Canopus ADVC-110/300 vs. ATI AIW/600 USB comparisons?
 

From what I know(I can be wrong) most people from this forum would say that to capturing the signal from the VHS cassette to the digital file the best option to get the best image quality will be: ATI AIW USB and followed by: ATI 600 USB. At the same time, the Canopus(DV) devices from the ADVC series, such as 110 or 300, are considered by at least some people to be worse than those of ATI. Yes ?

I know that there were probably people who compared these devices to each other. Is there any material on the internet showing image screenshots showing these image comparisons of the same picture frames: Canopus ADVC(100, 110, 300) vs ATI AIW USB (and/or ATI 600 USB) ? I would like to see the difference (if any) in the image between these devices. [It is known that using the same any settings, equipment only changing the converter.]

[Does it make a difference for the image quality which type of device(USB, PCI, AGP, PCIe) - for example ATI 600 USB - is used to capture the image(from VHS to digital file) ?
Maybe not a USB converter(and specifically those from ATI) is the best in terms of the best quality of the image being obtained only some other... ?]

EDIT:
I forgot to add that it's about converting the PAL signal and as far as I'm concerned about the comparison, it is just for the PAL signal.

The actual difference is that DV capture devices use a lossy DV codec that is outdated, incompatible with new devices and capture color at half the resolution for NTSC, While a PCI or a USB capture device gives you the option to capture lossless first if you plan on editing or saving as master footage and you can always compress later to a video format of your choice with decent color resolution and decent encoding bitrate.

Thank you for your response.
I already know this theory well(I think), now it would be nice to see a comparison about which I wrote in the first post. :D

Here is a lossless sample:
http://www.digitalfaq.com/forum/atta...t-gatheringavi

From the exact same PAL tape, here is a DV capture:
http://www.digitalfaq.com/forum/atta...athering-dvavi

Though the first sample is from JVC S-VHS using S-Video cable, the second is from a Toshiba VCR using composite cable, Not really a good comparaison.

Yes, so it is not good comparison.
In general, the first movie captured losslessly to AVI I can't be strangely played. Pops: "no encoder - decoder; this element has been encoded in a format that is not supported.; 0xc00d5212 ".

You need MPC-HC or VLC in your computer, Windows media player won't play such files.
The most important thing to take from those two files is that chroma noise is horrible on the DV sample, while the lossless sample has somehow better color uniformity.

definitely a unique request

comparing a lossy format to a lossless format (DV to Uncompressed), by definition the DV will be missing information from the Uncompressed.. so all you could compare by inspection would be the analog filters on the input.. and the DV would always start out at a disadvantage

to my inspection when I had to "see for myself" the DV always appears "fuzzy" or out of focus on all captured video compared to an Uncompressed sample of the same video. I think the technical term is "softer" or lacking detail

where Canopus ADVC got its fame was in stablizing the video and dropping fewer frames and making "smaller" file sizes.. until Huffyuv and Lagarith came along and elimited the smaller file size advantage

None of this invalidates the better "tolerance" that some capture gear had over newer and less expensive gear for bad and unstable analog video signals. ADVC came from an era plagued by marginal analog signals so was and "is" very tolerant of bad analog video signal "acquistion". Its major downfall however is it can't output "Uncompressed" video capture over the Firewire connection.

Here are some test files with PAL test patters.

A few months ago I have create a test file which I have burn onto a dvd. To play the dvd I have connect the DVD Recorder Panasonic DMR-495 with S-Video connectors to the s-video input connectors of the different capture cards.

I have left the capture settings from each card at default settings.

To compare the files you should use a Vectorscope & a waveform monitor.

yes

there's that fuzz again, although its a bit more pronounced than I recall.. i was stubborn at hoping it would be "not that big a difference" but finally had to admit it was

ugh.. these samples just put it all in stark relief all over again.. great captures.. good samples

DV white levels are very bad, Although it's PAL chroma is almost the same as the other samples because it's 4:2:0, You should post a NTSC file for DV it will reveal its weakness.

Thank you very much for this comparison. It is really helpful.

I tried to take the same screenshots of the image(using the video from the movie, which appears from 6 to 9 seconds of the movie) and by watching full movies.
So:
I don't see the differences between ATI AIW USB (ATI Wonder USB) and ATI AIW 9600.

ATI 600 USB seems to have clearer signs than ATI AIW USB (ATI Wonder USB) - for example: letters, numbers - which can be caused by the contrast that creates the background colors with the presented characters, but I'm not completely sure this better characters in ATI 600 USB - maybe its value is not caused by contrast, or not only by it.
ATI AIW USB (ATI Wonder USB) has in my opinion better presentation of colors than ATI 600 USB which is manifested in the fact that as if they are more natural, defined and delimited, where in ATI 600 USB as if at the borders of colors are created blurring, transition from one color in the other.
ATI 600 USB as if better defining different shades of white, where in the same situation ATI AIW USB sees the final of shades as one white color.
And the black color seems to better delimit the ATI AIW USB, where the ATI 600 USB does it worse. Regarding these black and white colors, evidence for this we have in the fields under the numbers: 235, 16, 1 (I also refer to comparisons between other devices) and on a long line showing the transition from white to black (and vice versa) in the middle of the graphics.

