What's the best MPEg format for DVD

[Devil]

You can edit a DV AVI clip and re-render it with practically zero loss of quality. You lose quality if you need to render an edit in MPEG. If you don't mind bad quality, by all means use MPEG.

[Gra]

I've been following this thread with interest as I've had similar questions in my mind as I'm never sure I've got things set at optimum.

From the last mail can I confirm that if I set to DV - Type 1 when capturing from my DV camcorder, it automatically sets the capture bit rate, etc and you don't need or can't make adjustments? Then in Video Edit, if I choose DV (PAL, 25 fps, Audio 48kHz), I don't need to alter anything? Or should I look to change some settings in edit, e.g. from DV Audio PAL 16bit to PCM or something else - at the moment my video settings are 6k bits, constant)?

Sorry for being a dunce again but I'm never really sure when some of the threads go into tech speak, with equations, etc.

[Devil]

Just use the DV Type 1 default settings for your standard (PAL/NTSC): change nothing except possibly the Field order (Lower field first if your final viewing is TV, via VHS. VCD or DVD etc., Frame setting if it is for viewing ONLY on computer monitors).

[sjj1805]

Sorry if any of you are confused hope this may clear things up.
If you intend to edit your video then you should "capture" in DV which is a form of AVI (AVI is a term relating to several different lossless formats)
When you connect your camcorder to the computer you are not capturing as such but transfering. Therefore you don't alter any settings at this stage.

If you do not intend to do any editing and your computer is powerful enough to capture direct to MPEG2 - the format used to create the VOB files for your DVD, then to avoid any further conversion (rendering) and thus further loss of quality, it is best to work out in advance what the MPEG settings need to be.

If you then capture your MPEG file using those settings you can create the VOB files without any further loss of quality.

Please bear in mind that MPEG files are intended for viewing and not editing, you also need a powerful computer to be able to capture directly to MPEG.

I hope this clears up the confusion.

[Gorf]

...when capturing from my DV camcorder, it automatically sets the capture bit rate, etc and you don't need or can't make adjustments?

The latter.

Simplistically speaking (this is how I view it, anyway) your DV tape goes past the head at a constant speed, so the amount of data coming off it is constant.

You should always maintain your highest quality source, and only compress at the very last stage. This is why you edit in DV and then transcode to DVD at a later stage - the compression/recompression thing is a bit of a red herring. With features like smartrender, simple cuts result in no recompression. Remember that most operations in the editor will recompress even a DV stream, such as video filters, transitions, moving paths etc. so there is quality loss with DV as well as DVD. You're starting with a higher quality source with DV, though, so it's less likely to be noticeable.

When you use a DV source to make a DVD, you should render to DVD-compliant MPEG-2 from your DV source on the timeline, IMO.

Why?

Because if you render out to DV AVI first, and then transcode that file, you're compressing your transitions, filters etc twice - first to make the DV AVI, then next to make the MPEG. Only the bits that are green above the preview bar are not recompressed during the DV output. (Even that's not a completely reliable indicator, because you might get areas that have had a temporary render to the MSP Preview Files directory.)

If you encode the MPEG file from the timeline, those non-green parts of the timeline do not go through an intermediate DV recompression, so they maintain their quality.

[Gorf]

If you intend to edit your video then you should "capture" in DV which is a form of AVI (AVI is a term relating to several different lossless formats)

Further to Steve's post clearing things up - "Lossless" is often a confusing term. It relates to the state of the digital video on the hard drive during capture and edit.

Certain editing actions (e.g. video filters) cause the video to be recompressed, so those actions are not lossless, even on DV AVIs. A project with a highly compressed source file such as MPEG-2 may also get recompressed from simple editing actions such as cuts, if the project settings do not meet or exceed those of the source files.

Whenever I talk about lossless, I restrict it to the capture process, and it relates to the actual bytes of the captured video data being identical on the hard drive as they are on the camera media. This means that capturing analogue video, or firewire-to-MPEG etc, is "lossy".

