Discussion:
Display artifacts on //e
(too old to reply)
Doug Dingus
2020-09-16 05:10:23 UTC
Permalink
For the most part, I didn't pay a whole lot of attention to these, but on my //e I see what looks like extra pixels, or echos of pixels.

This image of some text shows it pretty well:

https://imgur.com/vCkjxl8

And here's another one from the Total Replay title screen:

https://imgur.com/bdomSch

The teardrop shaped object near center should look more like this:

https://imgur.com/1MQbeJr

Or, should it?

I was just taking some close up shots for someone to see what Apple artifact color really looks like off a CRT and it seems odd to me.

And it definitely is!

Here's one last shot from a simple monochrome amber screen that seems to align with my memory of how a color image should look:
https://imgur.com/CmiY34U

Are any of you running PVM displays with your Apples?

This isn't a big deal because things look really great overall. But, it's odd and seems like the very out of spec Apple signal is exposing some of the advanced filtering these sets can do.

I'm mostly curious about your experiences with modern, higher end CRT displays, if anyone is even using one. I like a CRT and scored mine while they were inexpensive.
b***@gmail.com
2020-09-16 18:05:00 UTC
Permalink
Hi Doug,

I think the ghosting or shadowing in the images you posted is caused by a
poor analog cable. If the impedence is not well matched then this will
causes signal reflections, which can manifest themselves as ghost images
offset horizontally from the actual pixels that are illuminated.

Try a different cable and see if it makes any difference to the image
quality.

All the best,
Doug Dingus
2020-09-17 04:40:53 UTC
Permalink
Post by b***@gmail.com
Try a different cable and see if it makes any difference to the image
quality.
Oh, I should know better! Excellent suggestion. I'm not sure that's the entire issue as there are pixels below! But it's something to start with.

More when something new happens.
Michael J. Mahon
2020-09-17 05:08:13 UTC
Permalink
Post by Doug Dingus
For the most part, I didn't pay a whole lot of attention to these, but on
my //e I see what looks like extra pixels, or echos of pixels.
https://imgur.com/vCkjxl8
https://imgur.com/bdomSch
https://imgur.com/1MQbeJr
Or, should it?
I was just taking some close up shots for someone to see what Apple
artifact color really looks like off a CRT and it seems odd to me.
And it definitely is!
Here's one last shot from a simple monochrome amber screen that seems to
https://imgur.com/CmiY34U
Are any of you running PVM displays with your Apples?
This isn't a big deal because things look really great overall. But,
it's odd and seems like the very out of spec Apple signal is exposing
some of the advanced filtering these sets can do.
I'm mostly curious about your experiences with modern, higher end CRT
displays, if anyone is even using one. I like a CRT and scored mine
while they were inexpensive.
This is not an artifact of Apple video. It is a result of poor convergence
adjustment of your color monitor.

Proper adjustment of color monitor convergence is non-trivial, involving
multiple interacting adjustments to static and dynamic convergence.

That said, your photos seem to show that the green electron beam is
statically misconverged, since the misconvergence appears to be the same at
all screen positions. The static convergence adjustment is most likely a
moveable magnet near the neck of the CRT by the green electron gun.
--
-michael - NadaNet 3.1 and AppleCrate II: http://michaeljmahon.com
Doug Dingus
2020-09-17 18:26:09 UTC
Permalink
I am actually quite skilled in CRT setup and convergence, purity, etc. Back in the day, I got a fair chunk of my Apple computer paid for by doing these services using a Atari computer and cassette loaded with programs to generate the necessary patterns. Did a lot of 60's, 70's and a few 80's era TVs. Most of them would serve as respectable monitors, with some 60's era TV's requiring the occasional trade off, if used for that purpose.

You are quite right about the whole affair being non trivial!

Side note:

The Atari machines will do a full screen display, 48 byte DMA and had grey scales, all the stuff one needs to do a reasonable setup and calibration. I tweaked the Atari a little to get its color reference close to standard and the reset was more than close enough for the caliber of TV in question.

End side note.

Anyway, look at the total replay picture. There isn't any gross error that I can see. When rendering 80 column text, some minor league misalignment can be seen near screen extents, but nothing like the green text image might imply. I plan on a fine adjustment soon. The entire frame is rotated just enough to annoy me, but due to the fine pitch CRT, a current non issue.

I will grab another photo of it with the chroma turned off maybe later today. It really seems the set is adding some detail. I do need to match a cable to it better. I have my doubts about it being an issue, but it really could be, though that does not explain the pixels below a given row of pixels...

This is my first high end pro grade CRT. And I am quite impressed with it. The fine pitch CRT makes a big difference of course, but overall it really does deliver better than the best consumer grade sets I have refurbished and calibrated. Some of those are close though, and I suspect a better grade CRT would close the gap for all but the most demanding cases, and for sure anything retro computing related.

