Conventional analog capture workflow vs alternative methods - comparison

[FLP437]

Yes the JVC as TBC/NR ON the settings are as on the picture in one of my previous posts.

Yes I am aware that only some hdmi splitters are fit to the task .I hope I have choose the right one.In a few days I will know.However I choosed the one indicate in the German forum that I referred.

It's interesting the LG dvd recorder as no problems at all through hdmi with the startech

[FLP437]

vaporeon800 as in a previous post stated that probably JVC VCR disable luma DNR regardless of the TBC/DNR button being switched on, thus allowing DNR "off" with line TBC when EDIT mode is on.

I don’t know if it can be confirmed with the sample already included in post 28

I am also posting similar capture but with the Digital TBC/NR OFF to eventually ease the identification to this situation

Unfortunately I´m unable for myself to identify if luma is disabled or not and so I can’t decide if I must capture with the TBC/NR on or OFF. I ask your help to identify if the best setting to achieve maximum quality is to have TBC/NR off or on


I was also interested in clarifying a doubt related to color space conversion and losses . I have seen everywhere recommendations to capture in YUY2 I think as is the more related with the nature of analog video .However when we clean or restore in virtualdub almost all the filters work in rgb. So we are converting YUV to RGB and again to YUV when we save the results. Eventually frame serving directly to the encoding software perhaps, I don’t know for sure avoid the last conversion.But on the NLE probably similar thing happens.

Could it be useful to capture in RGB as probably all over the workflow from tape to dvd is perhaps possible to avoid space color conversions and stay in rgb.

The ATI 600USB for instance doesn’t allow to save in RGB but the startech does allow to save in rgb32 but only uncompressed not in Huufyuv .But it´s possible after to open the file in virtualdub and save in Huufyuv , without losses I think and maintaining the colour space (rgb).

Probably the capture to yuv is more efficient due to the nature of analog video and this is enough to guarantee best results even with additional limited color space conversions, but I´m not sure and was interested to know if this is really always the best solution.

However when the source is in RGB ( what is the case sometimes with HDMI ) does the recommendations to capture to Yuv still remain true ? in this situation could it be better to capture directly to RGB and if yes what will be the better solution RGB24 or RGB32.

I have tremendous difficulty to identify the better capture, or the better clip when they are not too different ( I have the same problem with wine ) is there a utility that can quantify variables quality related that help identify between two clips what is the best one. I have found some utilities for instance MSU VQMT 5.0. As anybody tried it , is it off any use for this purpose? Is there any thread with recommendations/advices best practices on how to compare video quality?

[jagabo]

The samples you've been uploading have no straight vertical lines or edges so it's difficult to tell if a line TBC is working. There's not much in the way of colors, shallow gradients, flat shaded areas, known colors, etc. on which to base any opinions. About the only thing I can tell is that something may be clipping the darkest darks.

[FLP437]

what will be the most appropriate samples for evalution , only including straight vertical lines or edges or there are additional requirements to make possibel a good evaluation between both JVC tbc/dn off and on with the pana dmr-eh65 in pass through.

For now i´m capturing only using a svideo capture card (ATI 600usb) as I have a conflict between the pana and the hdmi capture device (startech) seems HDCP related but even with sources that are not hdcp procteded.I hope a hdmi splitter solves the situation

[Skiller]

It's interesting the LG dvd recorder as no problems at all through hdmi with the startech

It's because the Panasonics have HDCP enabled all the time – even if there is no requirement for it (like playing an encrypted DVD). You absolutely need an HDMI splitter that "forgets" to re-insert HDCP to capture off any Panasonic.

Could it be useful to capture in RGB as probably all over the workflow from tape to dvd is perhaps possible to avoid space color conversions and stay in rgb.

No, you cannot capture in RGB. All that would do is convert YUY2 to RGB at an earlier point. Capturing in YUY2 is optimal, it doesn't get any better than that in terms of color space.

[Brad]

I feel like the latest "TBC/NR Off" capture is still noise-reduced compared to the DV capture.

Did you turn off the "AV-in NR" of the DMR-EH65? Pg 37 of the manual. To get the relevant menu, you need to first switch the input to whichever one your VCR is using.

However when the source is in RGB ( what is the case sometimes with HDMI ) does the recommendations to capture to Yuv still remain true ? in this situation could it be better to capture directly to RGB

It doesn't matter, because the USB3HDCAP doesn't actually support real RGB capture. I mentioned in its thread how the RGB capture mode is described as "pseudo RGB".

[FLP437]

Did you turn off the "AV-in NR" of the DMR-EH65? Pg 37 of the manual. To get the relevant menu, you need to first switch the input to whichever one your VCR is using.

My Panasonic didn’t came with a physical manual but I download both the user and the service manual. In my manual that information is on page 43. However after selecting AV2 ( I´m using scart plug in svideo as apparently has a more stable signal ?..then the svideo plug ) in the picture menu I have only a comb filter I suppose is the one referred but with only on/off and I have always set it to off .

Due to the difficulty to correct evaluate the video samples I prepared a VHS and a Video8 test K7´s. I recorded test patterns from a dvd player to the video8 and vhs VCR´s through svideo and I will be using these tapes for the next samples, I think they will provide more information and it will make easier to identify more accurately problems in order to find the best capture (and workflow) with your help.

