Agreed, but I want to know if they are dealing with the same amount of bits on a disk as we are, how do they get that quality.

And how much $$ is spent.

What do they use to compress the media to MPEG-2.?

Commercial DVD movies are double-layer. Bitrates are typically in the 5-7Mbs range.

Not to discount the other answers, but there are some straightforward technical reasons.

1. Many commercial DVDs are pressed on dual layer discs. This lets them encode at 70-80% higher bitrate. This makes a huge difference.

2. Many (certainly not all) DVDs are from film, which is 24p. My experience with encoding (and I'm definitely NOT an expert) indicates that 24p encodes much more cleanly than the same duration of 60i (i.e., 29.97 interlaced). The difference is not subtle.

3. Commercial DVDs are encoded with technology that is perhaps somewhat better than what is available in Vegas. However, I believe that this last one is overstated somewhat. I would love to see a side-by-side comparison of the MPEG-2 files encoded by multiple encoders, that we could download and burn onto a test DVD to see if we can detect the difference. Having said that, these ultra-expensive commercial MPEG-2 encoders do offer one key thing: manual overrides. I think this is really important. If you have ever done really low bitrate MPEG-2 encoding (like 2,000,000 bps instead of 7,000,000) you'll find that some of the video looks OK, but that the first thing to break down is dissolve transitions, scenes with smoke, and of course really fast movement of objects with strong contrast. The VBR encoders may or may not do the right thing (increase the bitrate) for these sections of video, but a manual override allows an operator to flag these areas for more bits.

Of course if all you are talking about is the overall quality, then the points in the earlier posts are exactly correct, and you can add about 1,001 more, each one having to do with the skill of the crew and the talent, and the quality of the equipment.

Unlike most of us, one common setting does not work the best for all scenes.
They look at a scene, say with plenty of motion, and encode.
Then next dialog only scene may have different settings. Much more work.

The engineer also has the ability to "pan and scan" (correct me if my term is
not accurate) from the master.

Hollywood movies have a specialist spending a few weeks going through every scene and doing a lot of manual work to squeeze the last performance out of every bit going through the Cinemacraft (or one of a few others) MPEG-2 encoder.

The rest is high end source video formats, much better lenses, and much better lighting than most people can muster.

Thank you this is exactly what I wanted to find out. Is there a mainconcept MPEG-2 upgrade available in Vegas 8, or is the one in Vegas their best. Is there a way to render regions in different setting (Bit rates) and edit them back together without re-compression?

Jim

I took as much time working on picture correction for my recent film as I did editing the rough cut. The difference was night and day.

If you want to see "what they start with" and be completely convinced on the value of post production picture correction, rent Santa Claus 2 and watch the deleted scenes or The Shipping News and watch the making of (it contains pre-corrected footage. You will literally wonder how an executive producer watches dailies without having a heart attack.

K

Jessariah is right.
I was in a post house in the Detroit area a few years ago watching a guy doing colour correction on a movie for DVD release.
It was done on a scene by scene basis (and sometimes finer than that) with all the parameters stored on a computer for the final run.
As I recall, the monitor he was using was worth as much as a mid-size car and was calibrated daily :-)

Mike

I'm not a professional, but I think it's obvious that the size and quality of the original matters quite a lot. That and the fact that film is progressive. Try throwing into a Vegas timeline a bunch of photos you took with a decent consumer photo camera at not even a really high resolution, something like 4 MP. Do a slideshow video of that and render to MPEG 2 in DVD format, whether it's widescreen or 4:3. Then encode something you shot on SD video using the same module, same bitrate and everything. Burn those two to a DVD and play it. Even if you don't have an HDTV you'll see that the timeline with the photos is going to look far better than the video one. It's because even though both are 720x480, the one with the photos came from a source with much higher resolution than the video one.

A similar thing happens when I burn to DVD format video that I took from my HD camera. It looks so much better than video I shot in SD. The higher quality the source has, the higher picture quality you get, obviously as long as you use a high bitrate to reproduce it as faithfully as possible, which is especially important these days with so many people using upscaling DVD players or Blu-Ray players on HDTVs, because this setting always shows mosquito noise much more than tube TV sets.

While this is true the advantages of that only seem to kick in at lower btirates. From my testing I have to go along with JM.

I captured a Digibetacam dub of one of the Planet Earth episodes and made a DVD of it at 8Mbps CBR. It sure did not look too shabby at all compared to the original off tape. So it isn't the MC encoder, it would be if you're trying to fit a full feature length movie onto a SL DVD plus all the extras etc. Hollywood DVDs do seem to be at lower bitrates than I can get a good looking encode at but that's a different issue.
What I strongly suspect made my test look so good was the content, the way less compression in the master and the low noise floor.

I'd also point out something it took me a long time to grasp. Two images can look identical...until you try to grade them or do anything with them. Someone above mentioned how bad dailies can look, they're typically straight from the camera neg with no grading. You take a DV or HDV image and try pushing it that far with CC and it'll likely fall apart. This is where those uber expensive cameras and their massive amounts of data being recorded have it all over the stuff we can afford. We can for a reasonable price record images that looks just about as good, until you futz around with them.