This happening is just odd, and not something I have seen before.
Post by Michael J. Mahon
Proper adjustment of color monitor convergence is non-trivial, involving
multiple interacting adjustments to static and dynamic convergence.
That said, your photos seem to show that the green electron beam is
statically misconverged, since the misconvergence appears to be the same at
all screen positions. The static convergence adjustment is most likely a
moveable magnet near the neck of the CRT by the green electron gun.
--
-michael - NadaNet 3.1 and AppleCrate II: http://michaeljmahon.com
Michael J. Mahon
2020-09-17 19:24:28 UTC
Permalink
Post by Doug Dingus
I am actually quite skilled in CRT setup and convergence, purity, etc.
Back in the day, I got a fair chunk of my Apple computer paid for by
doing these services using a Atari computer and cassette loaded with
programs to generate the necessary patterns. Did a lot of 60's, 70's and
a few 80's era TVs. Most of them would serve as respectable monitors,
with some 60's era TV's requiring the occasional trade off, if used for that purpose.
You are quite right about the whole affair being non trivial!
The Atari machines will do a full screen display, 48 byte DMA and had
grey scales, all the stuff one needs to do a reasonable setup and
calibration. I tweaked the Atari a little to get its color reference
close to standard and the reset was more than close enough for the
caliber of TV in question.
End side note.
Anyway, look at the total replay picture. There isn't any gross error
that I can see. When rendering 80 column text, some minor league
misalignment can be seen near screen extents, but nothing like the green
text image might imply. I plan on a fine adjustment soon. The entire
frame is rotated just enough to annoy me, but due to the fine pitch CRT,
a current non issue.
I will grab another photo of it with the chroma turned off maybe later
today. It really seems the set is adding some detail. I do need to
match a cable to it better. I have my doubts about it being an issue,
but it really could be, though that does not explain the pixels below a
given row of pixels...
This is my first high end pro grade CRT. And I am quite impressed with
it. The fine pitch CRT makes a big difference of course, but overall it
really does deliver better than the best consumer grade sets I have
refurbished and calibrated. Some of those are close though, and I
suspect a better grade CRT would close the gap for all but the most
demanding cases, and for sure anything retro computing related.
This happening is just odd, and not something I have seen before.
Post by Michael J. Mahon
Proper adjustment of color monitor convergence is non-trivial, involving
multiple interacting adjustments to static and dynamic convergence.
That said, your photos seem to show that the green electron beam is
statically misconverged, since the misconvergence appears to be the same at
all screen positions. The static convergence adjustment is most likely a
moveable magnet near the neck of the CRT by the green electron gun.
--
-michael - NadaNet 3.1 and AppleCrate II: http://michaeljmahon.com
The fact that the monochrome monitor displays the video properly rules out
any video signal problem, including the cable.

Looking at the color text display, it’s clear that the green gun is
displaying about a pixel’s width to the left and down from the red and blue
guns.

This is a convergence problem.

And while you have the case off, a slight yoke twist will fix the raster
rotation error.

It should be a pretty straightforward fix for you since you have
experience.
--
-michael - NadaNet 3.1 and AppleCrate II: http://michaeljmahon.com
Doug Dingus
2020-09-18 03:57:04 UTC
Permalink
On Thursday, September 17, 2020 at 12:24:35 PM UTC-7, Michael J. Mahon wrote:
(snip)
Post by Michael J. Mahon
This is a convergence problem.
And while you have the case off, a slight yoke twist will fix the raster
rotation error.
It should be a pretty straightforward fix for you since you have
experience.
Respectfully, it's not. I know how to deal with those. This particular PVM has a small error and does need the yoke tweaked just a shade, but that isn't what I'm asking about here.

Frankly, I'm making a little test signal generator for the purpose of tuning up a few CRT's I've accumulated. Should be fun, and this one will get the treatment. I'm holding off, because it's a better display than an old Atari can calibrate. Rotation issue aside, I would not get better than it is right now with that signal. I've always wanted to tune a very fine pitch device with good circuits with a really precise and high resolution signal, but that is also an aside.

Here is the same image with the chroma turned off, and the artifacts are still there:

https://imgur.com/bWtIs0C

And another one with some more filled in color areas, where that pattern is harder to see.

https://imgur.com/dO3lS3b
Doug Dingus
2020-09-18 04:13:41 UTC
Permalink
On Thursday, September 17, 2020 at 12:24:35 PM UTC-7, Michael J. Mahon wrote:
(snip)
Post by Michael J. Mahon
The fact that the monochrome monitor displays the video properly rules out
any video signal problem, including the cable.
I agree with you, tried a variety of cables, and the //e is fine obviously. I haven't seen these in the past on other displays.

I'm going to inquire among the high end CRT geeks too. See what they have to say.
Doug Dingus
2020-09-18 18:02:57 UTC
Permalink
Post by Doug Dingus
(snip)
Post by Michael J. Mahon
The fact that the monochrome monitor displays the video properly rules out
any video signal problem, including the cable.
I agree with you, tried a variety of cables, and the //e is fine obviously. I haven't seen these in the past on other displays.
I'm going to inquire among the high end CRT geeks too. See what they have to say.
Tiny update: This appears to be something to do with how this CRT processes signals and the basic nature of the Apple signal. I am going to connect other retro computers to see whether it's present.

I did try a consumer grade TV I am aligning right now and these artifacts aren't present. Weird.
Tempest
2020-09-19 12:39:56 UTC
Permalink
When I hooked my 12" PVM up to my Apple II I thought it looked kind of odd. I don't recall having the display artifacts like you did, but the actual artifacting (extra composite colors) looked off. I wonder if the Apple brand monitors were specially tuned for the Apple somehow because no other monitor I've used ever looked as nice as the colors do on my Apple Color Composite monitor.
Michael J. Mahon
2020-09-20 00:55:37 UTC
Permalink
Post by Tempest
When I hooked my 12" PVM up to my Apple II I thought it looked kind of
odd. I don't recall having the display artifacts like you did, but the
actual artifacting (extra composite colors) looked off. I wonder if the
Apple brand monitors were specially tuned for the Apple somehow because
no other monitor I've used ever looked as nice as the colors do on my
Apple Color Composite monitor.
The only thing “special” about the Apple Color Composite monitor is its
monochrome mode, which expands the luminance bandwidth to double the 3MHz
bandwidth of color mode.

Analog NTSC monitor signal processing is pretty simple, and doesn’t have
any way to create vertically displaced artifacts (like digital processing
can).
--
-michael - NadaNet 3.1 and AppleCrate II: http://michaeljmahon.com
Doug Dingus
2020-09-20 02:10:31 UTC
Permalink
Post by Michael J. Mahon
Post by Tempest
When I hooked my 12" PVM up to my Apple II I thought it looked kind of
odd. I don't recall having the display artifacts like you did, but the
actual artifacting (extra composite colors) looked off. I wonder if the
Apple brand monitors were specially tuned for the Apple somehow because
no other monitor I've used ever looked as nice as the colors do on my
Apple Color Composite monitor.
The only thing “special” about the Apple Color Composite monitor is its
monochrome mode, which expands the luminance bandwidth to double the 3MHz
bandwidth of color mode.
Analog NTSC monitor signal processing is pretty simple, and doesn’t have
any way to create vertically displaced artifacts (like digital processing
can).
--
-michael - NadaNet 3.1 and AppleCrate II: http://michaeljmahon.com
I agree. This set is doing some processing. I need more technical detail on it. When I run a DVD player, it is insane how great it looks!

