Panasonic DMR-ES10 review, with proc amps
 

Hi,
Been a while since I've posted here, hope you're all well.

Currently my setup includes the DMR-ES10 in passthrough for its monster TBC. The most rock-solid line correction and frame sync I've had the pleasure of using. I've tested the model thoroughly and there is just one downside, in that it does not capture the full luminance range. Basically, it 'cooks' the video.

Scope Comparison
Here are 3 waveforms for comparing the levels each device has captured. No adjustments have been made whatsoever.
PAL video from JVC HR-S9600EU.

JVC DR-M10
Attachment 5766
Accurate. Captures full range which can be pulled down to restore detail.

ADVC-110
Attachment 5767
Accurate. Near identical to DR-M10.

DMR-ES10 > ADVC-110
Attachment 5768
Obviously clipped. Because the signal is already digitized within the ES10, it is impossible for the ADVC (or any device) to restore since that information is already gone.

Solution
To pull levels into a safe range before they reach the ES10. Before I continue, let me say that I have never owned or used a dedicated proc amp. The closest thing I own is an AVT-8710, which does include a proc amp, but as a frame sync it renders the ES10 useless because the signal is already locked and video errors permanently embedded.

So keeping in mind my limited knowledge on them, what dedicated proc amps can pass a signal that the ES10 would still be able to stabilize? I don't know if there was ever a wide selection for PAL, but certainly not these days. I recall reading that the Elite BVP-4 is PAL compatible (correct?), but would prefer a simpler unit like the SignVideo PA-100, something with luminance metering. Perhaps some of the old Vidicraft models?

If anyone's willing to do me a favour and place their proc amp before their ES10, and post the results, I'd greatly appreciate it. Really hoping I get some advice on how to resolve this issue.

PS:
Please no opinions on the DV format if it caught your eye. I think we can all agree that unless video is correctly timed, stabilized and levels adjusted (among others), capture format is the least of the problems.

First test glitch was placing the AVT before the ES10 in circuit instead of after it. Second factor is using an analog to DV method (not recommended anyway) that offers no control over input levels, which even on the JVC are too high. Everyone has known for a long time that the AVT raises gamma and brights and requires adjustment before capture. The test illustrates why the hookup used is seldom recommended for analog capture.

I don't always have to use a proc amp, but when I did use my PA-100 it was placed after the ES10 and metered a second time with VirtualDub's capture histogram. Since most DV cards have no similar feature, and since I gave up on VHS to DV years ago (your posted test showing one of many reasons why), I wouldn't have anything to post that I didn't discard more than 10 years ago. I have a ton of VHS captures using the AVT and various pass-thru units, but they are all lossless and adjusted for correct levels during capture with either a PA-100 or VirtualDub's hook-in to the capture driver's level controls.

In this case a proc amp is your best choice. Or use the proc amp on the AVT which I used to do a long while back, but I don't know a way to monitor the input with a Canopus card. The assumptions being made are that VHS and DV use the same color system, interpolation matrices, and IRE levels, but problems with those assumptions have been documented for a long time. Even analog cards will give different results, and so will different players.

It's a bad idea to place an external frame TBC in line preceding a line-level tbc. A line tbc always comes first.

More tests and details with various pass-thru devices: http://forum.videohelp.com/threads/3...hat-do-you-use

Jarvis
To test but take the ADVC110 the color bar

You've Sony Vegas, which version?

Here is an example image
... TBC1000 ---> Datavideo DAC 7

Tomorrow Sunday I get the ES10 from the camp
and try it once with the TBC-E, see picture

Sanlyn is right ... in front of the ES10, the level should be adjusted okay.

At the end of sanlyn's link, I posted images that show the same issue you noted, albeit with the NTSC DMR-ES15 rather than ES10 PAL.

I agree that the way to correct this would be a Proc Amp without frame sync functionality.

Sanlyn,

As above, please do not derail my thread by focusing on DV as usual. I am not denying the format has some drawbacks but it is largely irrelevant here. Neither is it the only method I use.

DMR-ES10 > AVT-8710 > ADVC-110
Attachment 5771
Not a test glitch. It speaks for itself. The AVT attempts to correct the exposure but cannot restore clipped values.
Note: Same frame as previous waveforms, of an overexposed scene for the purpose of more evident clipping.

Exactly the point I was making...

To provide a clearer explanation -
For all intents and purposes, the DMR-ES10 is the line TBC, the frame sync (a function of external/full TBC) and the capture device all in one. The levels the ADVC receives from it have already been internally clipped by the ES10. At that point, any proc amp you use after it is no better at recovering lost luma detail than editing software. You can't recover what isn't there anymore. The best you can do is just prevent further clipping by whatever you feed the signal to by adjusting levels into a safe range.

