Which capture of my old VHS looks the best?

[Krazern]

Hi, I have tried to capture an old VHS tape from 1989 and im not entirely sure about the final results. At first I did transfers in places that "specialize" in transfering VHS in Warsaw in Poland but the results were far from being good enough for me. Only two transfers seemed fine but still lacked in some areas for me. I've looked at many guides and threads on digitalfaq and here as well. In the end i decided to try to do it by myself and i got 3 different VCRs (i had a crappy samsung combo from 2005) and tried to make the best capture i can with the gear i got. My capture chain is now VCR (Edit mode)- DMR-ES10 - Hauppauge USB Live 2 - AmarecTV/Virtualdub 1.9.11 - HuffYUV AVI. I'm using my old pc from 2008 with Windows 7/Windows XP. Now i have JVCs (PAL) HR-J668, HR-J658, HR-S7500. I like the final result the most from hr-j668. JVC HR-S7500 isnt tracking this tape as good as hr-j668 and generally the tape doesn't play good on all vcrs now (it's one and only original tape shared with family and watched many times by myself as a kid ). I know the VCR with TBC and 3DNR would give the best results, but being in college, i can't afford better equipment that is pretty rare here. I'd like someone to rate these captures. Are they any good? On which one should i base my further restorations? Can someone suggest what filters should i use in Hybrid Selur or avisynth? I'm attaching three samples below from two transfer services and one from mine hr-j668. The first transfer was captured without compression with Intensity Pro 4K (read it isn't really good for analog sources) and probably with normal VCR. The second one was transferred with S-VHS JVC VCR and Canopus (all details i know), and the third is mine hr-j668 described above. The rest are captures i got from other transfer service places that didn't tell me what hardware they used. Audio for me seems to sound the best from "other2" sample.

[johnmeyer]

Other2 and Other 3 are obviously awful in every way.

The Canopus is head and shoulders better than the others: it is the only one that gets the levels right (you have some seriously bad blown out highlights on the others). The detail is good (look at the man's sweater at the end of the clip).

The Hauppauge is over-sharpened but is otherwise pretty decent.

I did not look very carefully at the actual motion. There might be some interlacing issues. I had the sense that some of these might have already been converted to progressive. I would recommend you not do that during the capture process.

[Krazern]

Other2 and Other 3 are obviously awful in every way.

The Canopus is head and shoulders better than the others: it is the only one that gets the levels right (you have some seriously bad blown out highlights on the others). The detail is good (look at the man's sweater at the end of the clip).

The Hauppauge is over-sharpened but is otherwise pretty decent.

I did not look very carefully at the actual motion. There might be some interlacing issues. I had the sense that some of these might have already been converted to progressive. I would recommend you not do that during the capture process.

I told the place doing the transfer with canopus to adjust the contrast after watching samples they sent me before capturing whole tape.
It seems its a little darker than other transfers i got.

Unfortunately dmr-es10 blows highlights on hr-j668. The HR-S7500 captures with s-video have normal levels with dmr-es10 but tracks the tape worse. When i try to adjust levels on my j668 capture the transfers levels look pretty similar (highlights blown similary) to canopus. I didnt use proc amp with my transfer because of dmr-es10, left it as it is to be adjusted later.

More detail is preserved in hr-j668 hauppauge capture but i cant really get denoising right. There is something in that canopus capture that makes me like it. On the other hand the dmr-es10 is doing a great job with its line TBC on hr-j668 making image really stable. I attach a sample of denoised hr-j668 made with hybrid.


A few transfers i got were progressive. I didn't post all of them. The "other1" was made with some clearclick copy called grabstarAV or something like that if i remember correctly and has horrible bitrate and no 4:3 aspect ratio. It was done by the same place as the Intensity Pro 4K, so it's possible that they used the same vcr for capture. They normally transfer tapes with normal vcrs, better or worse, using that grabstarAV grabber with its own display and with cheap composite cables. The guy there was testing the tape in vcrs, while i was looking, and it looked fine on the last one. I feel they used the same cheap composite cables with intesity capture too. The capturing cost was the same as with canopus one... They didnt really want money from me for Intensity capture, they had already noticed im not an another casual dumb customer who don't care about quality.

