Testing DVD progressive scan (MPEG decoder + deinterlacer)

[atessadri]

Hello everybody,
I am trying to setup a simple test bed for comparing performances of several DVDs, all equipped with HDMI output and Full HD compliant 1080P (as declared by the supplier)

I bought the DVD Demystified test disc and I am playing with it.
A video is repeated using different rates, from 1.0 Mbps up to 9.8 Mbps.
Surprisingly, when I play the video on a DVD connected via HDMI to a big LCD screen, 1080P resolution, I cannot see any difference in quality for rates higer than 5.0 Mbps.
Do you think the chip (SUNPLUS SPHE8203R) mounted on the DVD is not able to handle high bit rates and it simply doesn't afford to decode completely the image ? (I am not an expert of MPEG, sorry, but I am trying to improve ...)
I can say that the link to the TV via YUV or via HDMI gives the same result. How is possible ?

About the deinterlacing function, I run few tests with black and white patterns containing thin horizontal lines. I see those lines flickering ...
How can I check if the deinterlacer is properly working ?

My idea was to capture the digital video using a device like BlackMagic Intensity Pro, but then how can I check that:
1) the MPEG elementary stream is not properly decoded
2) the deinterlacer is able or not to detect a stream in film-mode

Thank you in advance for your help.

Andrea

[Ghitulescu]

I bought the DVD Demystified test disc and I am playing with it.
A video is repeated using different rates, from 1.0 Mbps up to 9.8 Mbps.
Surprisingly, when I play the video on a DVD connected via HDMI to a big LCD screen, 1080P resolution, I cannot see any difference in quality for rates higer than 5.0 Mbps.

I don't own that DVD, so I cannot directly understand what do you mean by no difference? Are you implying that there's no room for improvement or that there is but the DVDplayer cannot pass over this threshold?

A good DVD encoder (and professional HW encoders are extremely good) can yield an almost perfect image by 6-6.5Mbps and extremely few artifacts at 9.8. This latter bitrate has been chosen as a panel of specialists found it out that increasing the bitrate over 10 will bring no further advantages.

[ramicio]

9.8 Mbit/s wasn't chosen because of anything relating to quality. 10.5 Mbit/s is the 1x DVD speed. Add some overhead to 9.8 Mbit/s and some headroom for safety and you have your 1x speed.

[Ghitulescu]

9.8 Mbit/s wasn't chosen because of anything relating to quality.

9.8 Mbps was selected as being a sort of threshold, the experts panel considered that increasing the bitrate over this value will give little improvement for the image quality.

You probably think of VCD, where the bitrate was chosen to fit 80 min of film, and since it was still low, they dropped the third error correction layer, too.

[ramicio]

No, it was chosen because the DVD mechanism reads at 10.5 Mbit/s. You also have container overhead and a safety net in case of scratches, to keep the buffer full. DVD specification limits bitrate to 9.8 Mbit/s for a physical reason. Surely there are instances where the picture CAN and WILL benefit from much more than 9.8 Mbit/s, but the DVD mechanism cannot read faster than that, so they imposed a limit.

Oh, correction, it's 10.8 Mbit/s.

[Ghitulescu]

Of course the reading speed played an important part in the equation, but if the drive would have allowed only 5 Mbps (for the sake of argumentation), the maximum would not have been set at 6Mbps, but the drive would have been improved. And now it reads even 24x (at least at margins). Decisive was the quality of the image.

[ramicio]

No, it is limited to 9.8 Mbit/s because of the drive's physical capability to read faster than that. What part of 1x do you not understand? Just because there are drives that read at 24x doesn't mean anything. Bit rate caps on video are ALWAYS based off the media's 1x speed. Video is averaged at ~4-5 Mbit/s anyway... If the drive would have been limited to 5 Mbit/s and they felt the quality was bad they would have not redesigned the physical media around that, they would have adopted a better compression algorithm. Why do you think Blu-ray started out with MPEG-2 on ALL titles and are now mostly h.264 with a few VC-1 titles here and there?

[Ghitulescu]

Because that HD material was already on MPEG-2 format (MPEG-2 was used mainly for documentaries, commercials/demos and stuff). It was sent through air as DVB-S (old norm, not DVB-S2), it was displayed using proprietary devices at various showrooms and fairs, and so on. On the other hand, the films were scanned, precessed and easily reencoded, as they are stored uncompressed.

