Calibration: initial sensor based adjustments

In my first set of sensor based readings the results were somewhat less than optimal. Since then I’ve tinkered with my video processor and overall video chain, and toyed around with astig and focus (but have yet to actually do any serious tweaking in that regard). Tonight I got a chance to see if I could improve things using the SpyderII sensor and HCFR software.

I set out to do two things:

  • Check and set my G2
  • Do an initial grayscale

Setting G2: There are several ways to go about setting G2. The suggested Ampro way is outlined here. I used this basic method – cut-off all 3 tubes, then set G2 on each such that the raster is barely lit (usually checked by looking into the tubes). An alternate manner is suggested here – I used this to verify my G2 setting.

An interested observation I had when setting G2 was that while my brightness is set to 95% (normal for an Ampro) some of the AVIA test patterns for black level were not resolving properly. The almost black and nearly black bars were basically invisible. Raising the brightness to max didn’t help – I need to investigate the output of the new Sony DVD player or possibly change the Quadscan to compensate for this. The net effect is that my shadow detail isn’t great at the moment.

Initial grayscale: I focused on two primary measurements 30IRE and 70IRE. The controls I used where the green, red and blue sub-brightness (pictured here), and the 3 small RGB trimpots by the BNC connectors on the back of the same card. If we look way back in time to this posting I had issues with these trimpots being turned way down and had moved them all to max.

Using the HCFR software in continuous measures mode I then setup AVIA with the full gray field test patterns. For 30IRE – I used the sub-brightness controls to try to get the RGB levels as close to 100% as I could. The sub-brightness trimpots are very sensitive, an almost inperceptible change affects the image. For the 70IRE pattern I used the 3 small RGB trimpots to reduce the incoming signal to get the RGB levels to close to 100%.

Gray Scale 2Colour Temperature 2

This is much better and the color temperature seems relatively sensible as well. Ignore readings below 20IRE as the SypderII sensor really doesn’t give you reasonable numbers there it seems.
Having finally watched Pirates of the Caribbean: Dead Man’s Chest the other night, I popped it in for a quick review while I sat down to write this. The black detail is improved over what it was, and the colors appear much more realistic than then did in the past. Some very dark scenes feel like I’ve got too much green, but that may be my imagination. Green forests stand out as looking better than they had, and peoples faces seems more naturally coloured – the costumes also seem seem to have more definition and texture. This seems like a big improvement.

Edit: A few notes from the morning after.  It is interesting that while I tuned for an accurate 70IRE, the final grayscale chart doesn’t show the same level of accuracy that I seemed to have with the continuous measurement.  The overall shadow detail is improved, but my near black is overly green (obvious by looking at the tubes) – this is likely because I raised the green vs. lowering the blue and red.  This is something I’ll have to go back and adjust.  Defocusing the blue a bit will probably help with the higher IRE levels for blue.

Focus

Some of you may laugh – but in investigating doing the astig on my Ampro I discovered that I had never understood the best way to optimize the electronic focus control.

Accessing the internally generated patterns:

  • Press 1, [test] and use [step] to select the dot pattern
  • Cut off all but one tube. (repeat for each colour)

Now by looking at the dots – you will see a huge difference when playing with the electronic focus. In the past I had used a crosshatch pattern to see the effect of focus – but this really hides the differences. Maybe I simply had not looked closely enough, but using the dot pattern makes it really obvious. I honestly in the past had thought that the electronic focus control on my Ampro had limited value – I was wrong. This also helped me dial in the mechanical focus a bit tighter too.

Now in terms of astig. I clearly have both beam centering and astig issues – not really bad, but enough that working on it is going to should have a noticible effect. I, like many, run my CRT with a 16:9 screen – this means that I need to vertically squeeze the image. This vertical squeeze has caused the default 4:3 setup of the CRT with round dots, to have squashed oval dots. Additionally it is clear that the beam centering is not perfect. More on this when I get to tweaking it.

In regards to the DVD player switch and moving from component to SVideo on the quadscan – it was a good change. I haven’t tried alternate component cables yet, but having done some searching around on the net it appears that at least some folks in the past had had issues with the quality of the component inputs.

Lines..

Well, for quite some time I’ve been having some noise lines appear on the screen – primarily with red, but it appeared that all 3 tubes were affected. I wondered if this was the projector or the signal. Having recently played around with checking the grayscale – it was pretty obvious my line problem was getting more serious – the lines were pretty obvious during the test patterns. Normal video material was mostly ok, but some movies such as The Incredibles really showed the problem in some scenes.

My HDTV source is a satellite box (starchoice) and it didn’t appear to have the same problem, but with the amount of MPEG2 compression artifacts it is hard to be certain that there isn’t a problem. Of course if it was my source – then it might be my DVD player, or my quadscan, or the cables.. so many items in the signal path that I couldn’t easily isolate.

I had figured that using the transcoder I have plus my ‘extra’ Sony DVD player [DVP-NS50P] running in progressive mode I could bypass the quadscan. This was relatively easy to try out, and it seemed not to have any problems – other than the somewhat questionable transcoding. No lines.

I then tried the Sony DVD player via SVideo to the quadscan, and again – no issue. Hooking up the Toshiba DVD player [SD-2109] via SVideo instead of my usual component connection the lines also seem to go away. This is an interesting A/B comparison since the Toshiba has both outputs active in parallel, so I can switch between the two outputs and verify that the signal is cleaner on one than the other.

The Toshiba seems to have a very subtle noise in the signal (almost like a ground loop hum) that I don’t see with the Sony. While the trusty 2109 has served me very well for many years, maybe its time to retire it. The 2109 was a great (quality) player, but at this point in time it seems like the (cheap) Sony has an edge on image quality.

After digging around to remove the 2109 from the rack (what a mess back there) – and getting the Sony in place – it appears that with component input I still have the same noise problem. The Sony SVideo is still cleaner than the Toshiba – so I’ll stick with it for a while. The component cables may be the problem, but it may also be the quadscan’s ability to process component input.

CRT Forum

I finally decided to sign up to the CRT Forum hosted by Curt Palme it seems to have a pretty active CRT projector section and not all of the discussions end up being mirrored on AVS in some way. I’ll likely just lurk, but I intend to at least visit regularly.

There was a post by Scott (tse) who knows a lot about Ampro systems. In one posting he comments:

The six devices attached to the registration amp heatsink have thermal shutdown circuits built in. They will turn off if they overheat. Once cool they will turn back on. So if the green vertical amp overheats the green will jump to the position it is in when registration is turned off. Same for blue horizontal, red vertical…..

Check that the screws are snug, they can loosen over time.

Now in this context he was talking about an Ampro 4200, but I’ve seen this same thing happen with my 4000 and hadn’t put 2 and 2 together. I’ve seen the convergence suddenly ‘jump’ to have a huge drift. Entering the help/setup menu would show that things were not really that bad – I suspect that the setup menus reduced image levels result in the system cooling off somewhat. In these cases – the room temperature had risen quite high – probably a bit more than 30C up at the ceiling.