I bought this cable for my transfer. It really made a difference.
https://www.amazon.de/dp/B01DW8M3S8/ref=sr_1_1?__mk_pl_PL=%C3%85M%C3%85%C5%BD%C3%95%C3...B&currency=EUR

[Sharc]

As I see it:
Intensity Pro: Levels correct (limited range), colors slightly greenish tint, details ok. Bad flagging (missing line TBC).
Canopus: Dark level a few notches too low. Brights too low and clipped. Color tint greenish (white shirt is greenish), rainbows.
Hauppauge: Dark level ok, brights clipped and blown out, tint and details ok. -> Fix the bright levels (-> try proc amp adjustment, check equipment settings or use an external attenuator depending which device causes the clipping ....).

Other x: poor

It is basically questionable to compare encodes of totally different bitrates / compression ratios.

[Krazern]

As I see it:
Intensity Pro: Levels correct (limited range), colors slightly greenish tint, details ok
Canopus: Dark level a few notches too low. Brights too low and clipped. Color tint greenish (white shirt is greenish), rainbows.
Hauppauge: Dark level ok, brights clipped and blown out, tint and details ok.

Other x: poor

It is basically questionable to compare encodes of totally different bitrates / compression ratios.

Im considering only three main captures for further restoration.

I added "others" to show how tape playes in different vcr/equipment and what really made me try capturing the tape myself. No one really cares about the quality where i live no more... It's kind of sad that a complete noob like me, that didn't even live in VHS era, can get better results with 20$ vcr than 4.8 star place with >2k reviews

The Canopus and Intensity captures were made after "others" when i really tried almost every place i could to do that and before i got my equipment to do my transfer.

[JadHC]

out of curiosity, do you connect canopus to computer via firewire or do you use it as passthrough and connect to PC via hauppauge, please? im intrigued by the results.

[Krazern]

out of curiosity, do you connect canopus to computer via firewire or do you use it as passthrough and connect to PC via hauppauge, please? im intrigued by the results.

The canopus transfer wasn't done by me but by transfer service place. Only the hauppauge one is mine.

[Krazern]

out of curiosity, do you connect canopus to computer via firewire or do you use it as passthrough and connect to PC via hauppauge, please? im intrigued by the results.

I guess it would be dumb to use canopus as a passthrough as it gives DV video output but can't say for sure.

[Sharc]

.... Only the hauppauge one is mine.

It's pretty decent actually. The blown out/clipped brights are not really severe.

Edit:
Your edited clip in post#3 looks good to me. You may want to encode it to AVC (x264) to reduce the file size.

[JadHC]

out of curiosity, do you connect canopus to computer via firewire or do you use it as passthrough and connect to PC via hauppauge, please? im intrigued by the results.

The canopus transfer wasn't done by me but by transfer service place. Only the hauppauge one is mine.

I guess it would be dumb to use canopus as a passthrough as it gives DV video output but can't say for sure.

yes, i was bout to say that if its a professional company, then they most likely use it via firewire port, not as a passthrough.

Looking at your results, i like "HR-J668-DMRES10-HauppaugeLive2" results the best... It has great colors, contract, perfect TBC and no horizontal jitter. Other 1/Intensity has great TBC too but strong horizontal jitter (waving left to right) which worsens the picture. Audio-wise, i think "other 1" is the best as there is least noise in the audio here...

ps [EDIT]: does your DMR-ES10 support SECAM-L? or just PAL/NTSC?

[Krazern]

I'm not sure but i think it does. I think the options of AV2 input were PAL/SECAM or NTSC. I will check it tonight.

[Sharc]

... Other 1/Intensity has great TBC too but strong horizontal jitter (waving left to right) which worsens the picture.

'Great TBC' with strong horizontal jitter/flagging ???