I have already the Osprey 210e card borrowed from a friend to test .It uses a Conexant PCIe A/V decoder CX23888-32Z. I think is a chip not very recent but in video sometimes this is not a problem .It integrates an antialiasing, noise rejection filters, and automatic gain control (AGC) and also includes an integrated analog IF demodulator. Video capture is accomplished by using 10bit video ADCs, a full 10bit video data path, and an adaptive 3D comb filter. It accepts only composite, svideo and unbalanced audio. The driver seems to have been updated frequently and works with windows 7,8,8.1 and 10. I used Windows*Driver Class 7 (Version 5.0.7.32, 11/2014 .the driver properties are quite detailed and I didn’t found bugs for now. Also the manual documents extensively the card and drivers properties it´s quite good and almost pedagogic. The card works very well with VirtualDub as it as VFW driver.

For now I am posting 4 samples , the workflow is the same as before

VHS->JVC HR-DVS3->svideo-> Panasonic DMR EH-65->svideo -> capture device->VirtualDub-> Huufyuv

The JVC as all NR off including TBC/DNR and the Pana as the comb filter off

As capture devices I will be using

1-Osprey 210e
2-Osprey 210e again but with the with the TBC/DNR on the JVC on to try to find out if it´s better or not to have it on or off
3-ATI HD 600 usb
4-Startech usb3hdcap ( analog capture svideo plug)

I have remarked with all the devices that even when I have a clean virtualdub histogram if I use the avisynth histogram to visualize I will have always something out of the 16-235 boundaries perhaps it is more sensitive .

I will receive tomorrow the hdmi splitter if it´s the right one and it works with the Panasonic I will provide after the HDMI capture to compare with the best of these ones and find the optimal workflow with these type of equipment.

I am still in love with the matrox mxo2 mini or LE but I will only buy it if really I will get best results , I begun to have a more clear idea about pros and cons and I will post my thoughts about this in the next days

[FLP437]

well, I found the av-in NR it was not available from the main menu and it was in automatic, working in picture input from a video tape . I put it off now but the captures have been made with it in automatic

[FLP437]

I have some doubts related to NR settings related to the JVC SVHS deck and Panasonic DVD recorder , can you please clarify.

I think the ideia is to have only line TBC on, if possible in the SVHS deck and disconnect all NR filter in both devices. I assume is better to denoize after in virtualdub or avisynth and avoid the NR filters implemented in the hardware devices included in the workflow. I assume also that even with all NR filters off in both devices line tbc like function and jitter filtering remains active in pass through.

Or are all the NR filters off only to get a raw capture for comparasion purposes and after for efective capture are some NR filters to be actived as they could provide best results then equivalent post processing software denoizing

As the JVC deck doesn't allow TBC on only, when I do it I will get also NR on, in this case what is the balance , does the improvements due to tbc on override the cons to have NR on or as it is impossible to use only tbc is better to disconnect it as the tbc like function on the Panasonic does the same job equally well .

I just received my hdmi splitter and it works fine with the panasonic .I can now record over hdmi from this device.To avoid posting a lot of samples it was interesting if I could have your opinion about the results from the previous three analog capture devices ( osprey 210e, AtI HD 600 usb and startech usb3hdcap-svideo input).

In your opinion between this three devices is there a winner that provide better results . In this case I will compare it with the similar hdmi version to try to get to a final conclusion

[Brad]

Your new test pattern recordings, recorded on the same VCR as you're playing them back on (and around the same timeframe) won't provide a good sample of time base errors.

The general opinion around here seems to have moved toward capturing "raw" with the noise intact, and then filtering in software. Lordsmurf would disagree with this, feeling that you're not putting the high-end features of your high-end VCR to use.

As the JVC deck doesn't allow TBC on only

I don't think this has been confirmed, since your TBC+EDIT sample from before was using the Panasonic noise filter. But it's entirely possible that using the Panasonic TBC could be preferable to the internal one.

[FLP437]

I hope these last samples can provide the necessary evidence as to allow for the identification of the pros and cons of the two main paths in discussion in this thread and based on that which workflow can eventually delivery best results for this type of equipment’s and video sources.

My target if I can get your support and consequently yours opinions and comments is to ultimately get the necessary insight for answering these questions

1 -Can analog captures using a DVD recorder in pass-through (as A/D Converter and tbc like ), to a HDMI capture device deliver good results and compete and eventually surpass the more established external full tbc and final video capture over svideo ?

2 - If the answer is no from the three different approaches using svideo capture , ATi Hd 600 usb, Osprey 210e and startech usb3hdcap ( svideo input ) which is the one with potential to deliver best results based on results presented ? have DVD recorders used in pass-through for it´s TBC like and filtering features potential to deliver better results then normal full frames external tbc´s ?

Additionally but in second level

3 - Does the samples provide evidence confirming the Panasonic delivers better results then the LG?

4 - Does the Deck Sony SLV T 2000 (combo video8-Hi8 /VHS) provides best results for playing video8 sources then the digital8 Sony DCR-TRV238E camcorder ( the camcorder does gives a sort of color bleeding with strong colors namely red on the right side of some pictures only, the SLV T2000 doesn´t , however picture detail from the D8 seems better then the one from the deck)

The clips included s are video8 samples most from real footage and one or two with test patterns . But even the video8 samples with test pattern have been recorded and played in different devices .The recorder device for the test patterns has been in this case the combo video8/hi8 – VHS SONY SLV T 2000 and the player the digital8 camcorder Sony DCR-TRV238E.

