Who uses a DVD recorder as a line TBC, and what do you use?

[danno78]

Not to be Captain Obvious, but it looks cleaner with NR off.

The source is very clean, so the NR make the picture to look a bit worse. What you see is not noise, is picture changed by NR processing. Save a image (or whole clip) from both and compress with the same ratio and will see that the later have smaller size because NR removed some high frequency (noise/detail).
Later I will do with VHS tape.

jagabo the attachments from 53 post don`t work.

Anyone have an ES20 or ES25 to try this with?

Ask gshelley61 and trhouse.

[jagabo]

jagabo the attachments from 53 post don`t work.

For some reason the images don't show up in the post but if I click on the links they open. If they're not working for others I'll fix the problem.

I did some further testing. They work as described in Firefox. They don't work in IE. I'll see what I can do...

I think they're working now. Be sure to view them full size.

https://forum.videohelp.com/threads/319420-Who-uses-a-DVD-recorder-as-a-line-TBC-and-wh...=1#post1979748

[lordsmurf]

I would be careful not to say "always turn NR off". Some crap tapes benefit from the NR, even when it does some damage. But benefits outweigh damage.

I need to open mine up again, doublecheck chipset, and then do some tests. No time right now, though.

[danno78]

ES15 with passthrough performance with S-VHS tape:
I used the following settings:
BEST=Off
Picture control=Edit
Digital 3R=OFF
TBC/DNR=OFF

JVC + T550:
http://www.mediafire.com/?lwmmtuy1muy

JVC+ES15 with DNR=Off+T550:
http://www.mediafire.com/?j21idxwijme

JVC+ES15 with DNR=On+T550:
http://www.mediafire.com/?evnthznzund

At this file TBC/NR from JVC vcr was activated
JVC with TBC/NR=on+ T550:
http://www.mediafire.com/?lnzlzmjwrhm

There is some motion artefacts when NR was activated. This is also true for JVC NR. The issue occurs when the two consecutive frames are completely different by color. Can be see only frame by frame and only the first two/tree new frames are affected. On normal speed I can`t see anything.
So, without NR, ES15 passtrough performance is close to excellent.

[lordsmurf]

Be careful comparing JVC VCR NR to Panasonic DVD recorder NR.
Some people may think this means it's "the same" -- it mostly certainly is not.

JVC does not have obvious motion oddities to it.

Both have temporal artifacts, if you know what you're looking for.

[davideck]

I wonder what the difference between the the units using the Matsushita chip and the LSI chip looks like. If LS is right about the LSI implementation being bad, perhaps it's better to have an ES10 or ES15 than the ES20 or ES25. Anyone have an ES20 or ES25 to try this with?

I had an ES20 for a while, and its TBC performance was no better than my Datavideo TBC-3000. IIRC, the praise for the ES10 dealt with its ability to eliminate vertical jitter, not horizontal jitter.

[2Bdecided]

Both have temporal artifacts, if you know what you're looking for.

I've looked at both sets of captures from danno78 (DV and VHS), and I can't see anything wrong with them. I've compared side-by-side, frame-by-frame.

It's not aggressive processing (there's still noise in there), but it's not doing anything wrong either. It's basically transparent for VHS (except for the improvement!). Can you see any faults on the VHS NR-off clip?


Whereas I can see "faults" with my Panasonic VCR's NR, watching frame-by-frame. It's quite subtle though - it sometimes averages away (or at least reduces) feint moving picture detail.

Cheers,
David.

[BrainStorm69]

I have a Panasonic ES15 in my hands for a few days. I made a test pattern video (MPEG2) with all 256 possible luma shades. I'm playing these on a Philips DVP 5990 and capturing with a Hauppauge PVR-250. I'm bypassing the MPEG2 encoder on the PVR-250 by using a GraphEdit graph and capturing from the Preview pin. I "calibrated"* the PVR-250 so that Y=16 and Y=235 in the source video are captured at Y=~16 and Y=~235 in the caps. The following images are converted from YUY2 to RGB with AviSynth's ConvertToRGB(matrix="PC.601") so that the luma values don't get stretched. I'm using the VideoScope() filter in AviSynth to show the luma levels. I'm using composite cables because I'll be comparing to a cheap VHS deck later.