ATI AIW USB (ATI Wonder USB) has a better definition, color delimitation than Diamond 500 USB.
Diamond USB 500 as if better defined different shades of white, where in the same situation ATI AIW USB (ATI Wonder USB) sees all of the shades as white. In shades of black Diamond 500 USB it's like a bit (very slightly) worse than ATI AIW USB (ATI Wonder USB).
The Diamond 500 USB seems to have clearer signs than the ATI AIW USB (ATI Wonder USB) - for example: letters, numbers - which can be simply because of the contrast of the background colors with the presented signs, but I'm not completely sure of not being that much more clear in Diamond 500 USB - maybe its value is not caused by contrast, or not just its value - it certainly don't matter as compared to the ATI 600 USB with ATI AIW USB (ATI Wonder USB), where the color difference is greater there. Colors in Diamond 500 USB and ATI AIW USB (ATI Wonder USB) are similar in terms of naturalness.

The Canopus NX looks the same as the ATI AIW USB (ATI Wonder USB) - although it can be very gentle if the Canopus was better, but almost imperceptibly in terms of the quality of the characters presented.

Canopus NX vs Diamond 500 USB - Diamond 500 USB as if it had clearer characters and better shades of white and slightly inferior blacks. Canopus NX - better color separation, where the Diamond 500 USB seems to blur a bit.

I don't know why the video for Canopus ADVC 300 is vibrating. It's really hard to tell if the ADVC 300 is better or the ATI 600 USB ... In some reason, probably one is better, the other worse and vice versa (I see it that way).

I may be wrong, but I would probably put together a list of the best of these devices in terms of the quality of the presented image, where the first four devices are very close to each other:
1. Diamond 500 USB
2. Canopus NX
3. ATI AIW USB (ATI Wonder USB)
4. ATI AIW 9600
5. ATI 600 USB [or vice
6. Canopus ADVC 300 versa]

-- merged --

In my opinion to show in practice differences would be the best option(captured VHS and showing from it the same screenshots or/and the same parts of movies). After all, the best that would be showing image screenshots showing these image comparisons of the same picture frames: Canopus ADVC(100, 110, 300)(DV AVI or DV) vs ATI AIW USB and ATI 600 USB(AVI, lossless) or the same parts of movies - for all - to VHS record converted to digital file - using the same equipment and settings everywhere on these connections and only changing converter(for example Canopus ADVC 300[with turning off all filters], 110 to ATI AIW USB). All in PAL system.

I have upload the samples without any comment and I'm really surprised what you see.

Never trust your eyes/hardware (monitor). I highly recommend to compare the capture files with the original mpg file. Put the files on the timeline of your editing software and use the vectorscope & waveform monitor to compare them.
Here for example pics from the black/white scene with the super black (0-16) & super white (235-255) luminance range.

Attachment 10407

Here the demonstration what you should see on this picture.
Attachment 10408

the diamond 500 clip the super black at 14 & the super white at 245
Attachment 10409

the canopus nx clip the super black at about 2 & super white at about 254
Attachment 10410

the canopus advc 300 clip the super black at about 2 & super white at about 254
Attachment 10411

the ATI AIW 9600 clip the super black at 16 & the super white at about 252
Attachment 10412

the ATI AIW USB clip the super black at 16 &the super white at about 252
Attachment 10413

the ATI 600 USB clip the super black part at 16 & capture the full range to 255
Attachment 10414

this are the capture result with "DEFAULT" capture settings

you can try to make the super black/white part visible (to broadcast safe range 16-235) with the contrast/brightness sliders in virtual dub.

with the other test patters you can compare if the capture card, capture the full resolution (720x576) or you can see how the agc control work or the colours, sharpness, if the capture window is centered (have a look at the AIW 9600 file)....

The ATI 600 USB looks sharper as the other cards because the sharpness is at level 2 at the default settings, this is just digital sharpening which you should better do in the post production.

If you compare the files, remember the Canopus ADVC 300 file (DV) is bottom field first.

I have upload the test captures from a dvd and not from vhs because I can upload the original file too and you can see what the different cards capture.

I apologize for my lack of professionalism in assessing your previous materials. I do not know if I understand you correctly:
"I have upload the test captures from a dvd and not from vhs because I can upload the original file too and you can see what the different cards capture."

If so, how would you be able to extract from the VHS cassette sample fragments of some material and put here as a comparison of the same screenshots and / or the same parts of the film for at least: Canopus ADVC 300 (DV AVI, DV) (all filters disabled eg by Picture Controller 300 and image settings like: color, on the default), ATI AIW USB and ATI 600 USB (AVI, lossless for both) and for the rest if you want and you can do it.

So far, all nonsense. This thread has nothing to do with capturing from analog sources. MPG isn't analog. See, guys, analog video has constantly varying waveforms, changing noise patterns, has no pixels, and video moves around a lot which is what test patterns don't do. Apparently some people here don't know that mpg isn't analog. If you need samples of analog sources, try VHS, SVHS, VHS-C, video8, Betamax, or laserdisc. But MPG to make non-moving still images from pristine digital material and try to conjecture about analog to lossless or analog to lossy?????? No way. MPG doesn't even use the same collorspace as analog. Sorry to be hardnosed, but you're wasting your time.

For the past couple of days browsing this thread, I recalled now and then my adventure of a few years back with Canopus cards and retail and homemade VHS tapes. OMG, what a waste of time capturing that analog stuff to DV. It was so ugly and futile I almost gave up capturing for good until I found this forum's guides and learned to avoid getting suckered by video marketing hype, and commenced doing things the way the pros do. So, this thread was OK for a brief trip thru memory lane, but this repetition of lossless-vs-lossy debate every few months for the past 20 years is really tiresome.

The test files from dvd should show you the limitations of your hardware (capture card & monitor).
If the capture card can't record full range luminance (0-255) or can't record full resolution or just blur the picture or add other errors to the test files from dvd it will o the same with vhs content.