The base video8 player as been the Sony DCR-TRV238E a digital8 camcorder reproducing video8 with internal tbc on and dnr off . The Sony video8 deck doesn´t have tbc or dnr. With a full tbc (avt 8710) and svideo capture cards the results from the deck were not good in my opinion, but with this “hdmi variant” workflow with the dvd recorder as tbc the results seems better but probably worse then the D8 camcorder , But however it doesn´t have the “sort of color bleeding” problem.

In all these captures The Pana DVD recorder as been used with comb filter –off and AV-IN NR – off , the LG doesn’t apparently have the possibility to put NR to on or off.

Workflow variants

D8 Cam (video8)->svideo->DVD Recorder->svideo->capture card->VirtualDub->huufyuv

1 - ATI HD 600 usb
2 - Osprey 210e
3 - Startech usb3hdcap ( svideo input )

D8 Cam (video8)->svideo->DVD Recorder->hdmi->Usb3HDcap->VirtualDub->huufyuv

4 -Panasonic DMR EH-65 + hdmi splitter ( real footage)
4a- Panasonic DMR EH-65 + hdmi splitter ( test patterns)
5- LG HDD DVD Recorder RH199HS

SLV-T2000 deck (video8)->svideo->DVD Recorder->hdmi->Usb3HDcap->VirtualDub->huufyuv

6- Panasonic DMR EH-65 + hdmi splitter


I look forward to have your comments about this subject to go on with my capture project with the confidence that I have choose the right solution / workflow

ps- related to the JVC for vhs capture I will put after a sample with the pana av-in filter off as past captures have been with this filter on

[Brad]

What's the purpose of comparing three samples from the S-Video output of the DMR-EH65 vs. its HDMI output? The results match the theory: an analog capture is the digital capture + added ADC quirks of the particular capture card used. The ATI and Osprey captures are sharpened. The Osprey masks off the top and bottom of the image. No thanks. There's no advantage to the added D-A-D step.

Having said that, your HDMI captures look impure. Did you alter the Proc Amp settings on the StarTech for them? I would expect the black borders to be closer to black, and the LG capture is hard-clipped at Y=226. The clipping alone makes the DMR-EH65 capture preferable to the LG.

I believe we already went over the relative merits of the Digital8 vs SLV-T2000 in the StarTech thread. The SLV-T2000 has actual smear, which sucks, but the added apparent sharpness of the Digital8 comes along with grainy edges, which are an ugly distraction. Overall I would choose to watch the Digital8. I believe by "color bleed" you mean the "edge color issue", which is always present in the Digital8 analog playback sample but most noticeable on prominent colors.

The Video8 test patterns show AGC inconsistencies. Don't know which component is causing this. Possibly the recording machine.

[FLP437]

If I understand the capture process over hdmi using the dvd recorder it does transfers the A/D conversion from the final analog capture card A/D converter to the dvd recorder A/D converter. Also it eliminates one A-D-A cycle as the intermediate TBC also disappears and both TBC like and A/D functions are made only by the dvd recorder without the need to go to analog again and after to A/D .

However even if we spare one A-D-A cycle conversion, the A/D converter from the dvd recorder and the one from the alternative final analog capture card could be quite different in what quality is concerned , as they could have different bits resolutions , comb filters, antialiasing , etc and the final balance could depends on gains and losses related to both situations .

It´s why I have compared captures over svideo and captures over hdmi. to try to get evidence the final results over hdmi were consistent better.

In my case also the avt-8710 as given some poor results and the startech also seems to work better over hdmi then from the analog inputs which all together adds additional interest in the dvd recorder solution

You are not very clear but I think you agree the captures over hdmi with the workflow above mentioned are better

Related to the hdmi captures I don’t think the LG clipping at 226 to be relevant as I have done small changes to the processor amplifier to accommodate the histogram. It´s not difficult with the LG or the Panasonic to get almost perfect 16-235 captures. But I have problems to get them first time as despite having an apparent perfect histogram with virtualdub sometimes I get clipped values and I have to read after luma values with the avisynth histogram and do some subsequent corrections.

Yes we do have spoken about the SLV-T2000 in the usb3hdcap thread I think, but in a different workflow context with the avt-8710 as an external tbc and the startech capturing through the analog input . I think the results improved a lot with the Panasonic as a tbc and using also the startech to do the capture over hdmi and the perceived quality for me approached the one from the D8 camcorder .Also contributes the fact that it does not have the “edge color issue” presented by the D8 camcorder.

“The Video8 test patterns show AGC inconsistencies. Don't know which component is causing this. Possibly the recording machine”
the recording machine as been the SLV-T2000 perhaps is it the culprit

So for now the best workflow with this equipment will be I think you will agree~

JVC SVHS or Sony D8->svideo->DMR-EH65->hdmi->Startech->VirtualDub-Huufyuv

Only remains the detail if it will be better or not on the SVHS JVC to have the TBC/NR on or off

The only improvement I can see is to get a device with a better A/D converter with TBC included to replace the dvd recorder ,there are some brighteye from ensemble designs , analog to SDI converters that I think could fit in this type of device, also grass Valley has a device of this type, but I would have to get an sdi capture card or an sdi to hdmi converter

Everything together it will be a lot of money even if I can get some used equipment. Also I would probably be getting at the end a 10 bits uncompressed file and I will be obliged to use an NLE or other software that supports 10bits as VirtualDub only supports 8 bits, and this all together is a great paradigm change and with a heavy learning curve , however sometimes I like challenges.