I'll post more samples, including VHS, as time permits.

* I put calibrated in quotes because I have no way of knowing if the DVP-5990's output is to spec. I'm just setting the PVR-250 up so that the levels in the captured video match the source video.

So let me see if I am getting this. It looks like your DVD player won't output or your capture card won't capture IRE below 0, so that's being clipped in the samples. It looks like your DVD Player or your capture card is slightly crushing whites above 100 IRE, and that the ES15 is crushing whites even a little more. But other than that, ES15 pass-through doesn't appear to be affecting the luma that much.

Also, you are going to do some more testing with VHS.

ES15 with passthrough performance with S-VHS tape:
I used the following settings:
BEST=Off
Picture control=Edit
Digital 3R=OFF
TBC/DNR=OFF

There is some motion artefacts when NR was activated. This is also true for JVC NR. The issue occurs when the two consecutive frames are completely different by color. Can be see only frame by frame and only the first two/tree new frames are affected. On normal speed I can`t see anything.
So, without NR, ES15 passtrough performance is close to excellent.

danno78, your two tests so far have been with very pristine sources (DV and S-VHS), can you do once more with regular VHS?

I had an ES20 for a while, and its TBC performance was no better than my Datavideo TBC-3000. IIRC, the praise for the ES10 dealt with its ability to eliminate vertical jitter, not horizontal jitter.

davideck, that's not exactly what I was asking about (was directed towards the claimed PQ issues with the poor LSI implementation), but perhaps is more important. I don't really care how good the PQ quality is if it doesn't fix the waviness. So it sounds like ES10/15 are the ones.

[lordsmurf]

Imagine watching a scenic image of waves. But the small waves don't move, only the big ones. It looks weird. That's the NR issue I've seen on the ES10.

[jagabo]

So let me see if I am getting this. It looks like your DVD player won't output or your capture card won't capture IRE below 0, so that's being clipped in the samples.

Yes. I have no calibrated test hardware for testing the output of the DVD player, or any way of generating a known-to-spec signal to feed the capture card. So for these test I'm assuming the output of the DVD player is close to spec (at least you can pretty much see every step in the ramps between Y=16 and Y=235) and adjusting the capture card to suit. Given this, what's more important in these tests isn't the absolute value of the levels, but whether they change when the ES15 is put in the loop.

It looks like your DVD Player or your capture card is slightly crushing whites above 100 IRE, and that the ES15 is crushing whites even a little more.

Yes.

But other than that, ES15 pass-through doesn't appear to be affecting the luma that much.

I agree. Nor does there appear to be any significant posterization in the IRE 0-100 range -- at least in these static grayscale tests. Posterization would show up as broad flat (aside from the noise) spots in the ramps in the graph.

Also, you are going to do some more testing with VHS.

Yes, I'm running tests today and will post some samples later.

[BrainStorm69]

Imagine watching a scenic image of waves. But the small waves don't move, only the big ones. It looks weird. That's the NR issue I've seen on the ES10.

That would be weird. This happens when it's used as a pass-through device?

Yes, I'm running tests today and will post some samples later.

Thanks for the response and looking forward to the VHS samples.

[deuce8pro]

Both have temporal artifacts, if you know what you're looking for.

I've looked at both sets of captures from danno78 (DV and VHS), and I can't see anything wrong with them. I've compared side-by-side, frame-by-frame.

It's not aggressive processing (there's still noise in there), but it's not doing anything wrong either. It's basically transparent for VHS (except for the improvement!). Can you see any faults on the VHS NR-off clip?


Whereas I can see "faults" with my Panasonic VCR's NR, watching frame-by-frame. It's quite subtle though - it sometimes averages away (or at least reduces) feint moving picture detail.

Cheers,
David.

Yeah the clip I looked at with the JVC's TBC/DNR turned on looked fine, I would suggest thats the way to go when putting it on dvd.