[Krazern]

Hauppauge: Dark level ok, brights clipped and blown out, tint and details ok. -> Fix the bright levels (-> try proc amp adjustment, check equipment settings or use an external attenuator depending which device causes the clipping ....)..

I have a Sony RDR-HXD895 (bought for like 15eur with non working dvd) and it allows to control levels pretty good. Unfortunately it's far from being as good as Panasonic DMR-ES10 with its strong line TBC and (im not sure) little frame TBC. On sony i get more horizontal jitter and frame skips that don't occure on Panasonic.

I can't control levels with DMR-ES10, and don't want to discard it just by blown whites. Overall the transfer benefits from it are much more important for me i think.
I tried every input of ES10 but the effect is the same- blown whites. Would it make a difference if i adjusted them while capturing? The DMR-ES10 already does analog to digital conversion and then converts it again to analog to s-video output.

The HR-S7500 gives prefect levels with s-video input on ES10. That old crappy samsung combo also has right levels with it, even though im using it with scart/composite inputs. I didn't find a way to adjust levels on hr-j668 or on ES10 so basicly im stuck with what i got now.

For now im waiting for vhs-decode to get better and for me to understand it fully. I think it's the only way to get full proper capture in terms of levels and TBC. It's hard to find VCR that would meet all of requiremnts needed for perfect transfer in Poland and it seems just pointelss to do so for just one tape. Other ones play pretty good on HR-S7500. I don't really want to wear the tape even more for now. It already got more scratches on the tape surface because of all that trasnfers made so far.

[Krazern]

... Other 1/Intensity has great TBC too but strong horizontal jitter (waving left to right) which worsens the picture.

'Great TBC' with strong horizontal jitter/flagging ???

I think it has good frame TBC indeed. There are no frame skips or vertical jumping on these transfers. With DMR-ES10 image jumps occasionally but isn't as asnnoying as horizontal jitter for me.

[lollo]

Hauppauge USB-Live 2 picture (fixed) is the best of the 3

versus Intensity frame 50: https://imgsli.com/MTg2OTU2

versus Canopus frame 50: https://imgsli.com/MTg2OTU3

versus Intensity frame 242: https://imgsli.com/MTg2OTU4

versus Canopus frame 242: https://imgsli.com/MTg2OTYw

Concerning video (in motion), same results.

Concerning levels, Hauppauge USB-Live 2 has the best distribution, but suffers of crushed blacks (frame 123). Tune your procamp.









edit side note: BlackMagic card is quite disappointing in this specific experiment. I have seen slightly better captures from it, although it was the not 4K model

[Krazern]

Concerning levels, Hauppauge USB-Live 2 has the best distribution, but suffers of crushed blacks (frame 123). Tune your procamp.

Thanks for the reply!

Does it matter if i adjust levels with hauppauge procamp while DMR-ES10 already did the conversion to digital with wrong levels?. I explained the level issue in 13th reply.

Adjusting levels post capture wouldnt give me the same result as doing it while capturing in this situation?

[lollo]

DVD-R recorders suffer from brightness level issue. Some workaround here: https://gleitz-info.translate.goog/forum/index.php?thread/47744-tipp-digitalisieren-%C...=no#post461112

Hauppauge USB-Live 2 can only capture 16-254 range, so you might need to use its procamp to avoid crushing input Y levels < 16

[Krazern]

DVD-R recorders suffer from brightness level issue. Some workaround here: https://gleitz-info.translate.goog/forum/index.php?thread/47744-tipp-digitalisieren-%C...=no#post461112

Hauppauge USB-Live 2 can only capture 16-254 range, so you might need to use its procamp to avoid crushing input Y levels < 16

I havent come across this tip before, but im not sure how it should be done.
What do i need to get the same results as someone did in the first post in the link?
Im not really a technical person, yet

Now i also see the black crash, havent noticed it before.
It wouldn't be a big reason for me to do retransfer because of that, but the possibility to get the levels right? Definitly.

To correct black crash alone should i just up the brightness with hauppauge procamp and check it with historgram right?