All USB captures devices I have had can only capture the luminnce range from 16-255.
The ATi 600 USB just oversharpen at default.
The ATI AIW 7200,9000,9600 resize the picture if captured with 704x576 to 700x576.

The thread starter ask for samples with PAL content. I have upload captures which were made always with the same hardware. I don't think he can find better samples but maybe you could upload samples, sanlyn:salute:?

If necessary Lord Smurf can delete the nonsens I have post.

No, let's get serious. That has nothing to do with reality

LS, they're living in a different universe out there.
WTF is this guy talking about?

No, I lived thru the VHs->DV b.s./hype/fiction and have no patience or time for repeating the experience. Maybe someone out there has both analog and DV devices and can capture from the same source.

Only captures from 16-255? Really? On planet Earth? :question:
Would you like to see a USB device that captured below-zero blacks and brighter than y=255? There are plenty posted in the forum. It's considered a typical newbie mistake and is usually corrected in Avisynth/YUV before doing anything else with the video.

Oops. Lemme partially correct that. The ATI USB 600 clips at y=16 and won't capture anything below that. Technically we say it "crushes blacks". Otherwise it can capture illegal signal levels just as well as anything else, especially if you don't know what the hell you're doing.

-- merged --

OK, after a couple of cups of Tension Tamer tea, I'll try broadening my outlook on this thread. I get the idea from several people that a workflow from capture to restoration/repair to final encoding hasn't really taken place -- or if it has, what was learned from the process? Can anyone in this thread answer the question: what does restoration and repair consist of? What are the properties of the required media? What software and what filters are used, and why? What analysis tools or other objective/aesthetic analysis or measurement tools are used to determine what gets done and what gets fixed, and what remains as-is? What does "editing' entail, precisely? If the final product looks pretty much like the initial input, what's the point of capturing to a PC in the first place other than chopping things into pieces and putting them together again? What's the purpose of capture devices and computer codecs when DVD recorders and PVR's are available, along with smart rendering NLE's and lots of free disc burners, cheap HDD storage, and video servers? It seems that for many users the workflow consists mostly of memorizing which icons to click and in which sequence to click them. How do you judge your own results? What do you compare them to? You're not a professional, you don't have pro gear, you don't have pro training, but in light of what pro's do and the results they get, how do your methods and results compare? Where do your results fall short, and why? If the capture device is basically chosen for its ability to cover your butt and mask your visual and technical shortcomings, make all the decisions for you, and correct your mistakes and omissions so that no further work or improvement or learning is required, why do you need a/v forums when the product's user manual has all you need to know? Why not take your tapes down to Walmart? What would be different about what they do and what you would do? Other than cut-and-join and posting on YouTube, what comes after capture?

at Sanlyn: Have you download the files and compared them or you just call them bullshit?

Put them on the timeline of your editing software and compare them and then post your opinion or what you see.

For me I prefer a capture card which can capture the luminance range from 0-255 and the full resolution 720x576. What is lost during capture you can't get back in post production.

If you have read my comment I have written this are the "default" settings. You can try to make the range from 0-16 & 235 to 255 visible with the brightness & contrast slider. This is what you do with the histogramm with virtual dub. It looks like you always know what you record, where the blackest or the whitest part is on your tape to make the correction.

I correct the levels to broadcast safe range (16-235) in the post production.

ATI 600 USB: I have captured another sample with sharpness level 0. These looks more like the original.

I don't know why you always talk about Canopus=lossy DV?

I don't recommend to capture with the ADVC300. The ADVC have well known issues(like the audio bug and others....).

The Canopus NX capture in YUV and not DV!

I have made about 20 test captures from different cards (ATI AIW, tv cards with BT878 & Philips 7134, Viewcast Osprey 240e, Pinnacle USB 500, Blackmagic Intesity Shuttle, USB card with EMPIA chipset to name a few...) if I compare them on the timeline of my editing software you'll see that the ATI AIW capture is smaller. I have to stretch the 704x576 capture to 700x576 to get the them result as the original
file. These happens only to all the 3 AIW 7200,9000,9600 cards.

I don't think many people have the possibility to compare their captures with different cards.
The test patters include the "Hello World" part. If you can't read Hello world, your computer monitor or tv can't show the levels from 0-16 & 235-255. This I think is important to know if you edit your videos later.

Another thing I have left off is the audio part which would be important too for the capture card. But if the video part is worse....

at Bogilein: i don't know exactly wich is your capture flow, but remember that in PAL standard ATI AIW have resoultion/aspect ratio issue:

ATI MMC and 704/720 resolution thoughts?

to obtain correct aspect ratio for PAL standard, in VirtualDub is needed setting the capture resolution to 720x576 and than crop 16 pixel (8 for each side) to obtain 704x576 that fit AIW specification, but with correct aspect ratio.

The videos aren't BS. They're just videos. The BS is your doing. The test panels did help me decide that none of the cards you used except one would be suitable for the VHS sources I encounter. As for your past captures you mentioned, the use of cards designed for digital and digital HD sources has been well covered and discounted as less than ideal for analog SD work.