@tempest: I have never seen the Apple monitor. This PVM is amazing. Those artifacts aren't a big deal. I am more curious than anything.

If the special is a wide open bandwidth, then I may be seeing close to what that monitor can deliver as the PVM can go real wide. They claim 600 plus lines. I have yet to send a signal into it to test. My little amber screen will do that plus some.
Tempest
2020-09-20 13:41:43 UTC
Permalink
Maybe the issue was the sharpness then? PVMs by their very nature are made to be sharp as they were supposed to be used in professional video settings which is why all the CRT-ophiles want them (I own a few myself). However since the Apple II relied on composite color bleed and artificating to for its graphics, super sharp would mess that up. Maybe it wasn't so much that Apple monitors were 'tuned' for the Apple II, but that their consumer quality worked well with the Apple II?

I should hook my PVM up and take some pictures. Like I said it's not bad, just 'off' compared to the Apple monitor.
Doug Dingus
2020-09-20 16:34:53 UTC
Permalink
Maybe the issue was the sharpness then? PVMs by their very nature are made to be sharp as they were supposed to be used in professional video settings which is why all the CRT-ophiles want them (I own a few myself). However since the Apple II relied on composite color bleed and artificating to for its graphics, super sharp would mess that up. Maybe it wasn't so much that Apple monitors were 'tuned' for the Apple II, but that their consumer quality worked well with the Apple II?
I should hook my PVM up and take some pictures. Like I said it's not bad, just 'off' compared to the Apple monitor.
Pictures and what PVM you have would be interesting. Might just be color reference differences too. I would like to see a detail photo of both myself.
Tempest
2020-09-20 22:06:10 UTC
Permalink
It was a PVM 14N1U. I'll dig it out this week if I get a chance and take a few pictures. It might very well have just been the color adjusts, who knows?
Doug Dingus
2020-09-21 02:04:46 UTC
Permalink
It was a PVM 14N1U. I'll dig it out this week if I get a chance and take a few pictures. It might very well have just been the color adjusts, who knows?
Cool. Looking forward to it. Mine is a SONY 1342Q. Has TTL instead of component input.

Soon, I will have a keyboard and mouse for my GS. Will try the RGB to see what that does.
Tempest
2020-09-21 13:04:46 UTC
Permalink
Soon, I will have a keyboard and mouse for my GS. Will try the RGB to see what that does.
Different kind of RGB. PVMs like the 1342Q use Digital RGB while the IIgs uses Analog RGB, so it won't work.
Doug Dingus
2020-09-21 14:54:40 UTC
Permalink
Soon, I will have a keyboard and mouse for my GS. Will try the RGB to see what that does.
Different kind of RGB. PVMs like the 1342Q use Digital RGB while the IIgs uses Analog RGB, so it won't work.
This particular one has both. And the GS has one of those PC cards in it. Who knows, I may use both inputs!
Thomas Harte
2020-09-24 18:48:15 UTC
Permalink
Maybe the issue was the sharpness then? PVMs by their very nature are made to be sharp as they were supposed to be used in professional video settings which is why all the CRT-ophiles want them (I own a few myself). However since the Apple II relied on composite color bleed and artificating to for its graphics, super sharp would mess that up. Maybe it wasn't so much that Apple monitors were 'tuned' for the Apple II, but that their consumer quality worked well with the Apple II?
For whatever it's worth, this would definitely be my guess. If you look at the coefficients of a digital high-pass filter they usually look a bit like a Gaussian in the middle but then dip below zero at the sides; that creates a risk of artefacts that aren't physically joined to their source since the amplitude of any given sample's input goes up, then back to zero, then back up again as they traverse from the edge to the middle of the window.
Doug Dingus
2020-09-24 21:47:16 UTC
Permalink
Post by Thomas Harte
Maybe the issue was the sharpness then? PVMs by their very nature are made to be sharp as they were supposed to be used in professional video settings which is why all the CRT-ophiles want them (I own a few myself). However since the Apple II relied on composite color bleed and artificating to for its graphics, super sharp would mess that up. Maybe it wasn't so much that Apple monitors were 'tuned' for the Apple II, but that their consumer quality worked well with the Apple II?
For whatever it's worth, this would definitely be my guess. If you look at the coefficients of a digital high-pass filter they usually look a bit like a Gaussian in the middle but then dip below zero at the sides; that creates a risk of artefacts that aren't physically joined to their source since the amplitude of any given sample's input goes up, then back to zero, then back up again as they traverse from the edge to the middle of the window.
I am not so sure it would.

In my past, I've tuned some TV's up for computer use. Frankly, I've gotten results on par with both my current PVM and Tempests Apple Color Monitor. Doing that can take modifying the TV some, opening up bandwidth in a place or two, and obviously it takes a good set with the bones to perform, but given those things are well executed, sharp doesn't do what is being seen on either PVM set.

These PVM's, maybe not all of them, but the two on this thread are processing the signal beyond simply executing a CRT display circuit well.

On mine at least, the raw bandwidth is totally there and then some. That set will knock out almost 800 lines. I've sent tests into it. What I need to do next is use a couple other machines. I have an 800XL that generates a "real" color signal. Very curious to see what it does with that one. The GS color signal is a mixed bag. Need to look at it again.

In monochrome, the GS is pixel perfect on that PVM, BTW. Goregous!

I think if I had, what I think is a digital processor, out of the look in my PVM, the result would basically be the Apple Color Monitor. The photos here compare almost spot on.

