 20 p setting, and we're changing it to be better quality. That's the default setting to faster encode, at which point you're gonna be able to take advantage of this speeds that we've been talking about. You can optimize mawr inside compressor. Now this is a true statement. If you are working with a a new I Mac or a Mac book pro. If you're working with a, um, a new Mac Pro, you may be better off exporting this is a master file, bringing it a compressor and change the settings to optimize for the Mac Pro. How can I upload to YouTube? Terry asks. At 10 p. The compressor default setting a 7 20 PM looks like there's no way to change it. Oh, yes, there is. Let's take a look here. Let's just add a one of my favorite generators, which is a blob, will set this to 10 80 p. Click OK, and now we'll export for YouTube, share YouTube settings. If your project setting is 10 80 you can export 80. If your project setting is not 10 80 you can't export 10 80. It won't allow you to, uh, press, because that's just gonna look ugly. It only allows you to export for YouTube at the size of your project or smaller. Oh, Sharon, If you're making a downloadable file, can you just use the YouTube setting and change the rate from, say, 10,000 to say, 4000? Yeah, it's the wrong question, but it's so you can do it. Continuing my question, the YouTube setting seems to give better video. Even when I try to make the impact force setting the same settings. The highest image quality comes from the fastest data rate. If you had ah data rate of 15 megabits per second, you're going to get great image quality. But the file is huge, and nobody wants to download files. So that big, especially on the cell phone, it's gonna kill your data rate. And for a computer, it's gonna take a long time. People hate to wait. They don't want to fill up their phones. She got to make the file smaller. So now you're in a trade off the lower you make the data rate the potentially lower the image quality is so you could make the file vanishingly small on DSEC. The data rate toe one megabit tiny file. But if you had a 10 80 p file, it would look ugly. Have rectangles, ugly rectangles running all over it, so you're always trying to balance. How do we balance file size with image quality and image quality is based on five factors. It's based on the size of your image. 10 87 20. It's based upon the frame rate. 24 25 30. It's based upon the Kodak Pro rez H stopped to 64 extra camp. It's based upon the bit rate number of bits per second, and it's based on the amount of movement between the frames. You can compress that talking head video down to virtually nothing because nothing's moving except the person's face. But if it's a dance video where you're on a hand held camera, you can't compress that at the same rate 4000 killer bits per second on a talking head. You're gonna have mawr than enough band with you could probably compress out to 1200 cabe. It's a dance video at 4000. You may be pushing it and still get artifacts in there. C'mon of movement between frames is the sleeper. That's why you want. If you know that you're going to the Web, putting your camera on a tripod is gonna lie to compress it smaller than if your handheld, because the way compression works. So it's a balancing act and the way that you balances frame size, frame rate, Kodak, amount of movement between frames and bit rate. Those five together yield image quality. What's your preferred method of compression? Compressing straight from an editing software or using an outside compression program. What I did is in compressor when you have a complete a job, see where it says the elapsed time. This will show you exactly how long it took for that job to compress. But there is no equally cool feature inside, um, media encoder because there's no elapsed time display. But there is something equally interesting. We need to see if I can pull this in here in a second. When a job is complete, there's gonna be a check box right here that the green check box indicates that compressor is done, compressing it when you hold the control key down or right mouse, click on the green check box it will I. It says Open log and log will show you when a job started all the different tasks inside a job, and there's lots and the time the job ended. So if you want to find out how long something took, right mouse, click on the green check box opened the log and do the math to separate the UN time from the start time, and I will show you how long it took to compress. Hello, Bruce. If you want to play the highest quality video but can't play pro Rest files on the device is, for instance, a raspberry pi. Can quality be improved with age stop to 64 with a greater bit rate of, say, 20 megabits a second Thes air videos to be played for art installations? I probably wouldn't go higher than 15 megabits per second just because of how each dot to 64 works. But, yeah, if you if you had aged up to 64 15 megabits a second, it's gonna look better than a five. So the answer is yes. Hello, Thomas. When using four K media, do you recommend editing in a four K sequence and exporting to 10 80 or editing in 10 80 having f. C P performed. The scaling for you obviously used cropping ability in the former. If you don't need to reframe your shots than shoot four K, edit four K and let compressor compress it from Fourcade of whatever your results are because you're bypassing, says step of scaling. But if you want to be able to reframe your shots, shooting four K editing 10 80 in out putting 10 80 allows you to reframe shots. So if reframing it's not important, you'll have higher image quality if you shoot, edit output four K and compress that in a single step down to whatever size you need for the Web. Hello, but what's the best compression for video? You used the YouTube settings in media encoder. You go take a look at the presets here, preset right here. You've got video settings. You've got YouTube settings. Those are all stored inside Web video. This is video. This is YouTube bond. There may be a Facebook as well, not just those two, but the settings are essentially the same. They just have a different name in compressor. Whether you're working with video or