[deuce8pro]

First set of files from an Ozzy VH1 Behind the Music VHS master with a Panasonic ES10 pass through with Line in NR turned on, and both input and output black level set to darker, more info:

JVC HRS-9800U S-VHS Deck (TBC/DNR Turned ON, Audio set to HIFI) >
AR Pro II Series S-Video Cable > Panasonic ES-10 > AR Pro II Series S-Video Cable > Audio wires from vcr >
Sony TR-740 Digital8 Camera (TBC/DNR Turned ON) > firewire > JVC DRM-100 DVD Recorder >
1 DVDR SP Mode

[deuce8pro]

Same method, different tape. Joan Jett VH1 Rockin the Rockies vhs master

[jagabo]

Once again, I'm using a Philips DVP 5990 DVD player as my source device. The output of that was sent via a composite cable to a Sony SLV 678HF VHS VCR and recorded on an old used TDK EHG T120 tape. Video was capture by a Hauppauge PVR 250 using a GraphEdit graph to capture from the Preview pin of the WMA Streaming Caputure Device. This bypasses the card's MPEG encoding. The PVR 250's proc amp was "calibrated" to the DVP 5990's composite output. The following caps were exported from AviSynth with ConvertToRGB(matrix="PC.601") to prevent the usual contrast expansion.

The video straight from the DVD player to the PVR 250 capture card:


Recorded onto VHS tape, played back, and captured by the PVR 250:


Played from the VHS tape, passed through the Panasonic ES15, captured by the PVR 250:


The VHS deck played the video a little too hot. The ES15 appears to have toned it back down.

Oh, I made sure the ES15 noise reduction was turned off. Caps were all HuffYUV in AVI. Let me know if you want the AVI files.

[victoriabears]

since discovering that a problem vhs I have , and i tried the "conventional" methods, s-vhs panny and jvc vcr's, tbc1000 to the jvc dvd recorder, was fixed by connecting two 680HS Panasonic VCR's in a chain to a jvc m100 dvd recorder, through a TBC1000 I admit, I have concluded that running round my garden naked and placing a large sausage in the connection chain may help with difficult conversions.

[jagabo]

Here's my low resolution luma/chroma test chart. Same details as post #75 in this thread. I'll address the luma levels first:

Y channel, straight from the DVP 5990 to the PVR 250:

For those who don't know, the sinusoidal wave at the bottom is from the grayscale bars in the image. The solid horizontal line is from the green/magenta bars. They are at the same intensity (about Y=48) but the chroma channels are changing sinusoidally. We'll see more about that later.

Y channel, from VHS to the PVR 250:

There's a little sharpening (increased amplitude of the sine wave in the right half) even though the VHS deck's sharpness setting was turned all the way down.

Y channel, from VHS, through the ES15, to the PVR 250:

A small levels drop in levels at the top end, about 6 pixels. Similar to the levels test in my previous post.

Now lets look at the U channel. I'm going to leave out the top part of the images because they don't change. Only the wave form monitor setting is changed in VideoScope().

U channel, straight from the DVP 5990 to the PVR 250:

There is a little rolloff as the frequency increases. Probably from a low pass chroma filter in the DVD player.

U channel, from VHS to the PVR 250:

As expected, the low color resolution of VHS causes a severe rolloff as the frequency increases. Even at the low frequency end there is a little loss of amplitude in the sine wave, but not in the flat bars.

U channel, from VHS, through the ES15 to the PVR 250:

There's a significant drop in amplitude of the U channel, about 16 units. If you look closely at the full Y channel image above you can see that the green and magenta bars are indeed darker and less saturated.

The V channel data looked substantially similar to the corresponding U channel data so I won't bother to post images.

[lordsmurf]

Okay, now translate that into English.

[jagabo]

Okay, now translate that into English.

ES15 passthrough causes slight color changes.

[davideck]

...IIRC, the praise for the ES10 dealt with its ability to eliminate vertical jitter, not horizontal jitter.

...I don't really care how good the PQ quality is if it doesn't fix the waviness. So it sounds like ES10/15 are the ones.

Does the ES10 eliminate horizontal jitter?

[BrainStorm69]

Does the ES10 eliminate horizontal jitter?