What actually happens as one increases the "sharpness" of an NTSC composite display, is the color bleed goes away completely. One sees the vertical bands, and a single pixel ends up the same size it does on a monochrome screen, and it takes on the color it does due to it's time position relative to the colorburst reference cycle. On very sharp sets, and this can be seen on my PVM, artifacts aside, is a red pixel will have a literally different position on the screen than say a green, or blue one does.

The Apple Color Monitor is opened wider than most consumer grade sets are, though later ones, say 90's era and above can perform at near that level with their sharpness cranked up.

If anyone is looking, the SONY WEGA CRT's can totally do that, especially the ones offering component input. The basic, well executed CRT circuit and sharpness are there.

I'll try and get a photo.

The other thing I'm going to try is the s-video input. Can split the Apple output and run one branch through a resister and connect both to the chroma. Sometimes this plays out very differently on some sets.

Sure has been interesting so far. I didn't expect any of these outcomes in the PVM scene. :D
Thomas Harte
2020-09-25 15:02:50 UTC
Permalink
Post by Doug Dingus
Post by Thomas Harte
Maybe the issue was the sharpness then? PVMs by their very nature are made to be sharp as they were supposed to be used in professional video settings which is why all the CRT-ophiles want them (I own a few myself). However since the Apple II relied on composite color bleed and artificating to for its graphics, super sharp would mess that up. Maybe it wasn't so much that Apple monitors were 'tuned' for the Apple II, but that their consumer quality worked well with the Apple II?
For whatever it's worth, this would definitely be my guess. If you look at the coefficients of a digital high-pass filter they usually look a bit like a Gaussian in the middle but then dip below zero at the sides; that creates a risk of artefacts that aren't physically joined to their source since the amplitude of any given sample's input goes up, then back to zero, then back up again as they traverse from the edge to the middle of the window.
I am not so sure it would.
In my past, I've tuned some TV's up for computer use. Frankly, I've gotten results on par with both my current PVM and Tempests Apple Color Monitor. Doing that can take modifying the TV some, opening up bandwidth in a place or two, and obviously it takes a good set with the bones to perform, but given those things are well executed, sharp doesn't do what is being seen on either PVM set.
These PVM's, maybe not all of them, but the two on this thread are processing the signal beyond simply executing a CRT display circuit well.
On mine at least, the raw bandwidth is totally there and then some. That set will knock out almost 800 lines. I've sent tests into it. What I need to do next is use a couple other machines. I have an 800XL that generates a "real" color signal. Very curious to see what it does with that one. The GS color signal is a mixed bag. Need to look at it again.
In monochrome, the GS is pixel perfect on that PVM, BTW. Goregous!
I think if I had, what I think is a digital processor, out of the look in my PVM, the result would basically be the Apple Color Monitor. The photos here compare almost spot on.
What actually happens as one increases the "sharpness" of an NTSC composite display, is the color bleed goes away completely. One sees the vertical bands, and a single pixel ends up the same size it does on a monochrome screen, and it takes on the color it does due to it's time position relative to the colorburst reference cycle. On very sharp sets, and this can be seen on my PVM, artifacts aside, is a red pixel will have a literally different position on the screen than say a green, or blue one does.
The Apple Color Monitor is opened wider than most consumer grade sets are, though later ones, say 90's era and above can perform at near that level with their sharpness cranked up.
If anyone is looking, the SONY WEGA CRT's can totally do that, especially the ones offering component input. The basic, well executed CRT circuit and sharpness are there.
I'll try and get a photo.
The other thing I'm going to try is the s-video input. Can split the Apple output and run one branch through a resister and connect both to the chroma. Sometimes this plays out very differently on some sets.
Sure has been interesting so far. I didn't expect any of these outcomes in the PVM scene. :D
Then the only other thing I can think of that may be helpful is that if you're planning to test other systems to verify whether the unnatural harshness of the Apple's signal is at fault: the NES also uses a square wave for its colour subcarrier. It's not artefact colour, it's more like the Atari in that colour is generated via amplitude + phase, but the colour part is square rather than a sine.