I notice in the videos that there is no motion. I know that sounds obvious, but you don't seem to think it's relevant. Don't you analyze videos that have motion? (I'm sure you do. It's a rhetorical question). I mean, for want of a better term, "real" video, the kind people sit and watch in front of TV or in a movie theater or with family videos and the kids? How do these cards respond with motion and noise that doesn't stay still? How do they react to tape noise? How do you judge skin tones and color balance in your timeline? Do you know how to work with video in anything other than an RGB timeline editor? I assume that your editor isn't the tool you use for repairs. What about excessive combing or tape noise levels or telecine effects? What does the tape noise look like and what type of filter do you think you'll need to calm noise on motion and camera pans? Excessive grain? How about noise in shadow areas? What would your timeline tell you about chroma noise, Chroma shift, bleed, or chroma displacement? Dropout problems? Spots? Horizontal ripples? Frame hop, jitter, luminance or chroma flicker, over- or under-saturated segments? How do your reds look -- are they supposed to be really red or do they tend toward yellow or purple? What can your timeline tell you about consistent white balance, gray balance, and black balance (how determined and corrected?). What about DCT ringing or sharpening halos? Any signs of chroma or luma ghosting at scene changes? Motion smearing or edge ghosts? Signs of previous digital processing in the mastering of the original tape (how can you recognize it)? Any posterizing, pixelation effects, clay-face? Hanover bars? Backlight effects or over contrasty front lighting? Camera exposure problems, CMOS noise, shadow range noise problems? Block noise, banding or hard gradient edges? How does your timeline editor correct for these and what do you do about it? Any border stains, skin tone discoloration/blotching or similar color problems? Edge noise with color anomalies along contrasting edges? Broken lines or edges, line twitter? Shifts in color balance and/or levels from scene to scene? Mosquito noise (common with DV)? How do you detect and fix these problems?

That can be tweaked later but 16-235 is the limitation during capture. You shouldn't try to make that correction in your RGB timeline. It has to be tweaked in the original YUV colorspace. Do you know why?

Canopus encodes everything to lossy DV internally, before capture. It's a DV card, dude. And what did your timeline editor tell you about the codec used for your Canopus captures? Canopus used their proprietary codec with a fourCC code of CDVC. It's otherwise known as "Canopus DV". It is a lossy codec. Not only that, but Canopus always renders botton field first, even if the original PAL VHS is top field first, and this field reversal often has noisy playback phase problems that can't be corrected later.

The text below is a MediaInfo readout of tech properties of your Canopus capture. The relevant data sections are in blue.
The readout for your Canopus NX capture looks incomplete to me. The NX cap seems recompressed with Lagarith. If it hasn't been recompressed, why doesn't it have the same kind or style of capture data as the other captures? Tech data on that NX capture looks rather suspicious to me. What happened to the original capture?

Isn't your MPG lossy YV12 and top field first? How would that relate to YUY2 analog video instead of pristine digital source? Do your capture cards respond in the same way? The MPG uses long-obsolete MPEG-1 L2 audio. What does your timeline tell you about the properties of your MPEG "original": ?
Attached is MediaInfo Reports.zip, a collection of MediaInfo text files for all of the videos you posted.

What are you talking about? DV is YUV. What type of colorspace do you think DV uses?

Correct. That's why people don't capture to lossy codecs for use in restoration projects.

This one made me laugh -- out loud, seriously. My ATI captures are NTSC 720x480. I have over a dozen frame choices available. Have you seen the Virtualdub dialog below?
http://www.digitalfaq.com/forum/atta...1&d=1563648301

If you're setting this up and you still get a different frame size, something is amiss and you should fix it. The ATI 600 worked properly for me and for millions of others in both PAL and NTSC. So do my other ATi cards.

700x576? Is that a typo?

To answer that one again, I did a lot more than that. I looked at them as YUV originals and examined them that way with various analysis tools. The image below is a frame from the ATi 9600 capture. It's not a "screen shot". it's a direct frame copy from the capture you posted. The data readings and the histogram are directly from YUV, not RGB. I want to know what the video is doing in YUV before your timeline editor screws it up with sloppy RGB conversions.


http://www.digitalfaq.com/forum/atta...1&d=1563648445


The test patch with the pluges was more valuable. I have no choice but to agree with you that if a card won't capture many image elements, the card is useless. The colors above look pretty soiled and poorly rendered, but we've known that about the 9600 for a long time. It does indeed capture values below y=16 and above y=235 bit does some clipping at both point (seen the YUv histogram and in absence of detail in the white and black pluges). In fact none of the black pluges in any of other test samples have satisfactory black detail, so I suspect the pluge rather than all the cards.

The ATI 600 USB test panel:
http://www.digitalfaq.com/forum/atta...1&d=1563648562

Notice that the 600 exceeds y=235 but still retains some detail in the white pluge panel. No detail in the blacks, but it's already known that the ATI 600 clips at y=16 before the signal reaches the capture software. If you glance at the matching dframe for the VC500 you'll see that the vC500 maintains bright details about as well as the ATi 600 does but doesn't clip blacks. Colors are more pure with the 600. Another assumption I'm making is that you've often used your editor's pixel reader to analyze actual RGB color values in those chroma patches.

As it is, neither my clients nor I spend much time with test patterns. We're more interested in learning what's wrong and what's right with lossless captures of conventional video and what can be done with them. For several days I've worked on a VHS movie-to-digital project for a local film society. I don't see how your editor or test panels could help with that, other than suggestions some cards to avoid. The French movie with subtitles is on Korean DVD but looks worse than the VHS tape, which itself has lots of problems. The Lagarith YUY2 sample attached has correct capture levels made with an ATI AIW 9600XT. It's typical of many warmish, dimly lighted interiors and night scenes in the film. One challenge is keeping a scene's startup levels when the scene later brightens (as shown in the small image below) when the room's drapes are drawn open and the scene explodes with brights...


http://www.digitalfaq.com/forum/atta...1&d=1563648678

....Also, the VHS is undersaturated in several segments. It will look even brighter and less saturated on TV. It will look even brighter and less saturated on TV. If you "correct" the color balance in the darker frames the colors will look wrong in the bright frames. And VHS has higher noise levels in dark frames. What can your editor and test panels tell you about working through those problems? What card would you suggest for a typical problem VHS like this?