For devices that accept the full range of levels however (ie. superwhite/overshoot), such as the DR-M10 or ADVC (direct), the information is able to be recovered through hardware or software means. Like so -

ADVC - Unaltered
Attachment 5772

ADVC - Levels adjusted to safe range, highlights recovered.
Attachment 5773

ES10 > ADVC - Levels adjusted, visible loss in highlights.
Attachment 5774

Note: Not reflective of proper image correction. The levels have simply been pulled into 16-235 and then expanded to 0-255 as a TV would display. Shows the maximum amount of detail recovery possible here.

So now back to the question -
Can anyone please recommend a solution that protects video levels before they reach the ES10, and sends a signal the ES10's TBC is able to access?

Pretty much this.

Sony Vegas Pro 13. Also thank you for the message and advice I've received from you. If you have any more tests to provide I'd be happy to see them.

Apologies for the 3rd consecutive post if it annoys anyone. I'm using this thread for the purpose of documenting also, a bit more readable not merging posts.

Something of interest I'd like to mention -
There are later Panasonic DVD recorders that also include a TBC, albeit a less powerful one than ES10. The main advantage they have is the ability to output via HDMI, as well as preserve the full luminance range. I do also own a Panasonic DMR-EX77 (HDMI) that sadly does not include TBC, however I have captured from it via a cheap HDMI-splitter that defeats the copy protection. This also negates macrovision on analog sources (I have proof), as the recorder will simply say "copy-protected signal was detected" but displays the video cleanly. Good alternative for those needing full-frame TBCs to capture commercial tapes.

I have yet to try one of these TBC-equipped HDMI models, but there are a few including DMR-EH65 and EH595. This method is well documented over on the Gleitz forums (German Doom9) that Goldwingfahrer is a part of. Basically it is the most modern, ideal way of capturing as it digitizes the analog signal only once, stabilizing the video and preserving the levels correctly, and feeding the clean digital output to any modern card with HDMI input. 576i/480i into lossless or whatever format you choose. It is no more difficult than capturing a regular digital source. I don't personally know how effective its TBC is, but its on my list to find out.

You can think of the HDMI method as a modern hobbyist equivalent to the old Pro TBCs with SDI-out. Same idea.

I bold-underlined one part here. Has someone tested these recorders with over-exposed tapes to ensure they don't clip like we have found with other Panasonic models?

For a while in my chats with premiumcapture it seemed as though his DMR-EH59 wasn't causing any clipping with the tapes he tried, but then he posted this.

Skiller (from German Doom9) mentioned here that all of the DMR-EH models use the same ADC, so in that case it would seem they all have this issue if the DMR-EH59 does.

Yes good question. I've read quite a bit on them at Gleitz but a bit difficult sometimes since I rely on Google Translate.

Based on what premiumcapture said though that is worrisome. I was thinking it could be possible, since even my DMR-EX77 is prone to the contrast problem too if I recall.

I also remember reading about some JVC HDMI DVD recorders with TBC, know anything about those?
I own the old JVC DR-M10 and DR-M100, neither have TBC although do soften the image on input. Their chroma NR is only active on recording/encoding I believe. They are stellar at reading levels though, which I imagine would be the same on the later HDMI models.

Anyway, a solution to the problem would be well worth it for the Panasonic models, I agree with premiumcapture they are better than any VCR TBCs I've used and frame sync is as good as my Datavideo (which is excellent). Just irons out analog sources with ease.

But here are already talking about small children German;-)

Young people speak in English, we old farts had at school only the 3 national languages [Switzerland] German, French and Italian.
Of course, even Latin
mea culpa.....

Small test.
Internal color bars from the ADVC 110 via S-Video ----> ES10 Front AV3
ES10 from Scart / Peritel ------> Datavideo DAC-7 ---- SDI-out map at Blackmagic Studio.
Enclosed 3 photos
Bin satisfied .... ES10 truncates anything.
I have Sony Vegas Pro 13 on the 2.PC
In Vegas, I can not adjust the level before capturing.
That's only in the Canopus Edius with NX card.

supplement
Sanlyn wrote above for ADVC110 ... that was a "Canopus card"
The ADVC110 is like the ADVC300 or ADVC1000 just a "box"

Sony Vegas shows it in studio standard at 16-235 [240]

My Monitor EIZO SX but I have adjusted to 0 to 255 [Image A + B]

I do not know any TV can display 0 to 16.

Possibly, a Vivanco bring help before ES10

Feeder------vivanco----ES10

Warning ... only a few have Vivanco S-Video
Most are only in CVBS

Thanks for the effort Goldwingfahrer, but I believe the test we need is a source with 110+ IRE, preferably 120. The test pattern only goes up to 100 IRE.

I wonder if you could artificially generate such a signal by running color bars -> Vivanco -> DMR-ES10 and cranking up the Luminance control on the Vivanco.