[Brad]

the A/D converter from the dvd recorder and the one from the alternative final analog capture card could be quite different in what quality is concerned

Right, but these latest files don't make that comparison. In each case, the DVD recorder's ADC is being used.

Related to the hdmi captures I don’t think the LG clipping at 226 to be relevant as I have done small changes to the processor amplifier to accommodate the histogram.

You should never adjust the Proc Amp for HDMI captures. There is no advantage. The capture device's Proc Amp is only adjusting in 8-bit, and unlike with analog input there is no possibility of grabbing some values that would be clipped above 255 or below 0. You can do the same thing (and better) in Avisynth after capture.

So for now the best workflow with this equipment will be I think you will agree~

JVC SVHS or Sony D8->svideo->DMR-EH65->hdmi->Startech->VirtualDub-Huufyuv

I believe so. We haven't seen direct comparisons of this workflow compared to, for example, D8 direct to the ATI 600, but I believe you've indicated that frame inserts eliminate that as an option.

[FLP437]

Right, but these latest files don't make that comparison. In each case, the DVD recorder's ADC is being used.

I think it´s not the case with the three first captures of the last bunch. They have been made using the DVD recorder only as TBC in pass through obviously increasing one A-D-A stage due to the tbc function , but the capture has effectively been made through the three final card own ´s ADC`s , ATI, Osprey and Startech (svideo-in) , so as to be compared with the other ones made through the DVD recorder ADC.

Related to the JVC TBC/NR On or OFF I think a post from Skiller in another thread is probably correct. He do think that two line TBC in serial could eventually do more harm than good and so it will be better to put this setting to off. Also there is no evidence that the JVC line TBC could be best then the one from Panasonic.

In what Panasonic DVD recorder comb filter is concerned I think that for svideo it probably has no special use. However I don´t know for sure, could it be of any use or it will be better to stay disconnected.

There is another point where I was interested to get yours advice . Could it be interesting to capture 10 bits hdmi vs 8 bits ,using for instance a Black magic intensity shuttle or a decklink mini recorder.

[FLP437]

We haven't seen direct comparisons of this workflow compared to, for example, D8 direct to the ATI 600, but I believe you've indicated that frame inserts eliminate that as an option.

Yes the ATi and also the startech do include a lot of inserted frames ( more the ATI then the startech ): however I will try the osprey it seems more permissive

[Skiller]

They have been made using the DVD recorder only as TBC in pass through obviously increasing one A-D-A stage due to the tbc function , but the capture has effectively been made through the three final card own ´s ADC`s , ATI, Osprey and Startech (svideo-in) , so as to be compared with the other ones made through the DVD recorder ADC.

That's the same thing. If you run video through the Panasonic it's being digitized and the Panasonics ADC circuit happens to be effective at repairing incorrect video timing (that's the TBC function), and it also buffers some full frames in memory. This always happens, no matter if you use it in passthrough or if you grab the video via HDMI.
Now if you do not grab this already digital data off the unit via HDMI all that will happen is the very same data just gets converted to analog again by the DAC and then you capture it again with the ATI or Osprey. There is absolutely no possibility that this would improve anything at all.

Could it be interesting to capture 10 bits hdmi vs 8 bits ,using for instance a Black magic intensity shuttle or a decklink mini recorder.

That would take the 8 Bits HDMI stream and padd it to 10 Bits during capture, no improvement.

[FLP437]

Here there are the final versions if anyone wants to compare. The workflow we have been discussing and the usual ones .Many thanks to vaporeon 800 and skiller that provided the necessary support and tips in order to fine tune the workflow.

WF1- D8 -> svideo-> panasonic DMR-EH65->Hdmi->startech usb3HDcap -> virtualdub-> Huufyuv

digital8 camcorder (TBC on / DNR off)
Panasonic DvD recorder (AV-IN NR off and Comb filter -off )
Startech hdmi proc. amp - default values

WF2 ->D8 -> Svideo -> osprey ->VirtualDub ->Huufyuv

Direct capture from the digital8 camcorder to the osprey analog capture card, without external TBC or DVD recorder in pass through. Only the osprey card has been used as it has been the only one to support the hiccups from the tape/camcorder without dropping or inserting frames. I only made an half hour record and the osprey as been stable without dropping or inserting frames. The ATI Hd 600 usb for this half an hour record inserted about 12 frames and the startech about 8, typical they insert a frame anytime there is a hiccup due to stop/start during recording.