That's only the beginning of the glitches including artifacts from a bad digital mastering job on the movie's journey from film to VHS. The original encoder imbedded jitter into many scenes. If you want more bad video to play with there is 480mb YUY2 at Mediafire in my pro-level account (no popups or ads on downloads) at https://www.mediafire.com/file/rtcyx...mple2.avi/file. One way your test patches helped me and my friends in the film society with similar movies and their noisy family videos is that I wouldn't choose any of the cards you used except the VC500, which I already own. The society projects movies onto a 100-inch screen in a small auditorium -- Would you advise adding DV data loss and more DV compression artifacts to this movie? Or using a capture card with anemic color (ATI 9000)? Or one that clips darks at y=16?

The Canopus NX is a pci or pcie (depending on the version) analog capture card, not a firewire box like the ADVC models.

The statement was that the NX version was "YUV, not DV". The NX version and the DV version are both YUV.

Ah, sorry, I misread. Disregard that.

Anyhow, I don't have an AVDC or a ATI 600, though I made a small comparison how DV compression can affect a PAL image. This is a shot from a lossless capture, followed by an image of a DV-encoded version of the same sample.
(I forgot to adjust levels before making a snapshot so there may be some slight clipping due to RGB conversion.)
The DV encoding was done in software with the Cecocida DV codec, though the DV encoding is simple and a constant bandwidth one so the result should be approximately the same regardless of what did the conversion.

5x Zoomed comparison, lossless on top, DV on the bottom. Notice the blocking and loss of details on the trees:
Attachment 10426Attachment 10427

Original:
Attachment 10422

DV:
Attachment 10423

Thread title renamed, as it's a request, not a comparison thread.

I have all of the cards, plus many more. But my time is finite, not always able to spend hours and hours with test comparisons. I'll consider doing an ADVC-300 vs. ATI AIW (Theatre 200) in the near future, when my time and equipment is freed up some.

But it is a huge difference, the DV loss is noticeable even to non-video folks. Most see it on their own, the rest once pointed out to them. Rarely does a person "not see the difference", and generally it's being facetious and purposeful/willful ignorance ("just video", to wit my response is generally that dog food is "just beef").

Maybe.
The ATI 600 PCI is completely different from the ATI 600 USB.
But the ATI AIW USB uses the same Theatre chipset as (almost all of) the ATI AIW AGP cards.

This is the correct answer. :congrats:

While good intentions, unfortunately a useless test, too many variables.

Not really.

Yes, that's all the test would show.

That's it. 50% chroma loss makes fuzzy color.

This is the same issue that dSLR cameras had for years. Only in more recent times have cameras had no OLPF (optical low pass filter). The OLPF was lossy chroma subsampling that kept images from being tack sharp without Photoshop sharpening.

Pfft. No. ADVC isn't more tolerant, that's just Canopus marketing myth.

Unfortunately, again, good intentions, but useless testing. Too many variables.

It's because you're double dipping. 4:4:4 patterns > 4:2:0 DVD MPEG > either 4:2:0 DV (different from MPEG) or 4:1:1 incurs extra loss.

Yeah, PAL DV isn't terrible, DVD-like acceptable loss, NTSC has the 50%+ color destruction.

I'd say wrong.

NX is (H)DV.

The VC500 has some darkness issues, some observed oddities with AGC, and potentially mid-production changes that explains oddities. sanlyn seemingly has a good card, but I've seen samples from others that have a miserable time with it.

When I finally had time to click into this thread, that was my observation as well. :screwy:

Not really.

I don't really agree with that. Sharpening is a complicated topic. Everything applies some degree of sharpening somewhere, be it in chipset, registry, or software. Sometimes a bad combination of choices oversharpens.

I'l be sure to do both PAL and NTSC when I have time to make samples. :)

Nah, not really nonsense, just potentially flawed testing, it's good to leave those posts there. :salute:

Don't forget, everybody has a different skill level. And with this, PAL vs. NTSC differences to consider. So any one of us may sometimes learn new info, update our knowledge. Overall, I think Bogelein is fine, has good intentions, but maybe some room for improvement in testing. He's not getting angry over this topic, so neither should you be.

Yeah, you were wound a bit tight in that last post. :laugh:

Hardware + software ideal to restore.

Lossless (or uncompressed) ideal for capturing for editing/restoring.

Filters used depend on errors present. But also noting some filters are destructive, so you'll often need compensation filters to undo damage. The end goal matters. Restored video is better, not necessarily perfect. But with the caveat that "not perfect" not be used as an excuse for shoddy work.

Good questions for others.

Facetious questions, amusing, but the answer here is what drove people to this site, unsatisfied with the cheap and shoddy work from the cheap item or service.

Audio is really bad on some cards.

:thumb:

However, that is from 2003 (posted here in 2005), and applies mostly to Theatre Rage/100 chipsets. I believe that was fully fixed in the Theatre 200. Maybe even later drivers that affects Rage/100. I know drivers can affect sizing, and multiple Rage/100 cards had alterations in later MMC versions.

So I don't know that any of this applies unless using the oldest 7000 ATIs with original on-disc drivers.

This is why I've never put much stock into test patterns. Limited usefulness.

Agreed.

Yep. 4:1:1, 4:2:0, and 4:2:2 are still YUV. :congrats:

I generally leave it as native UVYV, don't change AIW to YUY2,

My thought as well. Weird sizing, must be typo.

But still a DV card.

The problem here is that DV converters don't seem to subsample well, and just drop colors. Software subsampling can look better, but depends on the codec. Cecocida is somewhat crappy compared to Matrox.