I don't know. But if not, I have a "real" TBC (either the IVT-7 or the TBC-3000) to deal with that.


jagabo - I think you nearly blew my mind with the VHS stuff.
At the risk of oversimplification, the results of your VHS testing (with a VHS deck that is a little too hot on the luma and slightly lower in chroma than the reference) is that ES15 pass-through lowers both luma and chroma slightly.

[jagabo]

At the risk of oversimplification, the results of your VHS testing (with a VHS deck that is a little too hot on the luma and slightly lower in chroma than the reference) is that ES15 pass-through lowers both luma and chroma slightly.

It certainly did for these test patterns from this VCR. But I suspect the ES15 has an automatic gain control that is trying to correct the too hot signal it is getting from this VCR. I did run some other tests trying to understand exactly what was going on. From the DVD player I fed the ES15 from a picture that was too bright (~1/3Y=0, no peaks over Y=192), another that was too dark (~1/3 Y=255, nothing below Y=64), and another with too much contrast (~1/4 at Y=0, ~1/4 at Y=255, and the rest in between). In all cases the output was pretty close to the input, except for the clipping at the top and/or bottom. So the AGC doesn't seem to adjust the signal based on picture content. I suspect it's adjusting both the luma and chroma based on the horizontal sync pulse depth. I have no way of controlling that so I can't verify it.

I don't know what's inside the ES15, but this document:

http://www.datasheetcatalog.org/datasheet2/f/0cosr6uue0wl8ypepjp3cliyfucy.pdf

mentions that the chip's AGC can be based on hsync pulse, chroma burst, or picture properties.

[davideck]

Just to be clear about things...I already have 3 JVC VCRs with TBCs (7600, 9500, 9900) and 2 stand alone TBCs (Datavideo TBC-3000 and I.Den IVT-7). However, the standalone TBCs don't do a particularly good job of what the JVC VCRs' TBCs excel at, that is, correcting the timing of the scanlines so there is no waviness in the image. I specifically want to know if there is a device (for instance the ES10) that I can use as a passthrough device with VCRs other than the JVCs to obtain the effect of the JVC TBC (not considering the DNR aspects of the JVCs).

"correcting the timing of the scanlines so there is no waviness in the image" is what I mean by "eliminating horizontal jitter". Does the ES10 do this?

[danno78]

Yes, it does.

[deuce8pro]

What is horizontal jitter anyway? This has confused me for years as the jitter ive seen has usually been vertical.

[BrainStorm69]

"correcting the timing of the scanlines so there is no waviness in the image" is what I mean by "eliminating horizontal jitter". Does the ES10 do this?

I misunderstood you. As danno78 says, yes it does. Being the lay person that I am, I think of jitter as a back and forth movement or vibration. We need a darn glossary around here.

[jagabo]

What is horizontal jitter anyway?

https://forum.videohelp.com/threads/306272-Computer-video-capture-vs-vcr-to-dvd-combo?p...=1#post1882662

[lordsmurf]

"Jitter" is a technical video term, meaning waviness in the image.
Jitter as used to described something vibrating up and down is a laymen's term, and not a technical term. I forget the tech term offhand.

Thanks for the translation jagabo. I thought that's where you were going, but just wanted to be sure. This change to luma and chroma is standard on Panasonic DVD recorders. What I also noticed on the ES10 was the compression of the color palette -- it gets posterized.

[deuce8pro]

What is horizontal jitter anyway?

https://forum.videohelp.com/threads/306272-Computer-video-capture-vs-vcr-to-dvd-combo?p...=1#post1882662

Are there examples of it such as screen shots from a video (like the ones ive posted)?

[jagabo]

https://forum.videohelp.com/threads/306272-Computer-video-capture-vs-vcr-to-dvd-combo?p...=1#post1882662

Are there examples of it such as screen shots from a video (like the ones ive posted)?

That was from a pair of of videos (crops from screenshots from several consecutive frames enlarged 4x). Here's a full frame with mild horizontal jitter (from one of those videos):


It's a little more obvious when the two fields are separated and stacked:


Those lines should be prefectly straight. The amount each scanline shifts left or right varies with every frame.

Here are some more (along with TBC corrected samples):
https://forum.videohelp.com/threads/230650-Confused-Why-a-VCR-with-TBC-if-separate-TBC-...=1#post1346821