Finally, AVI is just a "wrapper" which tells the operating system what program should be used to open the file. The only constant thing about AVI files is the location of the codec information therein, which tells the playback program (or editor, or encoder, or whatever) which codecs are to be used to decode the video and audio data. DivX data can be contained within an AVI file, and the workings of DivX are closer to DVD than DV.

If you do not intend to do any editing and your computer is powerful enough to capture direct to MPEG2 - the format used to create the VOB files for your DVD, then to avoid any further conversion (rendering) and thus further loss of quality, it is best to work out in advance what the MPEG settings need to be.

I think it's worth noting that this will generally save time and disc space, but it's not advisable if you're not sure what you're doing or if you're not conviced your PC is up to the job. If you mess up an MPEG capture, the only real solution is to capture again, so one of its benefits has gone completely.

Capturing as DV will allow you to convert the video and if something's not right, you don't have to go through the capture again to put it right. If your machine is powerful enough to capture MPEG via firewire, it is powerful enough to transcode DV to MPEG faster than realtime, meaning your subsequent attempts to get the settings right take less time than a recapture.

[Devil]

There is another very important point that is often ignored.

Because, as has been mentioned, "capturing" DV does nothing to the signal, provided your IEEE-1394 connection through the various computer components to your hard drive is reasonable, the transfer of the signal in real time is easy.

However, with most "ordinary" computers, transcoding from the DV stream to an MPEG stream in real time is simply not possible with high quality because the CPU<>Northbridge<>RAM and RAM<>HDD simply do not have enough horsepower. Want proof? OK, take 10 minutes of unedited DV signal, put it on your timeline with the green and yellow lines showing. Now set it going to Create an MPEG file with whatever settings you consider as high quality. Time the operation. It will probably take more than 10 minutes. Therefore, it is not reasonable to expect this to happen in 10 minutes, at this quality, when capturing, is it? So, what gives? Yes, quality gives. Plus the fact that you cannot possibly have some options, such as 2-pass. I'm not sure about v. 8.0, but I do know that in earlier versions, you could not transcode the audio to AC-3, either, yet this is the most "economical" format for audio.

If the computer is marginally insufficient, the difference in processing speeds may be buffered in RAM and the operation completed after the transfer is complete. A message comes up to this effect. If it is largely insufficient, then the system will baulk with an error message.

If you need high quality MPEG-2 with a computer that has less than a 3 GHz single CPU and 1 Gb RAM and a good, dedicated hard drive, I would seriously suggest you forget about transcoding.

Gorf: There is no practical quality loss in recompressing DV, you are mistaken. Many years ago, I took a DV clip and re-rendered it 20 times. I then carefully viewed the results. For casual TV viewing (e.g., like watching a captivating drama on TV), my wife, daughter and I agreed that there was no significant quality loss up to the 10th-12th generation. At the 15th generation, the artefacts started to become obtrusive and by the 20th generation, they were pretty bad. However, with careful attention, we could observe no deterioration up to the 8th generation and doing frame-by-frame direct comparisons on a computer screen, there was nothing significant in the 5th generation. It is one of the beauties of DV that it is quasi-lossless for a few re-renders, provided you keep the same settings.

[sjj1805]

A further slight clarification of capturing to MPEG.
Firstly I must stress that we all agree tht wherever possible you should capture to DV (AVI) and later (after editing) render (create) a DVD compliant MPEG2 file before proceeding to the authoring of a DVD stage.

There are times when you do not have the option to transfer to DV.
One example of this is my Hauppauge TV card. It only captures in MPEG - There is no option of capturing to any other format. The Hauppauge TV card is equipped with a hardware MPEG encoder and so is able to keep up with the conversion from a TV broadcast to an MPEG file. It is a purpose built item. In these situations you would follow the guidelines to capture to an MPEG setting that would not require any further rendering to a different setting.

I think between us we are getting there

[mmpo]

We're learning a lot, thanks to you guys with practical experience.

Another challenge:
I'm busy with a project:
1. Most of the footage I took myself and captured it in AVI as suggested.
2. Unfortunately I need to use additional footage from a DVD (already MPEG) and I "captured" it with MSPs DVD Plugin.