RE Cedocia:

The Cedocida installation has a configuration menu that should be set up for the source video you're working with. Its defaults aren't ideal for everything. The configuration is accessed via Virtualdiub's "compression..." menu. Click the compression menu item, then click Cedocida, then click the "configuration" tab. Results will differ depending on the setup for specific video source.

In any case the loss and the difference will be evident, even with pristine source. It's inevitable. "Lossy" means "loss', period. You can't have data loss and end up where you started. It's not possible. Don't just count raw data loss. In digital video, data loss involves distortion as well as detail content. VHS color is also distorted -- if you have a sensitive color eye (many people do, many don't), you'll see that "something" in the original colors has changed and "something" seems off-balance. Even original DV source has plastic looking color IMO.

Not to sidetrack the conversation, but
- lossless can have loss (and so can uncompressed)
- lossy can be visually imperceptible to loss, especially higher bitrate lossy like H.264 and ProRes422

Stuff like that really screws with people trying to learn video. Mind blown. :laugh:

So always remember: every rule has exceptions.
But also remember: when you think you're the exception, you're probably not.

Honestly, when somebody starts to get overly dogmatic with video, or tech in general, I find it fun to throw these monkey wrenches in their gears, and watch smoke come from their ears (ERROR! DOES NOT COMPUTE! BZZT!)

Graphs and charts and benchmarks and test patterns are sometimes more for entertainment value than not, and it takes experience to navigate this complex topic. Those often oversimplify, sometimes to the absurd and unrealistic. The data can be skewed, and marketers know this. It's how myths get started, how BS is propagated.

This reminds me of baseball this season...
When a good batter walks up to the plate, good stats (good batting avg, slugging %, on-base), that's really all you need to know. Get ready for a likely hit, maybe HR. But no, some announcers stat you to death, % this and % that, and in the end none of it matters anyway. The guy hits well, probably gets paid a lot for it. It's those guys that get overpaid and can't bat that irk you, non-quality that costs $$$. Even worse are newbies/rookies that strand runners, complete amateur hour. If you follow baseball, you probably see the parallels here.

Yeah there will probably be some difference. I thought the DV spec didn't leave much room for differences in implementation, but looking at it now, I see that that was incorrect. In any case, DV compression will incur a noticeable loss of image quality, and make post-processing more challenging.

Regarding DV versus lossless, you can look at the VHS comparison shots here. To my eyes, the DV images clip some of the reds in the uniforms compared to the lossless image at the top. When color is clipped like that, it's gone. No amount of postprocessing will get it back.

There are DV codecs that would yield better qualty.....DV25 uses the 4:2:0 color space some some quality is loss colorwise, while DV50 maintains the 4:2:2 colorspace in which the color information is retained.

PAL DV(25) = 4:2:0, alternating halving, like DVD-Video, but different axis from DVD-Video MPEG
NTSC DV(25) = 4:1:1, quartering

It's why PAL color loss isn't bad, while NTSC is obvious.

Yeah, that's right on the 4:1:1 for NTSC DV25. I'm getting a little rusty lol. Either way DV50 or DV100 is superior compared to DV25

In my opinion, it would be very good to make and publish the results of such a test, if possible(VHS recordings captured on a digital file - lossless (eg HuffYUV), PAL standard, resolution: 720 x 576, via S-Video and so on vs in the same way in the DV codec (Canopus ADVC-300) (changing only the converter)).
[Screenshots of the same picture frames and / or movie fragments.]

Remember that I checked it myself (I can be wrong): all filters in the Canopus ADVC-300 can be turned off by the program: Picture Controller 300. (When connecting the signal via Composite - Y / C Sepration 2D (W) or 3D (H) unless always one of them will be active (which we choose) - probably. By connecting via S-Video these two options are inactive (although they can be switched in the program) - I think.)

Hello, I have thinked about one week if I should write an answer if nobody needs testsample or won't accept the results.
First let me say this; I'm from a PAL country (the home of the original Oktoberfest) and could only speak about PAL hardware and expirence with PAL content.

Another thing what I will remember. I wrote here not in my own language and if I'm in a hurry I could make some mistakes if I didn't checked back my writings. (@sanlyn: YUV,DV)

The Test patters are test patters and just should show if a capture card could record the full resolution for PAL 720x576 and if the card could capture the full range from 0-255 or if some parts are cropped or be clipped and if the capture window is centered. The luminace range from 0-16 is often called super black and the range from 235-255 is often called super white. The broadcast safe range is 16-235 but on analog sources like vhs the highest white & deepest black point could be different on each tape and could be in the super white/black area. With some capture cards you could make them visible with the brightness/contrast slider but not with every card. Some cards just clip this area and this will be lost during capture and it couldn't make visible in the post production. You should do this during the capture process with virtual dub (but I always want to know how you adjust this for the whole capture? You guys always watch the whole video before the capture just to know where the whitest/blackest part is in the movie?).

I just would like to know what's wrong with the test patters and the result.