Bob.

I believe Bob has hit the nail on the head.

This reminds me of when a band I was part of built a studio to release our CD's. This goes back to 1991. We continued to buy more and more equipment, work on arrangement, composition, playing, etc.. everything in order to achieve that "record label" sound.

Is it the equipment? Yes partly. But once in a moment of desparation I recorded a bit of the latest Prince CD (really solid production) onto a mere cassette tape. Puny 1/16" tracks moving at 1.75ips if I remember correctly. You know what? It still sounded great. Sure there was some tape hiss, a bit less dynamic range, and less high end but all that didn't matter. The SOURCE was fantastic and that is all that mattered. As we worked on our gear an chops things got better and better and I believe we got very close to a "record" in some cases but there is something to be said for great rooms, mics, producers, engineers, players, and most importantly songs.

So it is my opinion that the final pixels that are laid down in a commercial DVD count. Every one of them. And you could transcode them with simply a good encoder at a sufficient bit rate and they would still look fantastic.

Try it. Rip a DVD that you think looks great and re-encode a portion at say 5500kbps 2-pass. Unless there it is all action I bet it will still look great.

The old adage "garbage in garbage out" really is true.

- Mark

I think your underlying question is whether a better mpeg2 encoder would get you a DVD that looks like a big budget feature film release. Sorry, but I just don't think there's any way for a video to be improved in the mpeg2 encoding process. If it looks good to start with then a great encoder will preserve that good look, but it won't turn a sow's ear into a silk purse.

Really good equipment helps. Good sound, good lighting, good costume, makeup, art direction, lots of money, the best lenses, a good recording medium...

I used to light music videos for an up and coming DP. We did a lot in 16mm until he built a rep and started getting people to pay for 35mm film and gear. The first time I saw some of his raw 35 footage I was blown away at how much better the color was. Same film stocks, just higher density.

Top quality gear is important but I think it's more important to get top quality teams on the job. Actors and director need to work well together, director and DP need to be of one mind, DP needs to know how to direct his/her lighting and grip crews, DP needs to talk color with the art department, Gaffer needs to have mercy on the sound department, first AD and his/her production crew need to communicate with everyone to get the things they need to the shoot on the days they need them, and they need to manage everyone's time well. And almost all of these people need to do their jobs without distracting the Actors, Director,and DP.

Rob Mack

All the above comments are good. Problems with motion and "color" really impact encoding. "Colour" control is critical from frame to frame, the smallest colour shift can eat up bit rates and effect the viewing "focus" of the scene. This is why mixing different K degree lights sources are "no-no" without clear planning of the shot.

And shooting in non-subsampled 4:2:2 (or even 4:4:4 nowadays!) makes a big difference [this is not Panasonic's "prosumer 4:2:2," which is sampled off its CCD resolution of 960x540, making it 4:1:1 relative to 1920x1080).

Real 4:2:2 is especially much better compared to shooting in 4:1:1, where 75% of the color is thrown away before it is rendered to 4:2:0 MPEG-2 for DVD which leaves only 12.5% of the color information to try to figure out what the original scene looked like.

(Most Sony and Canon prosumer gear is 4:2:0 which doesn't have [much] loss when rendering the footage to DVD.

Now, crappy CRI lighting is very difficult to compensate for, it makes sense to give the lighting serious attention.

KinoFlos lighting for example looks gorgeous at CRI 95, especially skin tones. Big difference.

Vegas' MC encoder is very very good, it just can't do some of the things that top end encoders can do, like full manual control of i-frames and 10-pass rendering.

What I am trying to find out is what do the studios do, and is there any of it that I can reasonably do to improve my work.

What I am trying to find out is what do the studios do, and is there any of it that I can reasonably do to improve my work.Did you read ANY of the responses to your original question??

Sorry for the confusing post, I have read and I am looking at all of the great information. I was responding to a previous post that I forgot to copy.

"I think your underlying question is whether a better mpeg2 encoder would get you a DVD that looks like a big budget feature film release."

Thanks,
Jim

Jim,

With Vegas' MC encoder, getting a "better" encoder should be something like #25 on your list of things to improve.

It really is very good.

Seriously.

Since you're asking, I take it you're shooting NTSC DV?

As I explained above, that sucks for DVD delivery, regardless of the camera. Years ago I shot NTSC DV with a $12,000 Sony DSR-500 broadcast camera, and it still sucked on DVD.

Back then HD wasn't affordable, so I switched to standard definition D-5 (uncompressed 10-bit) which looks good indeed.

Today I would get a 4:2:0 XDCAM or HDV camera, these give excellent results also for SD DVDs.

I answered your question about the quality of MPEG-2 encoding, and I answered it in detail, in my first post above. As I stated -- and as others have confirmed -- the Vegas MPEG-2 encoder is almost as good as any other encoder in most ways, but the pro encoders let the operator manually choose portions that need more bits in order to be encoded cleanly. The 24p vs. 60i is the other big reason. Sorry to repeat myself, but your second question is basically a repeat of what you originally asked. If you need more help, you need to ask something more specific, rather than simply repeat your original question.

So, do most of you shoot in 24p to get best results?