So contrasting those two, if you have access to a Nintendo, might help with with further diagnosis. Also I'm not really a Nintendo person so I assume they come with direct composite outputs but I don't actually know. Obviously if they're RF-only then maybe that's not so helpful.
Doug Dingus
2020-09-25 22:28:52 UTC
Permalink
Post by Thomas Harte
Post by Doug Dingus
Sure has been interesting so far. I didn't expect any of these outcomes in the PVM scene. :D
Then the only other thing I can think of that may be helpful is that if you're planning to test other systems to verify whether the unnatural harshness of the Apple's signal is at fault: the NES also uses a square wave for its colour subcarrier. It's not artefact colour, it's more like the Atari in that colour is generated via amplitude + phase, but the colour part is square rather than a sine.
So contrasting those two, if you have access to a Nintendo, might help with with further diagnosis. Also I'm not really a Nintendo person so I assume they come with direct composite outputs but I don't actually know. Obviously if they're RF-only then maybe that's not so helpful.
Yeah, good call. I don't have a NES, but I can generate some signal types. A C64 would be interesting too.
Michael J. Mahon
2020-10-04 22:01:23 UTC
Permalink
Post by Thomas Harte
Post by Doug Dingus
Post by Thomas Harte
Maybe the issue was the sharpness then? PVMs by their very nature are
made to be sharp as they were supposed to be used in professional
video settings which is why all the CRT-ophiles want them (I own a few
myself). However since the Apple II relied on composite color bleed
and artificating to for its graphics, super sharp would mess that up.
Maybe it wasn't so much that Apple monitors were 'tuned' for the Apple
II, but that their consumer quality worked well with the Apple II?
For whatever it's worth, this would definitely be my guess. If you look
at the coefficients of a digital high-pass filter they usually look a
bit like a Gaussian in the middle but then dip below zero at the sides;
that creates a risk of artefacts that aren't physically joined to their
source since the amplitude of any given sample's input goes up, then
back to zero, then back up again as they traverse from the edge to the
middle of the window.
I am not so sure it would.
In my past, I've tuned some TV's up for computer use. Frankly, I've
gotten results on par with both my current PVM and Tempests Apple Color
Monitor. Doing that can take modifying the TV some, opening up bandwidth
in a place or two, and obviously it takes a good set with the bones to
perform, but given those things are well executed, sharp doesn't do what
is being seen on either PVM set.
These PVM's, maybe not all of them, but the two on this thread are
processing the signal beyond simply executing a CRT display circuit well.
On mine at least, the raw bandwidth is totally there and then some. That
set will knock out almost 800 lines. I've sent tests into it. What I
need to do next is use a couple other machines. I have an 800XL that
generates a "real" color signal. Very curious to see what it does with
that one. The GS color signal is a mixed bag. Need to look at it again.
In monochrome, the GS is pixel perfect on that PVM, BTW. Goregous!
I think if I had, what I think is a digital processor, out of the look
in my PVM, the result would basically be the Apple Color Monitor. The
photos here compare almost spot on.
What actually happens as one increases the "sharpness" of an NTSC
composite display, is the color bleed goes away completely. One sees the
vertical bands, and a single pixel ends up the same size it does on a
monochrome screen, and it takes on the color it does due to it's time
position relative to the colorburst reference cycle. On very sharp sets,
and this can be seen on my PVM, artifacts aside, is a red pixel will
have a literally different position on the screen than say a green, or blue one does.
The Apple Color Monitor is opened wider than most consumer grade sets
are, though later ones, say 90's era and above can perform at near that
level with their sharpness cranked up.
If anyone is looking, the SONY WEGA CRT's can totally do that,
especially the ones offering component input. The basic, well executed
CRT circuit and sharpness are there.
I'll try and get a photo.
The other thing I'm going to try is the s-video input. Can split the
Apple output and run one branch through a resister and connect both to
the chroma. Sometimes this plays out very differently on some sets.
Sure has been interesting so far. I didn't expect any of these outcomes
in the PVM scene. :D
Then the only other thing I can think of that may be helpful is that if
you're planning to test other systems to verify whether the unnatural
harshness of the Apple's signal is at fault: the NES also uses a square
wave for its colour subcarrier. It's not artefact colour, it's more like
the Atari in that colour is generated via amplitude + phase, but the
colour part is square rather than a sine.
The Apple video also generates color by creating a squarewave whose phase
determines the color. The amplitude is always 100%.
Post by Thomas Harte
So contrasting those two, if you have access to a Nintendo, might help
with with further diagnosis. Also I'm not really a Nintendo person so I
assume they come with direct composite outputs but I don't actually know.
Obviously if they're RF-only then maybe that's not so helpful.
--
-michael - NadaNet 3.1 and AppleCrate II: http://michaeljmahon.com
Michael J. Mahon
2020-10-04 22:01:23 UTC
Permalink
Post by Doug Dingus
Post by Thomas Harte
Maybe the issue was the sharpness then? PVMs by their very nature are
made to be sharp as they were supposed to be used in professional video
settings which is why all the CRT-ophiles want them (I own a few
myself). However since the Apple II relied on composite color bleed and
artificating to for its graphics, super sharp would mess that up. Maybe
it wasn't so much that Apple monitors were 'tuned' for the Apple II,
but that their consumer quality worked well with the Apple II?
For whatever it's worth, this would definitely be my guess. If you look
at the coefficients of a digital high-pass filter they usually look a
bit like a Gaussian in the middle but then dip below zero at the sides;
that creates a risk of artefacts that aren't physically joined to their
source since the amplitude of any given sample's input goes up, then
back to zero, then back up again as they traverse from the edge to the
middle of the window.
I am not so sure it would.
In my past, I've tuned some TV's up for computer use. Frankly, I've
gotten results on par with both my current PVM and Tempests Apple Color
Monitor. Doing that can take modifying the TV some, opening up bandwidth
in a place or two, and obviously it takes a good set with the bones to
perform, but given those things are well executed, sharp doesn't do what
is being seen on either PVM set.
These PVM's, maybe not all of them, but the two on this thread are
processing the signal beyond simply executing a CRT display circuit well.
On mine at least, the raw bandwidth is totally there and then some. That
set will knock out almost 800 lines. I've sent tests into it. What I
need to do next is use a couple other machines. I have an 800XL that
generates a "real" color signal. Very curious to see what it does with
that one. The GS color signal is a mixed bag. Need to look at it again.
In monochrome, the GS is pixel perfect on that PVM, BTW. Goregous!
I think if I had, what I think is a digital processor, out of the look in
my PVM, the result would basically be the Apple Color Monitor. The
photos here compare almost spot on.
What actually happens as one increases the "sharpness" of an NTSC
composite display, is the color bleed goes away completely. One sees the
vertical bands, and a single pixel ends up the same size it does on a
monochrome screen, and it takes on the color it does due to it's time
position relative to the colorburst reference cycle. On very sharp sets,
and this can be seen on my PVM, artifacts aside, is a red pixel will have
a literally different position on the screen than say a green, or blue one does.
This is always the case for an analog-derived display. Only coarse
phosphor dot grids and low chroma bandwidth obscure the color pixel
position.
Post by Doug Dingus
The Apple Color Monitor is opened wider than most consumer grade sets
are, though later ones, say 90's era and above can perform at near that
level with their sharpness cranked up.
Luminance bandwidth of consumer TVs was limited to about 3MHz by the chroma
subcarrier filter. This was actually seen as desirable, since it made solid
color areas look uniform rather than “dotted”.

