Displaying 625/50 PAL as 625/50 NTSC - UK Vintage Radio Repair and Restoration Discussion Forum

practical · 14th May 2010, 06:38 PM

The attached pictures show a test card (generated by a DVD player from a jpg on a photo CD) displayed in two ways:

1) output as NTSC and fed to a 1980's North American Philips NTSC-only TV through the composite video input (it does not have an S-Video input).

2) output as PAL and fed through a Chinese PAL to NTSC converter bought for $25 on eBay, then displayed on the same NTSC TV. As stated by the seller, this converter does not do scan conversion, only color system transcoding.

As can be seen from the pictures, the TV (Philips P0K221) is stable at 50 Hz (not all US TVs are) and has perfect geometry for the 625 line signal.

However, the colors are not perfect. The background white is too blue and the colors in general seem a bit washed out. I am wondering about the explanation of the color issues.

The converter contains two chips, one is a Samsung S1H2198A, the other has had its identifiers ground off. No datasheet for the S1H2198A is available on the Web, but I suspect it may be a renaming or update of KA2198BD, an RGB to composite/S-Video encoder. In addition, there are two crystals, one for the correct PAL color subcarrier, another for NTSC (M). However, there does not seem to be a delay line for PAL decoding.

Hypothesis 1: If PAL decoding is done in a crude way, phase errors may be to blame, right?

Hypothesis 2: However, even if PAL had been decoded into RGB correctly with the use of a delay line, wouldn't subsequent encoding of this RGB signal into NTSC necessarily result in color distortion due to the differences in color space, white point, and gamma? Can someone predict if these differences are sufficient to explain the color shifts seen in the examples?

Hypothesis 3: As far as I understand, the NTSC subcarrier frequency of 3.579545 MHz is not ideal for 625/50 in terms of frequency interleaving because it is practically an integer multiple of the horizontal frequency of 15,625 Hz (the ideal would be in-between two integers). Could an inappropriate Fsc result in the colors being knocked out, or is the choice of Fsc for a given scan rate determined by other factors such as avoidance of diagonal patterning on monochrome receivers, etc?

Cheers,

ENGLISH VICTOR · 15th May 2010, 08:25 AM

Perhaps there is a digital delay line built into the PAL decoding section of the converter, this is common practice these days, without a delay line you would have observed some line by line saturation and phase errors, there appear to be none in your jpeg. Crude decoding of PAL S that is without a delay line will produce Hanover blinds (line by line saturation and phase errors) roughly proportional to the phase error. With PAL S one effect without the other is not possible. The delay line in PAL D permits the separation of U and V thus eliminating phase errors up to 90 deg and converts the problem into saturation errors. Beyond 90 deg saturation begins to return but with inverse hue decoding.No U,V, cross contamination is possible with PAL D.

Victor

Patrick Dixon · 15th May 2010, 02:36 PM

To answer the last question first, yes the Fsc/FL relationship was chosen to make the subcarrier dot pattern less visible and shouldn't cause this problem.

It looks more like a DC (black clamp) error on the U and/or V signals, rather than a phase shift. If you feed a mono (PAL) step signal to the converter, can you see subcarrier on the NTSC step waveform using a scope?

practical · 17th May 2010, 03:09 PM

Thanks for the suggestion, Patrick, but we don't have any testing equipment other than a .jpg of a test card on a photo CD played back in PAL or NTSC through the composite output of a DVD player.

dominicbeesley · 20th May 2010, 01:04 PM

I loaded the two pictures into a photo editing program. The first can be made right by adjusting the hue. The second by reducing the blue channel gain or by changing the B-Y. Trouble with a test card either will make it look more or less right. How about some more pairs of real life pictures!

Dom

practical · 21st May 2010, 02:06 PM

It was a really good idea to deconstruct the pictures by bringing them into a photo editing software and play around with the levels.

Unfortunately, I cannot get any real-life pictures right now as the setup is with friend in another city who is short on time for this project.

However, I am expecting delivery of another PAL-decoder/NTSC encoder combo. I intend to connect it to another 50 Hz-stable NTSC-only TV (like my friend's Philips), which I need to locate first. The difficulty in finding a 50 Hz stable US-market TV is that nobody, possibly not even the manufacturer, knows whether a particular TV is 50 Hz stable, so it is all trial and error.

A Sony I tried did not lock up on 50 Hz, and a Panasonic did lock up but displayed the 625-line picture with incorrect geometry. I suspect the fact that the Japanese manufacturers have a strong NTSC-standard home market presumably with little interest in 50 Hz standards means that they have reason to develop NTSC-only componentry and chassis. I believe a better starting point is Philips (due to its corporate base in Europe) or a Japanese TV that shares it chassis with South American NTSC, PAL-M/N versions.

The purpose of my project is mostly academic. I wanted to determine whether, in the 1980's and 90's, it would have been possible to make a simple decoder/encoder to display PAL 625/50 in its native scan format on some NTSC receivers sold to the general public in the US. At the time, the only options were to either perform combined scan and color system conversion (still expensive and low quality at the time) or to get a multisystem TV destined for an overseas market. These gray imports were sold at a premium and lacked some features found on US models such as MTS/SAP (stereo/dual language) sound.