The first 2 test patters are a 3 second long black and white change. Where the black are with luminance 16 and the white scene with luminance 235. At the bottom of the test pic I have added the waveform monitor how it looks like under the waveform monitor from the original mpg file.
-- 404 link removed --
-- 404 link removed --

The 3rd test patters are a pal test chart from the belle nuit website. Which show the full resolution & full range. The super white part is all above 100 & the super black part is below 0.
-- 404 link removed --

the fourth test patter is a black/white picture with full range. In the range from 0-16 & 235-255 the font is written with range it ist.
-- 404 link removed --

the fifth test patters is in full resolution 720x576 & have some grids to compare.
-- 404 link removed --

the sixth test patters is the hello world test. hello world is only written in the super black/white part. It is useful to to understand if your hardware like dvd player/lcd can show the super black/white part. If you can't see hello world some of your hardware is clipping this part.
-- 404 link removed --

the seventh test patter are just some color bars.
-- 404 link removed --

Now I have tried to find out what captures card deliver a capture who looks like as close as possible to the original mpg file.
Limitations of this test are: I have made no adjustments during the capture and I have no dvd/blu-ray player ( I have tested about 17 different ones) who can play this test dvd correctly with all available outputs (composite,y/c (S-Video), component, hdmi). For example the Pioneer DVR-630 play it correctly but has no hdmi output. The Panasonic DMR-495 I have used offers all 4 outputs but the output from composite & Y/C are only in the resolution 704x576.

It might be useful or not. But it shows me that for example the default sharpness setting for the ATI USB 600 at 2 is oversharpening the picture. You can see sharpening artefacts. Here is the capture wtih sharpening setting at 0.
-- see attached Test File ATI USB 600 no sharpness.avi --

I have to correct something. I have said the capture from the ATI AIW AGPs is different and I have to stretch the files.
I have compared the 3 captures from the AIW 7200,9000 & 9600 again and the capture from the 7200 is close to the same resolution as the original mpg file. The capture from the AIW 9000,9600 I have to resize to the resolution to 700x576 to get as close as possible to the original mpg file. For this testing I have used the differenz filter in Edius 8.

The test files shows in my eyes that if you didn't adjust the settings during the capture that all usb card clip the super black range.

Another thing I have to say if you compare files is that you never should trust your eyes. If one capture have higher contrast it can look much better as another capture.

I must correct something about he Canopus NX.
The canopus nx exist as a PCI-X or PCI-Express card. With the card you can capture in YUY2 uncompressed and there is no internally compression to lossy dv codec before it capture to YUY2. To reduce the filesize from uncompressed I have use the lagarith codec after the capture with virtual dub. The card only works with Edius but offers an s-Video output where my studio control monitor is connected. You can find this card these days if you search a little for about 120,- Euro included Edius 3,4 or 5.

sanlyn asked how my workflow looks like?

Usually I capture with the canopus nx or the Blackmagic Shuttle (over HDMI) uncompressed YUY2 avi.
Then I made the adjustment for the broadcast safe range 16-235 with the edius YUV-Filter, zebra overlay & vectorscope,waveform monitor.
Edius should work internally with YUV and it should be no RGB conversation.
After this I cut and correct the color shift with the JPDSR Filter (which work in YUV colorspace) in Virtual Dub and save this master file with the lagarith codec. Then I do some noise reduction or ... with Avisynth. That's my way.

The test patters test file are only the first part of my testing I have do more test with real VHS video. With problematic tapes, tapes with macrovision. But you should remember the capture card is only on second or 3rd place in your capture hardware. Sooner or later you'll need an "TBC" and this could change the results again. Most people won't see big difference because they have nothing to compare.
My advice for PAL users is the Canopus NX (with Edius) if you could grab one for about 150,- Euro.
But you should make your own tests and at the end of the day it has all to do with money & availability
of the hardware in your country and if you want to use a laptop or desktop pc.

And now fire-free sanlyn & Lord Smurf

The YUV range for chroma is 16-240. If you use that range, you'll corrupt colors and lose details in the brightest areas in RGB. In any case, y=16-235 has more to do with proper display than broadcast safeguards. Saying it's broadcast-safe is an oversimplification.

Test pattermns are a big problem. The biggest problem is that test patterns neither look nor behave like real video. But some people love to stare at test patterns and pixel displays all day 24/7/365. I wouldn't last 10 seconds watching a test pattern, and still images of test patterns make me want to go to sleep immediately.

Thanks for mercifully not posting any samples of NX and Black Magic plastic videos. I've seen enough ugly video for one thread and really have seen too much ugly crap during the past year or so. I think I'd be better off staying away from newbie crap in a/v forums and just posting my own stuff, which are problems enough themselves. Noobs and BM fanboys never look at anything beyond their own stuff anyway, which is likely why they have so little to compare it to.

This is really depressing. BM? it just gets worse out there. I have a friend about 2 hours' drive away with a 35mm projection room and real film. I think I'll take a drive.

The attachments are corrupt.
Reattach. I'll merge and fix the post. Just let me know 1st/2nd/3rd/etc, to match the tests.

@sanlyn, stay tuned I will upload the same VHS samples made with different capture cards.

The biggest problem for me is your behavior about users who have made other experience as you. You call them fanboys, newbies or insult someone or just say the users haven't the same quality standarts as you. There are a few threads here on this forum where you do this. If you really want discuss and compare samples from different hardware with useful arguments your opinion is really important but if you just wanna insult someone or lower some hardware without testing it is better you keep silence.

I have started capturing vhs video about 20 years ago with a radeon vivo, radeon aiw, tv cards, canopus advc 50, dvd recorders, pinnacle usb, blackmagic shuttle, canopus nx, canopus advc 300 and a few other capture devices. I'm interested in how other capture cards works. Inspired from another forum thread here about a matrox capture card I have bought the Matrox Mini (for 11,00 Euro) to do some testing and to compare them with the other cards I own.

Makes perfectly good sense to me.
I'll be downloading some of mine, too.