Im really conservative about doing another transfers. As i said before the tape is pretty old, been damaged with horizontal scratch lines and probably more stuff i don't know yet. I really do feel like it gets a little worse with every play. Maybe its just placebo, overall the results looks pretty good. Tape physicly, besides the line scratches in a few places, looks fine.

[lollo]

What do i need to get the same results as someone did in the first post in the link?

Follow the instructions, and have a read here https://www.digitalfaq.com/forum/video-capture/12818-ohms-value-potentiometer.html and on other topics on the subject.

To correct black crash alone should i just up the brightness with hauppauge procamp and check it with historgram right?

Correct. Choose a significat shot to tune the procamp for best level settings. Also search the forum for some good guides from Alwin with VirtualDub or AmarecTV and GraphStudioNext

Maybe its just placebo, overall the results looks pretty good.

They are. Also consider that adding the ES-10 you're introducing a lossy A/D and D/A, and then the Hauppauge capture is marginally penalized compared to others.

It's up to you to decide if is worth to recapture with better input levels (the black levels < 16 may not contain significat details). If the conditions of the tape are critical, you may have only 1 attempt. Experiment with something else and get familiar with the workflow.

[Krazern]

Hauppauge capture is marginally penalized compared to others.

Before i got USB Live 2 i was using pinnacle pctv cards, they dont compare to clarity of hauppauge.
I also got Pixelview PlayTV Utra card with conexant chip that gave decent results (worse than hauppauge) but it wasnt giving a full frame capture.
Got it mainly because it was recommended on vhs-decode github page. Hope i will make use of it one day.

From what i read here and in other places Live 2 isnt the best device, but gives acceptable results.
I searched for other caputre devices like AIW usb or agp cards but they are impossible to get here.
I tried to find IO Data GV-USB2 but its hard to get and like 3x times more expensive than hauppauge so i settled with it.

It's up to you to decide if is worth to recapture with better input levels (the black levels < 16 may not contain significat details). If the conditions of the tape are critical, you may have only 1 attempt. Experiment with something else and get familiar with the workflow.

Probably arent that critical as it might sound, but at this point i feel like every transfer i make is lacking something, making me recapture the tape again and again.

[lollo]

From what i read here and in other places Live 2 isnt the best device, but gives acceptable results.

False, I do not know where your read that. In a proper workflow it's in line (or better) with the best cards such as ATI 600 USB, Pinnacle 710 USB, I/OData GV-USB2. There are several comparison across this and other forums. This also include (excellent) AIW cards, whose results and comparison have been reported by knowledgeable users like Bogilein here and on digitalfaq forums.

Other cards such as Canopus NX or high end cards like BrightEye75 perform well even if working in not ideal conditions (high end S-VHS player with TBC), because they have internal timing base correction, and give excellent results. For the last see captures from knowledgeable user dellsam34, for the first from knowledgeable user Brad.

A competitive approach is to capture the HDMI stream out of a specific DVD Recorder, which plays then the role of the capture device, avoiding A/D/A conversion to perform the TBC corrections. Search the results from user FLP347 (but that's another story).

[Krazern]

False, I do not know where your read that. In a proper workflow it's in line (or better) with the best cards such as ATI 600 USB, Pinnacle 710 USB, I/OData GV-USB2. There are several comparison across this and other forums. This also include (excellent) AIW cards, whose results and comparison have been reported by knowledgeable users like Bogilein here and on digitalfaq forums.

Oh, i didn't expect that. I think i read it somewhere here in videohelp forum. For me it looked really good from the beginning and didnt see how i could get much better results anymore.

A competitive approach is to capture the HDMI stream out of a specific DVD Recorder, which plays then the role of the capture device, avoiding A/D/A conversion to perform the TBC corrections. Search the results from user FLP347 (but that's another story).

From what i read the Panasonic DMR-EH65 would be good for hdmi capture and its TBC is like ES15 one or identical, but still slightly worse than ES10.