Now the big question(s):
1. Do I use both sources as is, mixed in the timeline, and render the final quality I need.
2. Do I convert the MPEGs to Similar AVIs?
3. Do I convert the AVIs to similar MPEGs?

This will only be a 10 min. promotional shown on DVD, so quality is not a great issue. I think I'll come right using both sources as is. Just wondering what the best would be when working on longer projects.

I must say, I've been keeping you guys thinking! But I'm sure we're learning a lot in the process.

[sjj1805]

I've mixed and matched several times with AVI/MPEG and still images all on the same timeline at the same time.
I Just render to the required DVD complaint MPEG2 format when I've completed my editing on the timeline.
The only hiccup you may have is Field Order as some items are LFF some UFF and some frame based. I did several tests and found that the best Field Order Option was to select the order of whichever clip was first - normally an AVI clip in LFF. I also right click the still images select Media Source Options and select the Flicker Reduction and in stretch mode - stretch.

Works for me.

[Gra]

This has been a very helpful thread to follow; thanks to those more experienced amongst us. I've seen some talk of compression levels and I wonder, in video editor, should I switch from 'DV Video Encoder Type 1' to 'none' or leave it as it is for best results - or will that affect instant playback, etc?

[sjj1805]

For an explanation of DV type 1 and DV type 2 please view this link
http://www.ulead.com/learning/video1/page1.htm

[Gorf]

This has been a very helpful thread to follow; thanks to those more experienced amongst us. I've seen some talk of compression levels and I wonder, in video editor, should I switch from 'DV Video Encoder Type 1' to 'none' or leave it as it is for best results - or will that affect instant playback, etc?

It will be best during editing if your project settings match the bulk of your source material. This means that if you have mostly DV with a bit of DVD MPEG in there, keep your project settings at DV.

This speeds up your workflow.

When you're happy with your edit, output from the timeline to your preferred codec, e.g. DVD-compliant MPEG-2.

Note that Brian has said you won't notice the first five recompressions of DV to DV, so even if you have a lot of video filters and/or transitions, you can get away with outputting your finished edit of DV source to a DV file, and then compress it as necessary outside MSP.

[pennstat]

Is the AVI then worth it if you can't set the bitrate and audio? (Sound given as DV Audio). Especially if you say that the final file needs to be Converted to MPG in any way for DVD and takes long to render from AVI to MPG.

DV uses a constant bit-rate for video (3MB/sec) and the audio is PCM 16-bit/48KHz/stereo. (Feel free to correct me on that if I'm wrong.) Those settings are defined standards and cannot be changed.

That being said, using AVI is far more convenient for editing because it's much faster and easier to use as source files than most codecs. Try it yourself. Capture a video using DV and then capture some more in MPEG-2, then try using both in the same project. Watch the rendering window and you'll see MSP slow down when it hits the MPEG-2 sections of the project as opposed to the AVI sections.

However, no matter what source codec you use, the resulting video will *have* to be converted to MPEG-2. Whether it's through MSP, DVD Workshop, or a third party package, it must get converted to MPEG-2 because that's what all DVD players use. (Some use other codecs, of course, but MPEG-2 is the universal DVD standard.) But using DV AVI is definitely the easier way to edit.

Rendering time also depends on a lot of factors. CPU power, of course, is the biggest, but how you render is also important. Variable bit rates take longer than constant; two-pass variable takes twic as long as normal variable; error correction and frame estimation also add rendering time. Normally what I do is edit, edit, edit and then render to MPEG-2 overnight in the highest quality possible while I'm alseep. A two-hour video is easily finished between the time that I go to sleep and get home from work, so in essense none of my time is lost waiting for the rendering to finish.


Edit: Of course, I posted this forgetting that there was an entire second page in the thread.. Oh, well.

[mmponline]

Pennstat

Thanks for your reply. I tried the AVI ones, and they do work faster.

However, when scrubbing the timeline I get a "out of memory" message after a whille, and then the scrubbing doesn't work.

I then went back to the MPEGs, as I normally work on shorter projects where rendering time is not such a big issue. Conserving harddrive space maybe more important (as my AVIs are 5 times larger).