The Apple Color Composite monitor bypassed the chroma filter when in
monochrome mode, so that 80-column text would be readable.
Post by Doug Dingus
If anyone is looking, the SONY WEGA CRT's can totally do that, especially
the ones offering component input. The basic, well executed CRT circuit
and sharpness are there.
I'll try and get a photo.
The other thing I'm going to try is the s-video input. Can split the
Apple output and run one branch through a resister and connect both to
the chroma. Sometimes this plays out very differently on some sets.
Sure has been interesting so far. I didn't expect any of these outcomes
in the PVM scene. :D
--
-michael - NadaNet 3.1 and AppleCrate II: http://michaeljmahon.com
Doug Dingus
2020-10-07 05:40:36 UTC
Permalink
Old display chatter mode = 1
Post by Michael J. Mahon
Luminance bandwidth of consumer TVs was limited to about 3MHz by the chroma
subcarrier filter. This was actually seen as desirable, since it made solid
color areas look uniform rather than “dotted”.
Indeed!

Later models, beginning in the later 80's began to open wide up. Broadcasts pushed it all too. I remember seeing TV graphics jump a resolution, from something roughly 320 to roughly 640, with the TV's able to employ better filters, "comb filter" comes to mind, to produce sharper images. An old indie station here stayed with simpler graphics well into the 90's and it was notable. I would look for the sets that showed the dot pattern, being just well executed, sharp circuits and CRT's, sans the filtering and such, and use them for games / computer use. Some people don't mind and seem to process the dots away.

Late 80's and most 90's era TV circuits ended up better than the CRT they were driving too. During that time period, one could get a smaller pitch masked CRT and upgrade a TV and notice! My personal set was a Zenith, hybrid. My fave TV ever! Had a vacuum tube H / V drive, transistor and IC circuits elsewhere. It was a 27" model, and I completely redid it with a 17" CRT that had basically a 2X finer pitch. When I got done, it was a full frame display, PVM "underscan" style. What I liked most about it was a pretty low noise color circuit, and it would widen right up. Performance was close to many of the better pictures in this thread, and rock solid stable.

The smaller CRT, and setup full frame made converging it near spot on. Took a long time, but it was worth it at the time. Being able to read 80 columns and get a great color display was expensive for this rural kid. Doing CRT setups and alignment was fun at the time. Like a puzzle. The more you put in, the better it looks. Then it all fades after a couple years, and do it again... I don't miss the do it again part.

The other thing I noticed was increasingly better art direction as we moved to DVD media and really began to get signals into the home that were better than the TV's and or many broadcasts. Luma changes at higher resolution, color changes at somewhat lower pushed NTSC right to the edge. Anyone connecting a DVD to S-video or component / RGB got the big jump, and watching movies on a CRT was kind of golden for a time. That "dotted" look got filtered and processed away some. Would be super interesting to compare some sets and broadcasts from 70's, 80's and 90's. The improvements were pretty dramatic! That's the "retro" look many are seeking today. It's distinctive.
Post by Michael J. Mahon
The Apple Color Composite monitor bypassed the chroma filter when in
monochrome mode, so that 80-column text would be readable.
Cool. On my "custom" Zenith, I set the color killer trigger right to the edge, and it basically did the same thing. Doing that meant marginal broadcasts ended up black and white. At the time, I didn't care much. The computer was where it's at! Many TV's included that particular adjustment, and it was set to permit color more than not to avoid "but it's not in color" calls on marginal signals. One difference is I basically widened the set up. That "dot" or "line" look actually appeals and in my view, is a big part of what makes great Apple graphics art great! The photo Tempest put here of the PoP title screen is an exemplary example.

On that note, older B&W broadcasts were sharper than we often see them today. That 3.58Mhz knocked what was roughly 500'ish lines down to roughly 300'ish lines. Old programs broadcast in mono looked crazy sharp on a set designed for them. I got to experience this a few times as a younger kid and it always stuck with me. Grandparents had an absolutely huge monochrome TV. It was fun to watch. Honestly, that thing would have done the 80 column text no problem. Hard to find those now. If I did find one, I would totally convert it for gaming! Great phosphors, will leave trails.

Old Display Chatter Mode = 0
Michael J. Mahon
2020-10-07 21:01:51 UTC
Permalink
Post by Doug Dingus
Old display chatter mode = 1
Post by Michael J. Mahon
Luminance bandwidth of consumer TVs was limited to about 3MHz by the chroma
subcarrier filter. This was actually seen as desirable, since it made solid
color areas look uniform rather than “dotted”.
Indeed!
Later models, beginning in the later 80's began to open wide up.
Broadcasts pushed it all too. I remember seeing TV graphics jump a
resolution, from something roughly 320 to roughly 640, with the TV's able
to employ better filters, "comb filter" comes to mind, to produce sharper
images. An old indie station here stayed with simpler graphics well into
the 90's and it was notable. I would look for the sets that showed the
dot pattern, being just well executed, sharp circuits and CRT's, sans the
filtering and such, and use them for games / computer use. Some people
don't mind and seem to process the dots away.
Late 80's and most 90's era TV circuits ended up better than the CRT they
were driving too. During that time period, one could get a smaller pitch
masked CRT and upgrade a TV and notice! My personal set was a Zenith,
hybrid. My fave TV ever! Had a vacuum tube H / V drive, transistor and
IC circuits elsewhere. It was a 27" model, and I completely redid it
with a 17" CRT that had basically a 2X finer pitch. When I got done, it
was a full frame display, PVM "underscan" style. What I liked most about
it was a pretty low noise color circuit, and it would widen right up.
Performance was close to many of the better pictures in this thread, and
rock solid stable.
The smaller CRT, and setup full frame made converging it near spot on.
Took a long time, but it was worth it at the time. Being able to read 80
columns and get a great color display was expensive for this rural kid.
Doing CRT setups and alignment was fun at the time. Like a puzzle. The
more you put in, the better it looks. Then it all fades after a couple
years, and do it again... I don't miss the do it again part.
The other thing I noticed was increasingly better art direction as we
moved to DVD media and really began to get signals into the home that
were better than the TV's and or many broadcasts. Luma changes at higher
resolution, color changes at somewhat lower pushed NTSC right to the
edge. Anyone connecting a DVD to S-video or component / RGB got the big
jump, and watching movies on a CRT was kind of golden for a time. That
"dotted" look got filtered and processed away some. Would be super
interesting to compare some sets and broadcasts from 70's, 80's and 90's.
The improvements were pretty dramatic! That's the "retro" look many are
seeking today. It's distinctive.
Post by Michael J. Mahon
The Apple Color Composite monitor bypassed the chroma filter when in
monochrome mode, so that 80-column text would be readable.
Cool. On my "custom" Zenith, I set the color killer trigger right to the
edge, and it basically did the same thing. Doing that meant marginal
broadcasts ended up black and white. At the time, I didn't care much.
The computer was where it's at! Many TV's included that particular
adjustment, and it was set to permit color more than not to avoid "but
it's not in color" calls on marginal signals. One difference is I
basically widened the set up. That "dot" or "line" look actually appeals
and in my view, is a big part of what makes great Apple graphics art
great! The photo Tempest put here of the PoP title screen is an exemplary example.
On that note, older B&W broadcasts were sharper than we often see them
today. That 3.58Mhz knocked what was roughly 500'ish lines down to
roughly 300'ish lines. Old programs broadcast in mono looked crazy sharp
on a set designed for them. I got to experience this a few times as a
younger kid and it always stuck with me. Grandparents had an absolutely
huge monochrome TV. It was fun to watch. Honestly, that thing would
have done the 80 column text no problem. Hard to find those now. If I
did find one, I would totally convert it for gaming! Great phosphors, will leave trails.
Old Display Chatter Mode = 0
I agree, some older monochrome sets had video bandwidth that was limited
only by the 4.5MHz “intercarrier” FM sound filter—you certainly wouldn’t
enjoy seeing the sound in the picture! This provided a nice luminance
resolution boost over color monitors without a more expensive comb filter.