Incidentally, a selection of used multisystem CRT televisions is now available in the local Craigslist classifieds for under $100. However, the I do not need a multisystem TV as my display needs are served by a couple of dual standard PAL/NTSC studio monitors. The studio monitors balk at 50 Hz with NTSC color, and I believe multisystem TVs would behave the same way, their design restricting them to broadcast standards, maybe with the addition of NTSC 4.43 or PAL-60 (a rare exception is Hitachi whose multisystem sets in the 2000's were specified to handle NTSC-50).

15th May 2010, 08:25 AM	#2
ENGLISH VICTOR Pentode Join Date: Feb 2008 Location: Sydney, Australia. N.S.W. Posts: 240	Re: Displaying 625/50 PAL as 625/50 NTSC Perhaps there is a digital delay line built into the PAL decoding section of the converter, this is common practice these days, without a delay line you would have observed some line by line saturation and phase errors, there appear to be none in your jpeg. Crude decoding of PAL S that is without a delay line will produce Hanover blinds (line by line saturation and phase errors) roughly proportional to the phase error. With PAL S one effect without the other is not possible. The delay line in PAL D permits the separation of U and V thus eliminating phase errors up to 90 deg and converts the problem into saturation errors. Beyond 90 deg saturation begins to return but with inverse hue decoding.No U,V, cross contamination is possible with PAL D. Victor

15th May 2010, 02:36 PM	#3
Patrick Dixon Hexode Join Date: Aug 2007 Location: Bristol Posts: 400	Re: Displaying 625/50 PAL as 625/50 NTSC To answer the last question first, yes the Fsc/FL relationship was chosen to make the subcarrier dot pattern less visible and shouldn't cause this problem. It looks more like a DC (black clamp) error on the U and/or V signals, rather than a phase shift. If you feed a mono (PAL) step signal to the converter, can you see subcarrier on the NTSC step waveform using a scope? Last edited by Patrick Dixon; 15th May 2010 at 02:41 PM.

17th May 2010, 03:09 PM	#4
practical Triode Join Date: Mar 2010 Location: Baltimore, USA Posts: 21	Re: Displaying 625/50 PAL as 625/50 NTSC Thanks for the suggestion, Patrick, but we don't have any testing equipment other than a .jpg of a test card on a photo CD played back in PAL or NTSC through the composite output of a DVD player.

20th May 2010, 01:04 PM	#5
dominicbeesley Octode Join Date: Nov 2004 Location: Hebden Bridge Posts: 1,039	Re: Displaying 625/50 PAL as 625/50 NTSC I loaded the two pictures into a photo editing program. The first can be made right by adjusting the hue. The second by reducing the blue channel gain or by changing the B-Y. Trouble with a test card either will make it look more or less right. How about some more pairs of real life pictures! Dom

21st May 2010, 02:06 PM	#6
practical Triode Join Date: Mar 2010 Location: Baltimore, USA Posts: 21	Re: Displaying 625/50 PAL as 625/50 NTSC It was a really good idea to deconstruct the pictures by bringing them into a photo editing software and play around with the levels. Unfortunately, I cannot get any real-life pictures right now as the setup is with friend in another city who is short on time for this project. However, I am expecting delivery of another PAL-decoder/NTSC encoder combo. I intend to connect it to another 50 Hz-stable NTSC-only TV (like my friend's Philips), which I need to locate first. The difficulty in finding a 50 Hz stable US-market TV is that nobody, possibly not even the manufacturer, knows whether a particular TV is 50 Hz stable, so it is all trial and error. A Sony I tried did not lock up on 50 Hz, and a Panasonic did lock up but displayed the 625-line picture with incorrect geometry. I suspect the fact that the Japanese manufacturers have a strong NTSC-standard home market presumably with little interest in 50 Hz standards means that they have reason to develop NTSC-only componentry and chassis. I believe a better starting point is Philips (due to its corporate base in Europe) or a Japanese TV that shares it chassis with South American NTSC, PAL-M/N versions. The purpose of my project is mostly academic. I wanted to determine whether, in the 1980's and 90's, it would have been possible to make a simple decoder/encoder to display PAL 625/50 in its native scan format on some NTSC receivers sold to the general public in the US. At the time, the only options were to either perform combined scan and color system conversion (still expensive and low quality at the time) or to get a multisystem TV destined for an overseas market. These gray imports were sold at a premium and lacked some features found on US models such as MTS/SAP (stereo/dual language) sound. Incidentally, a selection of used multisystem CRT televisions is now available in the local Craigslist classifieds for under $100. However, the I do not need a multisystem TV as my display needs are served by a couple of dual standard PAL/NTSC studio monitors. The studio monitors balk at 50 Hz with NTSC color, and I believe multisystem TVs would behave the same way, their design restricting them to broadcast standards, maybe with the addition of NTSC 4.43 or PAL-60 (a rare exception is Hitachi whose multisystem sets in the 2000's were specified to handle NTSC-50).