WF3 ->D8 -> firewire -> PC ->WinDV-> DV avi

I included two extras captures only for comparison purposes

WF4 - identical to WF1 but with digital8 camcorder with TBC OFF and DNR off , to try to sort out if two TBC in serial will do or not harm or benefit

WF5 - identical to WF1 but using the LG dvd recorder instead of the Panasonic dvd recorder

With the Startech Proc. Amp. set to the default values as recommended by vaporeon800, the luma values in the captures using the Panasonic dvd recorder have been always between 1-254. I hope this will be ok. However in the captures with the LG dvd recorder also with default values in the startech device the luma values have been between 16-240.Could it be that one was putting out RGB, and the other YUV

Due to my omnipresent incapacity to identify differences in clips not that much different I tried using the MSU video quality measuring tool and measured noise estimation and blurring. I´m annexing the graphics obtained with the 5 video clips here included

For noise estimation from best to worse

Osprey
DV
panasonic using digital8 with tbc on
panasonic using digital8 with TBC off
LG dvd recorder

For Blurring from best to worse

LG
panasonic with digital8 with tbc off
panasonic with digital8 with tbc on
DV
Osprey

I will not try to interpret these values I let it to the experts. I hope I have not made any error in this calculations and in the classification, if I have my apologies . I intended to estimate some other variables but I have to study a little more as some were not evident for me and others I have not discovered how to interpret the results.

[Skiller]

Comparing WF1 with WF5 it becomes obvious that the Panasonic is superior. Here is why: The LG clips the luma at 240 and there is detail missing because of this. The Panasonic does not seem to be clipping any of the highlights in this capture; the maximum is at about 251, so everything seems covered.

The histograms of both are smooth and without spikes or holes – nice.


Looking at the right black border the Panasonic is removing jitter/wiggle well (no surprise) while the LG does not seem to do much in this regard (if anything).

However it's surprising that there is a difference because afterall the TBC of the D8 was on for both, so the lines should already be straight and "uncorrectable" for anything which follows after. That's weird. Can't explain that. It's as if the D8's TBC wasn't even on.

Looking at WF4 the lines even seem to be slightly more stable than in WF1, so I'd say if the D8 TBC does do something, it's actually making things ever so slightly worse.


Conclusion: The LG is totally outperformed by the Panasonic. D8 TBC should be off.


I noticed the D8 camcorder puts a solid black border on the left when it's outputting analog. The DV cap [WF3] has a smaller border on the left (like it's on the tape) because it does not run through the D8's DAC. Not really relevant but I was wondering what was causing it because the border is too sharp to be on the tape.

[FLP437]

Thank you Skiller for yours very useful comments and for the additional information related with the TBC use in the digital8 camcorder. Now the video8 capture workflow is completely tuned.

In the video8 workflow the only way I think to get eventually some small extra added quality is perhaps to buy a top video8 deck. I am thinking about buying the EV-S880 or probably the EV-S9000. I have not yet decided as I know that the eventually added quality will be probably very small if any, the price is quite high, and I don’t know for sure the exact condition of the devices. However if I find one that seems to me less risky and with a slight more moderate price I eventual will be thinking very serious about buying one.

For the SVHS workflow I still have some doubts if it will be better to put the TBC/DNR on or off as the TBC on the svhs deck seems to be clear much better then the one on the digital8 camcorder, the subject I think as not completely been sorted out.

[Brad]

Looking at the right black border the Panasonic is removing jitter/wiggle well (no surprise) while the LG does not seem to do much in this regard (if anything).

However it's surprising that there is a difference because afterall the TBC of the D8 was on for both, so the lines should already be straight and "uncorrectable" for anything which follows after. That's weird. Can't explain that. It's as if the D8's TBC wasn't even on.

Looking at WF4 the lines even seem to be slightly more stable than in WF1, so I'd say if the D8 TBC does do something, it's actually making things ever so slightly worse.

On WF4, the image is shifted slightly to the right internally in the D8, such that its right-border black masking covers the original analog transition from image to horizontal blanking.

(On WF1 and WF5, you can see that the head switch skew extends into the blanking just slightly, before it's also cut off by the masking. On WF4, it's just cut off hard at the same point as the rest of the image. It doesn't extend into the border area at all. This is most obvious compared to the DV -- as with the left-border black mask that you noted.)

For these screenshots I attempted to compensate for shifts, and also did basic ColorYUV tweaks.







Because of this, the right black border can't be used to compare TBC effectiveness, at least not for samples where the D8 TBC is Off. Consequently, I politely disagree with your conclusion that was formed from this basis:

Conclusion: [...] D8 TBC should be off.

I'm also unsure whether the LG is truly worse than the Panasonic in regard to jitter, with these specific samples.

Each of FLP347's files was made from a different playback of the tape, so it could be that the D8 TBC simply performed worse in the particular playback you're comparing, or noise is confounding the issue.

The main difference I see comparing WF1 (On) to WF4 (Off) is in levels uniformity, or whatever the real term would be. The top of the image seems accidentally brightened to a small degree in WF1. I found the same thing in my NTSC comparison.

To be honest, I prefer the DV here...

Pros:
+ No added D/A/D step = no masking or added noise.
+ The edges appear less ragged and noisy than the S-Video captures -- due to the compression softening things?

Cons:
- The odd fields are vertically displaced downward by one scanline, due to Sony's apparent bug. If captured again, the bug may not appear.
- Mild macroblocking visible when single-stepping through fields at 200%. Not going to pretend I can see this during 50Hz playback.

