Info: Comb Filter Testing — Page 4

Spaced Ranger said:

althor1138 said:

Here is a clip of my Diamond Theater750HD PCI-E card. This is the composite output from my pioneer dv-360 to the Theater750HD.

AntcuFaalb said:

I guess the only thing left is to see if it exhibits the same AGC issues as the USB one. If it doesn't, then this is definitely the best capture card on the market.

althor1138 said:

Here is the "Pluge Pulse" test with a comparison of the digital file on the right.
...
So what was the verdict on the pluge test?  Has anybody checked it out yet?

Any conclusion on this? I so would like to get the best capture card on the market for this kind of work.  ;)

Unfortunately, no. I'm waiting on the results of poita's experimentation to make a final determination.

I started to use another CLD-D925 as main laserdisc player, and I wanted to test its comb filter, so I re-do some old tests that I previously made with another specimen.

My PC has an AverMedia BDA capture card, based on the Philips SAA713x chipset; I've done these tests using the CLD-D925 composite out to three different DVD recorder, used in passthrough mode as comb filters; the Panasonic DMR-ES15, the Pioneer DVR-320 (comb filter set for movement), and the Yukai DVDR-100B (a Mustek clone, for whom it may concern...), then from the DVD recorder to the capture card via S-Video cable; plus, a direct CLD-D925 -> capture card via S-Video; the test video is from the NTSC Video Essential.

I used also the only PAL laserdisc I own with a test card - actually, a CD-Video, and (sadly) it shows... - to test an USB AverMedia EZMaker capture card, just to see how bad it was!

The result capture were made with VirtualDub, at 720x576 for PAL, and 720x480 for NTSC, encoded with Lagarith lossless codec, then I use a simple avisynth script to show them. The NTSC screenshots have the same frame, while I didn't bother to do the same with the PAL, as they are so ugly... "thanks" to the CD-Video highest quality...

Well, I uploaded images as PNG, and they are quite big (few MB each) so everyone who is curious enough, could examine them better; here you are the links:

PAL CDV test

image 1 (test card) | image 2 (resolution)

NTSC VE test

image 1 (color bars) | image 2 (contrast) | image 3 (test card1) | image 4 (test card2)| image 5 (Snell & Wilcox)

Conclusion: none of them has a decent comb filter, so I should buy a better capture card and/or a device with a better comb filter, or... I could try the phase-inverted trick!

...good result, eh? The only fact that this phase-inverted trick is, actually... a trick! (^^,)

I took a frame when the circle was still, I took the following frame, and I overlaid them to obtain the third picture.

This is a demonstration that inverting the phase SHOULD work! But I tried every possible test to achieve it, without success... I switched off the laserdisc player, then on, capture, nothing... off, inverted the plug, on, capture, nothing... the same with the DVD recorder, at the same time or one at a time, plus with the PC too... I tried to invert the plug for everything, but nothing!!! I've done dozen captures, and the phase was always the same... why? If really it should work when the devices are switched off, if should happen before or after, but not never! Did you try to launch a coin for ten times? Did it happen that the same face appeared EVERY time? Yes, it's a possibility, one out of 2^10... actually, it's not impossible, it's one out of 1024 possibilities, but I should be really, REALLY unlucky - and probably I am...

OR... there is another explanation... maybe the Pioneer CLD-D925 finds everytime the "right" phase, as it has digitally *improved* analog video out? If so, I should try to redo the same test with a digital-free player, like the Pioneer LD-V4300D. Or, maybe I have to switch off the home electricity at all? If someone has an explanation, please let me know, I'm going crazy... poita, antcufaalb, if you read this post, and you did succesful a phase-inversion trick, post it here how have you achieved that task! I'd like to do some important laserdisc captures soon, and I'd really like to use this trick to eliminate all chroma noise before further processing...

metaleonid said:

To the previous poster _,,,^..^,,,_. Something is not right with your screenshots. How come that for the last Snell & Wilcox picture all screenshots have rainbows and dot crawls. Are you sure you didn't by mistake capture directly using S-Video from the LD player thinking that you routed via the DVD Recorders?

I'm sure. Maybe it's because the composite out of the CLD-D925 is simply a recombined Y/C signal and so could not be decombed properly? I should make further tests with another laserdisc player...

Take a look at this post, second image: as my ES15 have the same chipset of the ES10, it seems its comb filter has very low quality...

I should find a new capture card! Or, find the way to let the phase-inversion trick to work...

_,,,^..^,,,_ said:

Take a look at this post, second image: as my ES15 have the same chipset of the ES10, it seems its comb filter has very low quality...

I should find a new capture card! Or, find the way to let the phase-inversion trick to work...

Unfortunately, I think I was wrong with the phase-inversion trick. :-(

Also, the ES10/ES15 is the best 3D comb filter I've ever seen for NTSC material. I don't know how well it does on PAL material, however.

Check out the Pioneer DVR-550 for PAL: http://originaltrilogy.com/forum/topic.cfm/Laserdisc-players-screenshot-comparison/post/633778/#TopicPost633778

Well, the fact that I have rainbow in the Snell&Wilcox both wih still and moving pictures, and I use the NTSC Video Essential laserdisc!!!

_,,,^..^,,,_ said:

Well, the fact that I have rainbow in the Snell&Wilcox both wih still and moving pictures, and I use the NTSC Video Essential laserdisc!!!

Are you using the P/PC (America/Canada) model of the ES10/ES15? It's NTSC-only.

The other models have a different video processing chip. You can verify this by looking at the service manuals on elektrotanya.com.

As I bought it here in Italy, I'm quite sure it's NOT the P/PC model...

What about Phase-Inversion trick? poita seems to had good results using it...

The problem should be the ES15, as it *must* correct the rainbow problems with its comb filter; if not, it simply has a low quality comb filter... and I have not the DVE DVD to test...

So the last test I should do is to use a "pure" analog laserdisc player...

_,,,^..^,,,_ said:

So the last test I should do is to use a "pure" analog laserdisc player...

The LD player I used in my test is "pure" analog: Pioneer CLD-1010

Yesterday a friend brought to my home an old Samsung DVD-R131 DVD recorder... "it doesn't read any disc, try if you could fix it; it's yours now, anyway!"

Another DVD recorder, broken too... what I have to do with it? Thinking to throw it away without even try to clean/fix/test it, suddenly the idea bulb lighted up... "what if..."

So, I googled a bit, found it has a Techwell' TW9906 comb filte; from the paper:

The built-in three high quality 10-bit analog-to-digital converters (ADCs) convert inputs into digital signals for processing.

The TW9906 uses proprietary adaptive 4H comb filter for chroma and luma separation to achieve high video quality. The image enhancement includes horizontal and vertical peaking, CTI and BCS control.

The advanced synchronization processing can produce stable pictures for non-standard signal such as those produced by VCR trick mode.

The high quality scaler uses multi-tap poly-phase decimation filter to accurately scale down the image with minimum phase error. It can be programmed to scale-down the output picture to an arbitrary ratio with cropping.

The TW9906 supports flexible pixel interface. It outputs YCbCr (4:2:2) data stream over 10-bit or 20-bit data path. It also supports both free-running clock and line- locked clock output.

So what you guys think about it? Is the "adaptive 4H comb filter" = 4-lines comb filter?

Now I'm too involved in too many projects, but I'll manage to test it, sooner or later - maybe it could solve some laserdisc capture chroma problems!