@Jarvis I understand that you are documenting a fe methods that not many would recommend here, so it's quite useful to have a demo of why some methods are frowned upon.

I should add that for some reason I don't get the same problems with my AVT or ES10/ES20 setup that you show here. In the past I've also used the ES20 and Toshiba's for pass-thru, but with less powerful tbc results as you mention. Considerable work has also been done with the ES10/ES20 and other machines for pass-thru in another forum, but with different results than yours. Except for the ES10/ES15, my often used PA-100, and occasionally my old AVT when required, I'm not using any of the other hardware or software mentioned in any post in this thread. That might explain why I don't see the problems shown here (?). I didn't learn much from the images posted in #6 except that they all looked a little washed out for some reason, but that might be because they're screen captures (?). Are your waveforms taken from the original YUV or have they been converted to RGB before taking a reading? I ask because I'm puzzled that I don't have the problems you show. I did have years ago a single Macrovision tape that gave the AVT some trouble, but it gave a borrowed TBC-1000 the same trouble as well, so I attribute that one blowout incident to Macrovision effects. I no longer use JVC players or DVDR's or DV capture and I don't record tape to DVD recorders, so I'd have no way of duplicating your test.

But good to know that some setups don't work so well.

hello mohan
as accurately I can set only the external Osci or in Edius.

I have the original
and
a clip with the Vivanco
When Vivanco from the default null -------> maximimum 10

and another 2 photos
with us again slowly wirds night

addendum

Here the "Vivanco Clip" opened in Sony Vegas
and screen filmed with Mirillis Action

I'm sorry but I don't know what to tell you, this is the conclusion myself and others have reached. Using a proc amp after the ES10 will not restore the luma it internally clips off. That's just how the device works, I believe I've adequately explained why and provided proof. You're welcome to show me your own test results.

Nothing to do with screen caps.
#1 is the unaltered/unedited ADVC capture. #2 shows what is possible to recover in the luma range that is preserved in the captured file (superwhites). #3 shows what is possible to recover with ES10 in the chain (much less). Look to the clouds.

Vegas does not work directly with YUV, it converts to RGB values. The Vegas DV codec uses Studio RGB which is 16-235 (correct). I can show subsequent readings in software of your choice but it will show the same, YUV-native or not. Glenn Chan is a good source for understanding how colorspaces are handled in Vegas.

It works perfectly other than the luminance problem, if that can be resolved the ES10 ticks all the boxes. I can send the signal to anything including via HDMI. The ADVC has little bearing here (white peak is just slightly off).

Also anyone who wants to monitor and adjust levels for DV capture, Enosoft DV Processor has a built in proc amp that can do it on the fly.

PS: Goldwingfahrer, thanks for further testing. It's late and I'm tired but I'll look over this again tomorrow.

I don't know what you mean by this. :o

Thanks very much!

Sample #1 = ADVC 110 <YC> DMR-ES10 (Front Input) <SCART-YC> Datavideo DAC-7 <SDI> Blackmagic DeckLink Studio 2

The maximum luma value in any pixel of any frame of this sample is Y=251. The level of the white bar is around 242-244.
Attachment 5786

Sample #2 = ADVC 110 <YC> Vivanco <YC> DMR-ES10 (Front Input) <SCART-YC> Datavideo DAC-7 <SDI> Blackmagic DeckLink Studio 2

The maximum luma value in any pixel of any frame of this sample is Y=253 (a little additional overshoot at the start of each line caused by Vivanco passthrough). The level of the white bar is around 242-243. As you raised the luminance control, the luma gain increased and the other bars raised. Yellow hit the same level as white, but white never raised.
Attachment 5787

I've attached a few more screenshots.

If you're willing to provide another sample, please do ADVC 110 -> Vivanco -> DAC-7 directly (same chain as sample #2 but with ES10 removed). This should confirm whether the Vivanco is able to output higher levels that the ES10 is soft-clipping.

MP4 see with Edius and image from Oszi

yes....tomorrow

histogram
here it looks different

I think your UYVY decoder being used by AVISource is doing an intermediate RGB conversion. Mine outputs expanded RGB, so levels above 235 in the original are gone if I allow it to do this.

The workaround I've been using for this lately is to Fast Recompress in VirtualDub to a lossless format. Then <16 and >235 are intact as seen in the images I posted. VirtualDub uses its own internal 4:2:2 decoders instead of the system decoders.

Original it is Blackmagic UYVY 4: 2: 2
see Screen

cut with "Direct Stream Copy"

MPC-BE also shows "input and output" type YUY2

Original not cut..cirka 106 MB

ZIP on my homepage
www.ww-consulting.ch/DL/vivanco_Original.zip

Script

AVISource("H:\vivanco.avi")#,pixel_type="YUY2")
ConvertToYV12()
histogram(mode="levels")
ColorYuv(analyze=true)