[FLP437]

As asked by vaporeon800 I´m including samples of direct captures made with the osprey card from VHS sources with TBC/NR on and off on the JVC SVHS player
JVC settings (B.E.S.T. -off,Picture Control- Edit, Digital 3R-off,,VideoStabilizer -off) and Digital TBC/NR off or on depending on the samples
.
I included also 2 similar captures with the pana (changing the osprey with the pana and startech and capturing in hdmi)

A small detail if anyone wants to use the the hdmi workflow through the pana, even with OSD off after 30 min the HDD goes on sleeping ant that shows on screen and captures.Better to wait 30 min and capture after the hdd already as gone to sleep , I didn´t find other way.

[FLP437]

I was interested in moving forward and capture the bulk of my VHS cassettes ( for now I have only captured a limited number) , so I was very interested if anyone as arrived to a final conclusion related to the best settings to use in JVC SVHS deck´s particularly if it will be better to use TBC /DNR on or off for captures with the workflow discussed in this thread, based on the samples posted in the previous post.

JVC SVHS->svideo->Panasonic DVD recorder->hdmi-> Startech usb3hdcap->VirtualDub-Huufyuv.

I am sure vaporeon800 and Skiller that have actively participated in this thread have for sure arrived to a conclusion, and I would be delighted if you can share yours thoughts about this subject.

[Brad]

Sorry. I feel bad because you responded with the samples super fast after I asked, and I've been super slow replying. The last few weeks have been crazy for me.

Maybe I can post in more detail later, but for me the best remains what we suspected: Panasonic as TBC and ADC, with JVC's TBC off. The internal TBC doesn't seem to be as powerful.

[FLP437]

Thank you vaporeon800 for your last post. It was the necessary bit of information that I needed to move on with my VHS captures.

I have been searching how to get some additionally quality if possible from my old video8 tapes and for that I have been considering to get a player that could eventually deliver these extra additional quality as the workflow itself it´s already quite optimized , I think ( only if I buy a brighteye 75 or similar device with a 14 bits adc and Full Tbc included but they are very expensive much more than a new player and I will also need an SDI capture card or an adapter) .After studying what were the best decks/camcorders offered in the past for video8/HI8, and the ones that I could still buy in reasonable condition on Ebay or Amazon for an acceptable price I decided to buy the flagship Sony prosumer deck EV-S9000E (manufactured between 95-99 ) from a short list that included also the Sony EVO-9800P, c2000 and S880 .
It has almost everything that I will need F- Mechanism, edit function, TBC, Picture noise suppressing ( off, standard and maximum-only works if TBC enabled) picture control ( Color, Sharpness Y/C delay ),Digital comb filter , drop-out compensation, voice boost, PCM, Tracking adjust ( slow,-slow and x2 ) and a lot of other features that I will probably not use.

Even If I have some expectations I was prepared to get no better results that the ones I already have as it is a refurbished older unit and the condition even if announced as very good , could not be as announced.
.
Well The deck just arrived : I will put some samples in the coming days , but here are my first thoughts. Related to previous captures with a D8 Camcorder

Pros
. Less noise - It seems to provide captures with less noise ,even with NR disabled. If enabled NR works quite well and it can put the question if it will be better to capture with or without NR.

.Color restitution – It seems the color reproduction is more accurate

Vertical Lines – Quite effective even without tbc , much better than the D8 camcorder

.Color edge issue the D8 have a significant right color edge issue, this as a top color edge issue but less omnipresent then on the D8

.In a tape with quality problems where the D8 in a section almost captured no image and no sound , this one was able to capture the image almost without problems but was also unable to capture sound as the D8.

CONS
.It seems to provide a little loss of picture detail related to the D8 Camcorder even with the picture control sharpness increased. However the image in general seems better perhaps better bright and contrast. I still have to confirm this, if confirmed it´s a very strong negative point at least for me that can jeopardize obtaining better captures

It gets quite hot if working for more then 3/4h continuously even with a normal ambient temperature. It needs to be well ventilated.

[FLP437]

I bought some time ago a Magewell “ Pro Capture HDMI “ capture card .I was at the time searching for a card that could include a *time-base correction (TBC) with frame synchronization with suitable video memory in order to avoid an additional external TBC ( already difficult to find , expensive and usually not providing good results as in my case with an AVT-8710 ) ,having ADC´s with reasonable specs , pcie 3.0 or 2.0 and supporting if possible a broad range of operating system versions. I found this Magewell card that somewhat complied with these requirements, it has been developed around a Analog Devices ADV7842 chipset and a Xilinx Artix-7 FPGA chip including also 256 MB onboard memory. There was 2 lines of products 2k and 4k versions some with limited connections HDMI, SDI , etc others with multiple connections, AIO versions, most with pcie 2.0 and some usb 3.0 versions.

I bought a 2k hdmi pcie 2.0 card version that supports HDMI directly, and component, composite and svideo through a breakout cable. The card has been relatively expensive it costed in May 329€ at Amazon. It came in a cardboard box without any identification other than a small tag with the brand and reference. Inside the box only the card in a electrostatic bag ,a breakout cable and a BNC /RCA adapter and a low profile bracket.

From the first tests it was evident that apparently the TBC with frame synchronization seemed to work well, as it is was difficult to get dropped or inserted frames while capturing directly video8 or VHS tapes through svideo, only when turning on the player I got only sometimes 2-3 inserted frames but not if the player was previous in pause or I began capturing after turning on the player, so far I captured about 12 vhs / video 8 tapes and no inserted or dropped frames . Related to the quality provided it seems similar to the one I was getting from the more complex workflow previously discussed in this thread and that as been so far the best capture workflow I have been able to put in practice after almost an year trying several captures cards, players and workflows.

