Need help with understanding chroma in mainconcept

[Reggie]

Hello all, I have been trying out capturing and found MainConcept mpeg encoder software to be very good for capturing from with Geforce FX5600 vivo card. I need a little help understanding some settings in the software, and possibly the hardware. I have been reading up so I just need to confirm some things.

On a capture that I did, at the start of the capture the scene is of a star field panning left to right, which should be black with some white bits, looking at it on some some equipment it seems to be quite blocky in the black bits and also has a dark green hue to it. not what I expected, as the rest of the 108 minute capture was fine.

As I'm in the UK, I captured to mpeg2 using the PAL DVD setting in the software, looking on the web for a possible reason for this blockiness and green, I came across chroma settings, I've checked the capture file and it is using 4:2:0, I am using a composite cable to capture with, and if I understand correctly composite and mpeg2 need 4:2:2 to capture correctly?

Am I getting a green hue because the green is derived from the black/white and the red, blue? or am I barking up the wrong tree?

I can find the chroma setting in mainconcept, but it doesn't allow me to change it to 4:2:2, can anyone tell me why?

[SaSi]

Quite a few points you raise in your post.

To answer your question about chroma space (4:2:2 vs 4:2:0), Mainconcept won't let you use 4:2:2 as you probably (and correctly) specified MPEG encoding at Main Profile @ Main Level, which is using 4:2:0 only.

Further, the color space used has nothing to do with incorrect color presentation.

Black screen and starfields are one of the hardest things to encode. I've spent countless hours experimenting with some Star Trek introduction clips trying to understand Tmpgenc settings.

Regarding capturing, although Mainconcept v.1.4 introduced a really handy and good working capture facility, and although it allows you to capture in MPEG-2, this is not advisable for the following reasons.

1. MPEG encoding is very CPU intensive and capturing at full frame (720x576) and doing good quality encode are two contradicting things. You will either drop frames (as your CPU will choke) or use poor quality settings so as not to drop frames.

2. Good quality MPEG-2 encoding for long movies (more than 60minutes) you will put on DVD needs to benefit from 2 pass enoding during which the encoder will first scan the video to allocate bitrate settings and on second pass will actually produce the MPEG stream. Obviously you lose this with MPEG capture.

I would suggest that you download and use hufyuv codec for capturing. Try the black starfield part of the video with this codec. If it captures OK, then your problem was caused by poor (low bitrate allocation) MPEG encoding. If not, then play with the hue settings provided by Mainconcept's capture utility to correct the capture settings. Normally, you should use the default settings with good results.
Then, after - optionally - editing and trimming the captured file, encode with Mainconcept doing a two pass encode.

[Reggie]

Hi SaSi, thanks for the reply

I'm wondering if I chose the right setting in mainconcept, I'm recording from a Satellite digital TV box via a composite cable, and as I understand things, 4:2:2 is what I need, I chose PAL DVD as the 'video type' setting as this was going to be burnt to DVD when finished, would I have been better off matching the 'video type' to the source type which would be PAL DVB? Would this make a difference at all?

I know that PAL DVB will give me the 4:2:2 setting, and I understand that this has nothing to do with the 'green tinge', I'm thinking that this colour thing may be an issue with the monitor rather than my encode, as it seems to look black on every other device I've looked at it on.

Getting on to the 1st of the 2 points about encoding, what would be acceptable settings for mpeg capture?

I have an AMD 2800+ XP processor, 2x512MB of kingston hyperX low latency ram, and as I'm capturing to mpeg, HDDs aren't really an issue as it's a pretty low data rate.

as for your 2nd point, yup, I agree that 2 passes would be impossible to implement. Originally I had a 40GB drive for capture, I put an 80GB in the machine instead last night, so I should be able to capture plenty of AVI data without worrying about running out of disc space.

Saying that though, the starfield seemed to be the only iffy part of the capture, the rest of the 108 minutes is fine, even in high action scenes, so it would be nice to capture directly to mpeg2 if I could because it means less work for me to get it to DVD-R if I want to.

One more question for you, does quality of source make quite a difference to encoding? for example, I am using composite into my PC, but if I bought a scart RGB to S-video adapter (a proper one, not a cheap £3 thing) I could use S-video, would this make a significant difference in capturing?

[SaSi]

One more question for you, does quality of source make quite a difference to encoding? for example, I am using composite into my PC, but if I bought a scart RGB to S-video adapter (a proper one, not a cheap £3 thing) I could use S-video, would this make a significant difference in capturing?

I have not been able to really see a difference between a high quality composite cable connection and a high quality S-Video connection between my VHS VCR and the PC. I do capture using S-Video, (why not), but would not really worry about link quality.

I have also tried the Scart to S-Video cable I have as well as the Scart to composite video and audio cables. Again, for VHS content, I can't say I see any difference.

The main difference between composite and S-Video is that the video signal is not "multiplexed" (or better said composed) into a single carrier cable, but transmitted in two elements, chrominance and luminance. Theoretically, this certainly improves quality, however the output from any normal VHS tape, especially if it is not SVHS, let alone LP or EP recording, doesn't really "outperform" the bandwidth of the composite cable.

Perhaps, if you directly record from the TV-Tuner, the S-Video connection will produce better results compared to composite.

[Reggie]

I'm actually coming straight out of my Sky digital satellite box which is outputting Scart-RGB, not through a video Which is why I want to spend money on a little converter, about £70 I think, which will do scart-RGB to S-Video.

Any thoughts on the rest of my last post? you've been very helpful already, I don't get too much time to experiment since I started working, which is why I'm asking as many questions as I can think of.

[SaSi]

Any thoughts on the rest of my last post? you've been very helpful already, I don't get too much time to experiment since I started working, which is why I'm asking as many questions as I can think of.

Sorry but we said too much and I lost track of the rest of the questions. What were they?

[Reggie]

I just cut and paste it from the previous post........



I'm wondering if I chose the right setting in mainconcept, I'm recording from a Satellite digital TV box via a composite cable, and as I understand things, 4:2:2 is what I need, I chose PAL DVD as the 'video type' setting as this was going to be burnt to DVD when finished, would I have been better off matching the 'video type' to the source type which would be PAL DVB? Would this make a difference at all?

I know that PAL DVB will give me the 4:2:2 setting, and I understand that this has nothing to do with the 'green tinge', I'm thinking that this colour thing may be an issue with the monitor rather than my encode, as it seems to look black on every other device I've looked at it on.

Getting on to the 1st of the 2 points about encoding, what would be acceptable settings for mpeg capture?

I have an AMD 2800+ XP processor, 2x512MB of kingston hyperX low latency ram, and as I'm capturing to mpeg, HDDs aren't really an issue as it's a pretty low data rate.

[SaSi]

I think these questions were answered in my first response.

In a quick recap, the color space cannot be 4:2:2 as MPEG2 doesn't support that for Main Profile @ Main Level, (i.e. DVD profile) let alone for SVCD or VCD. Also, I wouldn't recomment a capture to MPEG2 directly with a software encoder, even if you have an ultra fast PC.

I am using a P4/2.8HT clocked to 2.9. Although it encodes single pass full DVD resolution at slightly higher than realtime speed, trying to capture with MC at MPEG2 even at a high and constant bitrate failed. There were times when CPU utilization was briefly so high that a few frames had to be dropped. This spoils the whole process.

Capture with a codec that requires low CPU usage to ensure that you don't loose frames. Then edit and encode to MPEG-2.

Also please read my first response for some more explanation on some sub-issues of your question. And if something is still unclear, please ask again.