[Sharc]

The "black crash" for the Hauppauge is so minor I wouldn't even touch it.
If you decode to RGB there is at the most one component (typically the Blue) which goes to zero. The others are still positive. Crop the black borders and head switching crud off and apply blur(1.5).blur(1.5) to see that even these excursions disappear, so the occasional infra-blacks are just noise as I see it.
(I refer to the file 'HR-J668-DMRES10-HauppaugeLive2.avi' which IMO is just perfect).

Edit:
Or run jagabo's commandline to visualize the waveform:

"ffplay.exe" -hide_banner -loglevel 24 -stats -loop 0 -i "HR-J668-DMRES10-HauppaugeLive2.avi" -an -sn -color_range 2 -vf "split=2[a0][b0];[a0]waveform=filter=lowpass:scale=digital:graticule=green:flags=numbers+dots:components=1:display=stack:envelope=instant[a0o];[b0][a0o]vstack[out]"

[lollo]

If you decode to RGB there is at the most one component (typically the Blue) which goes to zero. The others are still positive. Apply blur(1.5) and see that even these excursions disappear, so the occasional infra-blacks are just noise as I see it.

Excellent analysis Sharc! We (I) often forget the famous cube of legal RGB values inside YUV range.

[Alwyn]

@Krazern, here is my guide for using Graphstudio to control the procamp of your digitiser in real time:

https://drive.google.com/file/d/1DYxxRMdaVWF4nJpN_5_mbBKsnXcCoBlZ/view?usp=drive_link

[Krazern]

@Krazern, here is my guide for using Graphstudio to control the procamp of your digitiser in real time:

https://drive.google.com/file/d/1DYxxRMdaVWF4nJpN_5_mbBKsnXcCoBlZ/view?usp=drive_link

Thanks! Will be useful for other family tapes im collecting now, especially stubborn ones like this one (and the most important one...) that plays the best on j668.

I think i got lucky with that JVC HR-J668. It seems to be in pristine state. The seller told me it was sitting in his wardrobe for like 20 years at home. Looks like it wasnt really used back in time.

[Alwyn]

The seller told me it was sitting in his wardrobe for like 20 years at home. Looks like it wasnt really used back in time.

A dream come true these days.

[Krazern]

ps [EDIT]: does your DMR-ES10 support SECAM-L? or just PAL/NTSC?

Didnt find any option in ES10 reffering SECAM, but its manual mentions it. In German manual u can see it on page 39.
In my Polish manual its mentioned as well in the same way.
https://www.manualslib.de/manual/24783/Panasonic-Dmres10.html?page=39#manual

[JadHC]

ps [EDIT]: does your DMR-ES10 support SECAM-L? or just PAL/NTSC?

Didnt find any option in ES10 reffering SECAM, but its manual mentions it. In German manual u can see it on page 39.
In my Polish manual its mentioned as well in the same way.
https://www.manualslib.de/manual/24783/Panasonic-Dmres10.html?page=39#manual

oh wow! you actually found this. This is excellent! This gives me a comfort in ordering it for my SECAM tapes too (i have similar issue as your other 'non-live-2' recordings - great TBC but horizontal jitter, so will either order your DMR-ES10 or lollo's DVR-530H). Truly thanks for indeed checking this and getting back to me with a reply.

Plus, im happy that others were able to help you with your recordings. I agree with them that Live-2 has the best results. I have also been using hauppauge-live-2 based on advise of other forum members and im getting the best results possible. This forum is the best thing honestly...

[Krazern]

oh wow! you actually found this. This is excellent! This gives me a comfort in ordering it for my SECAM tapes too (i have similar issue as your other 'non-live-2' recordings - great TBC but horizontal jitter, so will either order your DMR-ES10 or lollo's DVR-530H). Truly thanks for indeed checking this and getting back to me with a reply.

I have Sony RDR-HXD895 (made by Pioneer i read) and its line TBC isnt ad good as ES10 one, "frame" TBC is even weaker. I'm not sure how HXD895 compares to popular Sony RDR-HXD870 and to DVR-530H, but they might be simillar.