Facebook or YouTube, use the settings inside video sharing services. And if you want the resulting file to compressed file to be 7 20 p. Use HD 7 20 or HD 10 80. You can even go to four K but posting a four K images simply bragging rights from marketing. Nobody is gonna be able to see that file playback and four K Russ. What does reset que accomplish? Where are we, uh, see is that compressor? What Reset Q. Does is. Remember I said, that compressor is just a front end for this background process. This demon that runs it's called compressor D. Well, you're queuing up putting a list of files together, that that background process is going to run and sometimes, and it isn't your fault, that background process gets completely confused. So what reset Q does is it? Deletes the file listing all the the master files that have to be compressed resets it back to square one. This is similar to what digital rebellions compressor repair does, which resets compressor in case it gets hung up and can't move. Resetting Q. Means that any jobs which have not been processed are going to be flushed, and you have to resubmit. Um, hello, Thomas. Any tips for compression using footage with fast motion profession, especially professional sports? The way you compensate for fast motion is you raise the bit right. If motions doubt is slow. Think talking had pulled a bit breakdown. If motions fast, the bit rate goes up. It's a lever, Um, assuming the frame size remains the same, the frame rate remains the same, and Kodak remains the same as motion increases. You have to increase the baby the bit right. How soon will age start to 65? Be part of our workflow? My guess is probably next year. Late this year at the earliest. Suspect was approved January 2013 and it's taken a while to filter down into software and and all the rest of it. Hello, Yuri again. Why, after uploading on YouTube, is it necessary to me and my client to wait for two hours to one day to see the best quality of my 10 80 p movie on YouTube? Remember, I said that YouTube always re compresses your video? Well, not only does it have to upload the video and process what you sent it, but it has to re compress. And it's creating about 20 different versions of video so that it's what YouTube does is it asks your browser what browser you're using, what frame size you running and what's your data rate? What's your Internet speed? And then YouTube dials in the necessary version. One of those 20 versions based upon your browser, your connection, speed the image size, all of this to get the highest quality image playback. Well, it takes time to get all those different versions created, especially if your movie is a long one, because it now has to create not just h 0.2 64 but a variety of other Codex that YouTube uses that different browsers support in different mobile devices. So the time is spent in getting YouTube's gear to get that file re compressed, loaded, the servers linked and able to be accessed. It just can't happen instantaneously because there's too much data that's gotta get moved around, Keith says. Larry. You don't need to render final cut 10 files before sending the compressor. Absolutely true. Final Cut will render anything that hasn't been rendered during the export process. So here's Here's a trade off. One of the questions we had at the Nine Oclock show was Well, what happens if I use ah Kodak like, say, a V C H D. Can I edit a V C H d? The answer is yes. But now when you export, whether you're going send a compressor creating a master file, we now have to deal with slowness of that Kodak. It's gonna be a much slower Kodak to export, and now you're gonna be waiting for that Kodak to render and export, which is slower than if you had converted it to, say pro rest for 22 So if speed is your goal, optimize your media for final Cut and converted to pro rez because you're gonna be able to edit faster, render faster and, most importantly, export faster than if you doing camera native. If speed isn't critical and spa storage space is critical, then leave it a V C h d. But you're gonna be paying a speed penalty every time you add, in effect at a transition or export. When you're working with camera native and in some cases you're gonna be paying a quality penalty as well. How can you determine the time it's going to take to compress a certain file? My preference is always to do. It is a two step process. Always export is a master file from final Cut and then compress it and compressor. And the reason is, I want to look at that export file. Does it look good? Did it EXP work properly? Is everything in sync that I have a typo? It's amazing how many typos I spot after I export that I never noticed when I was working inside Final Cut. So I use it as a quality check. Yeah, takes a bit more time, but it's a whole lot easier toe. Check a file when it hasn't been compressed and exported, compress it and then realize you've wasted an hour, two hours, three hours compressing, and there's a typo in it, so I always want to check my files first, then move them into compressor. It's not because it's more efficient, it isn't. It takes more time, but I like the quality control of being able to see whether the file output correctly met the technical specs and creatively looks the way that I wanted to look like. As I was researching this presentation, one of the things I discovered is that every time I started to talk about something, I realized it was something else I need to talk about, explain the setting and explain another setting, and it just got incredibly deep and way too complex. I've written a lot and talked a lot about compressor, but I've never talked about it from the point of view of how do you make it go faster and that depends upon the hardware you're using. The Codex you're using the settings that you're using, the software that you're using all of this stuff has an impact on how to make video compression happened faster, sometimes going fasters a trade off in quality. But more often it's a better understanding of the mix between Codex hardware and software that allows you to compress video faster and learning how to compress video faster is what today's webinar is all about.