The best sets were the very early models and the expensive late models. The
cheap sets of the 1960s were only worth what you paid for them.
I was fortunate to acquire an RCA TS630 chassis (the “30” in its number
referred to the number of vacuum tubes—being the first NTSC set, it had a
“lab-derived” circuit with no shortcuts. It even had separate video and
audio IF strips!)

Affordable color sets finally became pretty to view in the 1970s with
Sony’s introduction of the color-stripe Trinitron CRT, and we’ve never
looked back.

Most Apple II graphics, however, were designed on and for relatively
inexpensive color TVs, since they were the most prevalent Apple II
monitors, typically connected using a UHF modulator. As a result, the most
“authentic” and usually the most aesthetic display of Apple graphics is
seen on a mid-range period monitor (or its emulation).
--
-michael - NadaNet 3.1 and AppleCrate II: http://michaeljmahon.com
Doug Dingus
2020-10-07 23:52:00 UTC
Permalink
On Wednesday, October 7, 2020 at 2:01:59 PM UTC-7, Michael J. Mahon wrote:
[snip]
Post by Michael J. Mahon
Most Apple II graphics, however, were designed on and for relatively
inexpensive color TVs, since they were the most prevalent Apple II
monitors, typically connected using a UHF modulator. As a result, the most
“authentic” and usually the most aesthetic display of Apple graphics is
seen on a mid-range period monitor (or its emulation).
I've never subscribed to that view. To me, it was always about what could be done back in the day. And back in the day, I had a lot of fun getting the most out of various gear, radio, TV, computer. Today, I like to see it near peak. That's what I was looking at more often than not back then. Sometimes, that meant optimizing the display to the intended task. Might not want to watch regular TV on it after that, but who cares?
Post by Michael J. Mahon
RCA TS630 chassis (the “30” in its number
referred to the number of vacuum tubes—being the first NTSC set, it had a
“lab-derived” circuit with no shortcuts. It even had separate video and
audio IF strips!)
Nice :D

Scott Alfter
2020-09-21 17:21:01 UTC
Permalink
Post by Doug Dingus
For the most part, I didn't pay a whole lot of attention to these, but
on my //e I see what looks like extra pixels, or echos of pixels.
https://imgur.com/vCkjxl8
That looks like poor convergence adjustment...the green electron gun is
aimed below and to the left of where red and blue are aimed. Convergence
controls are often brought out for larger displays (such as projectors), but
are probably buried inside smaller displays (or, at best, will only show up
in a service menu).

_/_
/ v \ Scott Alfter (remove the obvious to send mail)
(IIGS( https://alfter.us/ Top-posting!
\_^_/ >What's the most annoying thing on Usenet?
Doug Dingus
2020-09-21 18:29:30 UTC
Permalink
Post by Scott Alfter
Post by Doug Dingus
For the most part, I didn't pay a whole lot of attention to these, but
on my //e I see what looks like extra pixels, or echos of pixels.
https://imgur.com/vCkjxl8
That looks like poor convergence adjustment...the green electron gun is
aimed below and to the left of where red and blue are aimed. Convergence
controls are often brought out for larger displays (such as projectors), but
are probably buried inside smaller displays (or, at best, will only show up
in a service menu).
_/_
/ v \ Scott Alfter (remove the obvious to send mail)
(IIGS( https://alfter.us/ Top-posting!
\_^_/ >What's the most annoying thing on Usenet?
Here is the same image with the chroma turned off, and the artifacts are still there:

https://imgur.com/bWtIs0C
Doug Dingus
2020-09-22 05:08:03 UTC
Permalink
GS composite shows the same artifact and is both at a significantly different color reference compared to the //e and a bit worse overall when doing DHGR and lower graphics displays.

GS composite displaying GS level graphics in color also has similar artifacts, and monochrome does not and is basically pixel perfect match with my old amber display.

Side note: I have always wanted to have some fun on a GS. This is a great little computer! Needs RAM tho.

@tempest: I am curious about your display should you get an opportunity to run it. From what I can tell, DHGR and below should tell a lot. I suspect this one over processes non interlaced signals somewhat.