[ramicio]

MPEG-2 for HD broadcast was used for just that, not archival things. It's encoded on the fly, too, with a sacrifice of quality. They even used h.264 in other countries for HD broadcast before Blu-ray even came out...so your point makes no sense.

[Ghitulescu]

MPEG-2 for HD broadcast was used for just that, not archival things. It's encoded on the fly, too, with a sacrifice of quality. They even used h.264 in other countries for HD broadcast before Blu-ray even came out...so your point makes no sense.

All the nice coloured demos (birds, landscapes, insects, country specials, tropical fishes etc.) they use to run on TVs in a loop to sell them are MPEG-2 files. MPEG-2 is not that bad, just a bit outdated, ie newer codecs are much efficient at the same bitrate.

@atessadri:
what device does the upscaling and how?

[ramicio]

So you believe the archive of those video are in MPEG-2? You are still broadcasting something if you are playing it. You're broadcasting it from the player to the TV.

[Ghitulescu]

So you believe the archive of those video are in MPEG-2? You are still broadcasting something if you are playing it. You're broadcasting it from the player to the TV.

I have seen the demo bluray disks for god sake, they are MPEG-2, all of them, Sharp, Pioneer, Panasonic, JVC ...

[ramicio]

Yes, because they came out when Blu-ray came out... It was all the Blu-ray standard allowed then. They also don't need to try to dazzle people so much by Blu-ray anymore, so they aren't going to release new demo discs with a new video format for free. I really don't understand your point here...or most any of your points for that matter. If some new standard for cinema came out, like an upped FPS, then you would be damn well sure that the Blu-ray demo discs would now be h.264.

[Ghitulescu]

It's highly off-topic, but I'll try to explain you once and for all:

Long before BD was launched, there were HD TVs and HD transmissions. I live in Europe, I don't know how the things were in the States, or in Japan. To sell those TVs that had an awful image with SD aerial/cable senders, most manufacturers had a proprietary device that fed the TV with HD images (Philips has one, Sharp had another one, Samsung too, Sony had the prototype - cartridged Bluray, and so on). They were MPEG-2. Some shops ran HD shows/loops from satellite, all of them being at that time MPEG-2, and sent using the old DVB-S protocol. Several years later, the same demos were put onto the Bluray disks, as now a unified standard exists. That's the story with MPEG-2 and HD. After movies begin to represent the bulk of the HD content, the other 2 codecs gained importance.

[ramicio]

The videos are not stored as mpeg-2, they are archived losslessly. The same thing that is broadcast is not the same as is what's on the disc. Anything that is broadcast is encoded on the fly, quite inefficiently too. They used mpeg-2 on the blu-ray demos because that's the format that was available to blu-ray at the time. They used mpeg-2 on this magical device and for broadcasts before that because that's what was majorly broadcast. The us still uses mpeg-2 for broadcast for fuck's sake. There were a few channels back then that did use h.264, but it wasn't widely adopted yet.

My point is: Media is first. From there they decide a read speed from the size versus a common title length, 90 minutes, then they choose a video codec. Mpeg-2 was considered decent and was supported more than anything else in the world. Avc took over once more computational power and hardware support came along.

The post was supposed to be in all caps because of how pissed I am. They don't base whole standards around a few demos in some retail stores.

[atessadri]

Hello guys,
thanks a lot for having started a nice discussion on the origin of the DVD.
Since you mentioned that MPEG-2 is fine and that in every shop the demo played on big LCD screen are all MPEG-2, I am really wondering why my DVD connected to a big Panasonic LCD screen is so bad.
How can I test if the deinterlacer is properly working ? or if the MPEG-2 at 9.8 Mbps is completely decoded ?
I mean, if I connect the DVD via composite video or via HDMI connector I don't see any difference.

Could you please help me to setup a proper test bed to compare several DVDs ?

Thanks again for your interest.
Sincerely

Andrea

[ramicio]

Maybe it's just the DVD source that is bad. There is no such thing as not fully decoding...there is...it's called no picture whatsoever, or totally corrupt (like when a set top box loses a bit of signal you see corruption.)

[Ghitulescu]

Bad is a dog that bites, bad is a dog that doesn't eat, bad is a dog that s**ts in the neighbour's garden ....

I asked you once, what is bad, what is no improvement, and remember that Pannies always have to be set for progressive stuff.