SVHS JVC deck ( VHS )/ D8 Camcorder ( video8 ) -> svideo-> Panasonic DVD rec -> HDMI-> splitter-> Startech usb3hdcap (usb 3.0 ) -> VirtualDub (HuffYUV).

This card could eventually interest people preferring a simpler workflow ( without an external TBC and or a DVD recorder ) without compromising quality and having the benefit of an item easier to buy and supported by current equipment/ operating systems .Has I have said the results, I think are very similar to the ones presented by the previous workflow.

A equivalent workflow with this card would be

SVHS JVC deck/ D8 Camcorder -> svideo->Magewell-> VirtualDub (HuffYUV)

Tests
Even if the Panasonic DVD recorder does a good job removing jitter I think I have detected a minor difference between the TBC´s included in the video8 player and the one in the SVHS player and have adapted captures in accordance. I kept the TBC on at the SVHS JVC deck and worked with TBC off for the D8 camcorder as it seems these provide the best results ( I used also the MSU video quality measurements tools to help me to decide, I hope the results are reliable) . However the tbc on in the JVC seems to shift the image a little and mask about 9 and 8 pixels on left and right. The drivers used for the magewell and the startech were the latest available respectively 1.2.0.3003, and 1.1.0.165 . Magewell Quantization used in all tests , input – Full, output-Full, saturation-Full

VHS
Test 1
SVHS JVC Deck->svideo-> Magewell-> VirtualDub (HuffYUV)

Test 2
SVHS JVC Deck->svideo->Panasonic DVD REC-> Hdmi-> Magewell-> VirtualDub (HuffYUV)

Test 3
SVHS JVC Deck->svideo->Panasonic DVD REC-> Hdmi-> Startech (usb 3.0)-> VirtualDub (HuffYUV)

Visually I don’t see much off a difference only some small shift and top 3-4 pixels masked by the magewell . Again if the results of the MSU video quality assessment tools are somewhat reliable the best of these tests in terms of noise is the one with the Panasonic and the startech test 3 and the better in terms of blur is the magewell svideo direct capture, as the noise in this case is almost the same ,perhaps the magewell svideo direct capture could be the choice.

Video 8

Test 4
D8 Camcorder->svideo-> Magewell-> VirtualDub (HuffYUV)

Test 5
D8 Camcorder->svideo->Panasonic DVD REC-> Hdmi-> Magewell-> VirtualDub (HuffYUV)

Test 6
D8 Camcorder->svideo->Panasonic DVD REC-> Hdmi-> Startech (usb 3.0)-> VirtualDub (HuffYUV)

Tests 3 and 6 have already been presented in this thread however as some variables have changed card drivers , tbc on/off, cables, etc I decided to repeat them in order to make it easier to compare with each other.

Here again the results are, I think , visually similar, by using the MSU test tool the better seems to be the svideo direct capture by Magewell and second the hdmi captures using the panasonic and the magewell or the startech as the values are very similar as expected but the magewell seems marginally better ( MSU test noise estimation and blur ).
.
All tests seems also to be marginally better then older similar tests previous posted in this thread even the ones obtained with identical workflow ( again if MSU results can be believed ) perhaps the small cumulative effect of better management of tbc on /off, new drivers and better cables.

Even if MSU values are reliable they change quite a bit during footage and per tape, so even if the values seem relatively consistent to me this could correspond to a wrong perception and lead to an incorrect evaluation.

Related to audio capture I can easily control and adjust it with the magewell ( windows mixer levels ) with the Startech it probably auto adjust the audio levels, I have not found a way to adjust the audio level record ( or I have forgot how to do it) , in the properties the audio adjustment doesn´t work and the levels in windows mixer also don’t work. It would be interesting in this case also to fine tune the recorded audio levels has they don’t always seem to get the right level.

[FLP437]

I forgot to include de ADV7842 specs and identification of on board card memory.
Also I have been asked to include VHS captures with the Magewell card with SVHS Deck TBC and DNR off ,so I´m including them in order to better evaluate the performance of the internal card tbc.

VHS
Test 7
SVHS JVC Deck->svideo-> Magewell-> VirtualDub (HuffYUV)

Test 8
SVHS JVC Deck->svideo->Panasonic DVD REC-> Hdmi-> Magewell-> VirtualDub (HuffYUV)

[FLP437]

I´m adding two more VHS capture samples but these ones were recorded with a different VCR, this time I used a SVHS JVC HR-S9500EH and the other samples have been obtained from a SVHS JVC HR-DVS3 ( combo miniDV, SVHS ).All the filter options on the VCR remain the same as on the previous captures.

These samples could eventually ease the magewell card and internal TBC evaluation.

I think it´s not difficult to see that there was a reasonable improvement. I was a little surprised because I thought that the other VCR would be better or at least they would be similar.

The tape used in all these tests is a banal VHS tape not a high grade version. The content itself is a copy made 15 or more years ago from a Video8 tape and it does have had some use overtime.