I need to connect another classic system, maybe my Atari to see if it is signal type, or timing or maybe both.
Tempest
2020-09-22 15:55:34 UTC
Permalink
I hope to get to that this weekend. Work has been busy and I need to clear room off the desk to hook up the second monitor.
Doug Dingus
2020-09-22 16:03:13 UTC
Permalink
I hope to get to that this weekend. Work has been busy and I need to clear room off the desk to hook up the second monitor.
Thanks for lugging it out. I got lucky. Parts for my gs came in and I had an early end of day leaving me several hours to play with my gs. No rush though. Retro time is precious. I blew a nice chunk of mine on some games.
Tempest
2020-09-22 16:34:14 UTC
Permalink
Ok I took some really bad pictures, but I can describe what's not in them as well. Sorry they're crooked and at weird angles, I did this quickly between meetings.

As you can see, the PVM has darker richer deeper colors and the screen is 'solid', there's no lines like there is with the Apple Composite monitor. So you'd think that I'd switch to the PVM right? Well there are also some issues that the pictures don't show. For one, the text is blurrier. It's really obvious when you see it and for some reason the Apple Composite doesn't have this issue (my camera couldn't capture it, but I think it has something to do with the lack of those lines). Secondly, when the screen switches resolutions or modes (text to graphics) there's a second or two where the monitor has to resync so you either lose picture or it goes B&W for second.

Apple Composite
------------------------------
Loading Image...
Loading Image...
Loading Image...

PVM Monitor
----------------------
Loading Image...
Loading Image...
Loading Image...
Doug Dingus
2020-09-23 01:11:25 UTC
Permalink
Ok I took some really bad pictures, but I can describe what's not in them as well. Sorry they're crooked and at weird angles, I did this quickly between meetings.
As you can see, the PVM has darker richer deeper colors and the screen is 'solid', there's no lines like there is with the Apple Composite monitor. So you'd think that I'd switch to the PVM right? Well there are also some issues that the pictures don't show. For one, the text is blurrier. It's really obvious when you see it and for some reason the Apple Composite doesn't have this issue (my camera couldn't capture it, but I think it has something to do with the lack of those lines). Secondly, when the screen switches resolutions or modes (text to graphics) there's a second or two where the monitor has to resync so you either lose picture or it goes B&W for second.
Apple Composite
------------------------------
http://atariprotos.com/temp/apple1.jpg
http://atariprotos.com/temp/apple2.jpg
http://atariprotos.com/temp/apple3.jpg
PVM Monitor
----------------------
http://atariprotos.com/temp/pvm1.jpg
http://atariprotos.com/temp/pvm2.jpg
http://atariprotos.com/temp/pvm3.jpg
Doug Dingus
2020-09-23 01:15:05 UTC
Permalink
Ok I took some really bad pictures, but I can describe what's not in them as well. Sorry they're crooked and at weird angles, I did this quickly between meetings.
As you can see, the PVM has darker richer deeper colors and the screen is 'solid', there's no lines like there is with the Apple Composite monitor. So you'd think that I'd switch to the PVM right? Well there are also some issues that the pictures don't show. For one, the text is blurrier. It's really obvious when you see it and for some reason the Apple Composite doesn't have this issue (my camera couldn't capture it, but I think it has something to do with the lack of those lines). Secondly, when the screen switches resolutions or modes (text to graphics) there's a second or two where the monitor has to resync so you either lose picture or it goes B&W for second.
Oh yeah. Your PVM does stuff mine does not. These things are definitely processing the signal. And mine, extras aside, looks more like your Apple color. My gut says well tuned TV's can deliver something close to the Apple color and cheaper.

More later. Thanks for those pics!

The Apples are too old school cool for PVM's!
Doug Dingus
2020-09-23 22:25:19 UTC
Permalink
Tempest,

Yeah, I got a closer look at those today. Did those come from a GS, or Apple 8 bit machine? Looks like both sets came from a GS, but the first set is hard to tell from what my //e does.

With artifact graphics, the //e seems better to my eye and on my displays. The GS has many subtle differences, and did require considerable adjustment from the //e to get roughly the same colors. Once I did that, it seemed to drive yellows and whites a little harder than the //e does with it's essentially monochrome signal. Your pictures aligned with my recent experiences pretty well.

Also, some of that is art choices. The textured nature of the artifact graphics plays a part in things. The GS signal offers some real contrast and had the art choices been made differently, we might feel differently about it today. And the beauty is we can run this stuff on what we want, so no worries.

I will say though, using monochrome mode is excellent! Pixel perfect. Love it.

I saw the same "time to sync and deliver color" when doing that on my GS too. The signal timing is a bit different.

Basically, my PVM adds the extra artifacts, but otherwise looks pretty much like your Apple color monitor does. I'm happy with that overall. The basic feel of the art comes through, and for gaming it's fine. I have a monochrome screen to verify whether something really exists. lol

Your Apple monitor nails the PoP title screen. Wow. :D
Tempest
2020-09-24 00:07:35 UTC
Permalink
Both are from a IIe.
Doug Dingus
2020-09-24 01:48:12 UTC
Permalink
Post by Tempest
Both are from a IIe.
Makes sense. The better PoP picture is very good. Your PVM processes differently than mine, enough to mask that being a //e!

I also think the Apple color monitor does no processing and is just a very well executed CRT display. Like a PVM without the "Pro" part.
Michael J. Mahon
2020-09-24 06:20:11 UTC
Permalink
Post by Doug Dingus
Post by Tempest
Both are from a IIe.
Makes sense. The better PoP picture is very good. Your PVM processes
differently than mine, enough to mask that being a //e!
I also think the Apple color monitor does no processing and is just a
very well executed CRT display. Like a PVM without the "Pro" part.
That is correct.
--
-michael - NadaNet 3.1 and AppleCrate II: http://michaeljmahon.com
Tempest
2020-09-24 13:26:53 UTC
Permalink
Yeah it's similar to the IIgs composite output, which is to say that it has issues with certain things. The deal breaker for me was that the text seemed blurrier. I think it has to do with the normal purple fringing that 40 column Apple II text has being 'smudged' somehow.
Loading...