That said I was expecting some contradictory or simple opinions on the magewell card itself, merits and demerits, and on the workflows used with the Panasonics and with this card and advantages/disadvantages of using the TBC OFF / ON in the SVHS JVC recorder to capture VHS with the Magewell an even with the panasonic dvd recorder , as it doesn´t seem to provide a clear advantage.

It would be useful for me and to whom read these forums some debate on these particular issues. I hope to have attracted your attention and interest and get some comments.

SVHS JVC HR-S9500EH Deck->svideo-> Magewell-> VirtualDub (HuffYUV)

TEST9_HR-S9500_tbc_on

TEST10_HR-S9500_tbc_off

[Brad]

Thanks!!! Huh. So it really is 256 MiB RAM. That makes more sense once I thought about the fact that it's a 1080p60 HDMI board that offers internal scaling, deinterlacing, etc. via FPGA.

Some quick thoughts on the VHS samples in posts 56-57 (not 58 yet).

The Magewell is performing at the level I would expect from the ADV7842: very good line correction, but not as good as the Panasonic. Probably the best horizontal stabilization available in a capture card being sold today, with the added benefit of frame sync that you mentioned. Looks like a great option for anyone in the market for a 1080p60 4:4:4 lossless capture card who also wants lossless camcorder* VHS capture. (All seven of those people.) The card is ultimately similar to the Intensity Pro 4K's claims, but unlike Blackmagic, Magewell actually came through as far as the analog side is concerned. The price is commensurately 1.58x higher though.

*For multi-gen tapes, the ADV7840 chip in my AVR is notably inferior to my DVD recorders and Digital8 camcorder TBC so I expect the same is true of this card. Possibly first-gen EP tapes too; haven't compared closely at this point. So before anyone goes running out to buy this card instead of a line TBC passthrough, know that this particular tape isn't the most challenging material out there for a TBC comparison.

for me the best remains what we suspected: Panasonic as TBC and ADC, with JVC's TBC off. The internal TBC doesn't seem to be as powerful.

Time to contradict myself somewhat. I would rate the TBC performance in these new samples as JVC > Panasonic > Magewell. The Panasonic does have the advantage of displaying a completely uncropped frame and also corrects below the head-switch point, unlike the other two. In that regard (flagging / skew error correction) the Panasonic TBC is more powerful than the VCR. But I guess it's at a disadvantage having to deal with the signal as a full PAL input rather than having lower-level access to the signal, internal to the tape player.

Which one provides straighter lines really depends on which field you're looking at; that's how close they are with this tape.

The JVC chroma DNR is doing a nice job here, so besides the missing borders it's the clear winner for me. Other scenes will differ.

Test1 has clipped brights. Test7 has the best levels. You only showed one proc amp setting for the Magewell in Settings.zip, so I'm assuming you never changed this for any of the captures and the difference is purely due to +DNR&TBC vs -DNR&TBC. Hopefully the levels can be recovered by lowering the Contrast.

Test8 (Panasonic as ADC) has slight clipping. You would need to add a proc amp to correct this, but it's minor with this content.

[FLP437]

I can confirm that I have used the same proc amp settings for the Magewell for all the captures ( default settings )

I could have used the option limited quantization on the Magewell and if so I will have got levels exactly between 16-235 . However it seemed to me that to accommodate luma between these boundaries it compress a little at the limits.

Using the option of Full Quantization I get levels exactly between 1-254 and the histogram seems to have a more natural aspect. As I can treat the files as PC 601 I believe I will not have clipped brights or blacks and the results seems marginally better.

d8c- limited quantization
d8d- full quantization

Related to vaporeon800 screenshots comparison, it seems the values I got from MSU video quality measurements are completely consistent with their evaluation.

Using this tool to compare the last 2 capture samples from the JVC HR-S9500 with the ones from the HR-DVS3 I got significantly better results for the fist one (higher values are better). The HR-DSV3 could have the drum a little more dirty as it as been used more frequently but I doubt that it´s the cause, the 9500 is probably real better , the combo unit´s have sometimes to compromise between the 2 players and the the final results can reflect this situation.

In the workflow using the Panasonic the hdmi stream seems to be YcbCr 4.4.4 would I have any interest to capture in 4.4.4 instead of 4.2.2, I could do it with the magewell if it would add any value

So my actual conclusions (with such equipment ) are that the best VHS captures are the ones made by the magewell through a direct svideo connection using a SVHS deck with TBC/DNR on ( between the 2 decks the 9500 is better )

For the video8 captures it seems the best is also a svideo direct capture by the Magewell card with tbc off and DNR off on the D8 camera.

In this context I didn’t know that svideo output on a digital 8 camcorder was adding a ADA cycle and I was not getting a direct analog signal from the tape ( even with all the filters off ). Obviously that situation will not add to the final quality and could be better to get a HI8 camcorder instead and not a digital8, as I´m not using the digital DV path. I have only doubts if the newest electronics inside the D8 could make any difference and counterbalance the final result. Also the difference between the 2 svhs decks gave me a signal that I could eventually earn also some extra quality ( I already tried with the deck HI8 but had no luck so far).

If anyone knows of a list of the best HI8 PAL camcorders in these forums please let me know , all the lists I have saw are NSTC models which have different references If I do find a good HI8 camera I will probably wanting to try this additional path to see if I can still squeeze some extra quality .

Vaporeon800 the way you present the data on yours screenshots is useful can you post the avysinth script .thanks