A lot of users don't have quality standards, period, and wouldn 't know quality if it bit them in the tush. Lately we're seeing more and more dudes who are mighty proud of their mediocre material, are not ashamed to say so, and they get pissed if you don't congratulate them for nothing. Sorry, folks. My standards are those that I learned from the pros and by observing their work. Most of us can't always go all the way up to that level because we don't have the resources that the industry pros have; no one expects perfection, and I certainly don't. But rather than asking others to lower their standards to suit yours, maybe you should raise your standards to try to match criteria that advanced users have learned from. You don't need a forum to make ugly video -- anybody can do it. It's nothing to be proud of. And frankly, I'm getting very weary of looking at more and visual crap. It ain't good for your brain.

He gets grumpy at times. So do I, but it rarely shows. And sometimes it's warranted. We all have pet peeves. Video folks have video pet peeves. If sanlyn goes too far, I'll clean it up. If anybody else thinks he's gone too far, that what the report button is for, available on every post, and Site Staff will look at it. Same for any forum member, even Staff posts. (FYI: I hate referring to other forum members as if they're not here, reading the post. I'm sure you're reading this sanlyn!)

However, sanlyn does have a valid point, even if sometimes stated with gruff language. Don't dismiss it with soft language like "not the same standards".

Lots of people have this false idea that VHS transfer quality must look terrible. And that myth stems from user error, bad software, and bad hardware. Many people can be educated about this, and can then either DIY better, or select non-amateur services for conversions. For those folks, problem solved.

But some folks are real head-in-sand (head-up-ass) about video. Video has rules that must be followed, specs that exists. Every rule has exceptions, but some folks just do whatever the hell they want, complete disregard for norms of quality and acceptability. That includes not just users, but even hardware/software makers.

The real "WTF" moment comes from capture cards especially.

You can buy excellent SD cards for under $100. Yet it seems folks want to spend either pennies on known-junk Chinese KO cards (Easycaps, etc), or go the complete opposite and spend $200+ for a "better card" that actually isn't better either (HD cards like Blackmagic, Magewell). It's as if everybody wants to only put either homebrew fuel/gas/petrol in car -- or premium grade. And the vehicle will probably hate both of them. Yet there's perfectly good regular unleaded gas easily available. So why not buy it? That's the source of this.

Understand that I wish it was as easy as picking a card based on price and features, and all of them gave you good quality. But it's just not reality.

Thus you're not a newbie or fanboy.

"Newbie" can be a derisive term, and I favor "novice" instead. A novice is less experienced, teachable, probably eager to learn. A newbie often thinks he's suddenly an expert. But the term newbie is often inclusive of novices, hence wider use.

"Fanboys" are generally people that are defending their choice/purchase. The defense is not based on quality of the item, just defensiveness for defensiveness's sale. This is the main reason I can't use certain sites, because it's a cesspool of "facts" and opinions that contradict reality (and I refer to non-video topics).

That's really what matters. I'm interested to see your tests. :)

You have knowledge and experience to share at this site, and we're glad to have you. Also realize some of the information regarding capture cards is very NTSC-centric. PAL differs. So if you're able to give us good tests for PAL gear, post it. We'll all look at it.

I'll try to maintain some equilibrium, cut down on the coffee (I hate Keurig's stuff anyway), and will use "novice" when it applies. When it doesn't, I'll just try to avoid people who don't care. They won't learn or change their minds about anything anyway.

I'll be waiting for comparison posts myself, from which we can all learn much. I've been collecting samples of how to work with challenging sources from several of my own projects, including caps made with less than perfect players and 3 different capture devices -- I can appreciate the proposed card comparisons, considering the effort and delays involved in getting up my own posts.

Tired is why I get grumpy, not overstimulated. MS causes chronic fatigue. Not to OT too much, but tea started to give me headaches, diet cola upsets guts, regular cola too fattening, so I may have to start drinking coffee. And I've never liked coffee. This is the sort of stuff that sidetracks me. :depressed:

Nah, give them the benefits of the doubt. Sometimes newbies/fanboys can be educated back into novices. Some can graduate to various levels of competency, while a few even go on to become true experts. I've seen that happen many times. That's really what education is about. I'm really humbled, and honestly somewhat jealous of, a person I was mentor to for years, starting a decade ago. A few years back, he started to surpass me. I know my health is largely to blame, as I've somewhat stagnated in that specific area.

I've been collecting cards. Lots and lots of card. Taking samples as I have time, but not done. Having to use my dev system for capture right now, overflow. :depressed:

This is good stuff and information. Yeah there are many folks that just cheapens out with the modern capture stuff (like the ezcap junk at walmart) that we know obviously never cuts the mustard, but those that support such weak hardware line them up with praise and think its the best since sliced bread. I even seen some try to capture VHS with the Elgato capture devices, which is mostly designed for video game capture.

Ever since my VH days, I have came a long way when it came to capturing VHS material, and getting the right hardware to do the job, which I learned on this forum. It also helps to learn the weaknesses of the VHS format to where we can make the best capture as we can with what we have. Sure, the quality will never be HD standard; its simply not designed to be that way. I would rather watch older SD material the way it was originally presented, even if not the best. For me it just captures the nostalgia of what things were like back in the day. Folks like to stretch SD material to match their widescreen HDTV, which looks ugly. Unless the material was formatted that (amaphormic) way its always 4:3.

And though I have sold one of my XP based systems, I still have one that I will keep on hand, as I have alot of older hardware that works best with it. I know Black Magic products aren't that popular here, and with good reason since alot of their stuff is mostly based in incorporating HD workflows, but I think with alot of these newbies (and some fanboys) they don't realize that upper-end gear requires a clean, synced signal from SD sources despite their 'support', and most of them don't know how to do that properly and therefore get crappy results. I have done a few test captures with the AJA LHi capture card to DV50 and it turned out fine. I can always do the rest of the finetune in software.

TBC's are obviously the most important device in the signal chain to get clean transfers, with line based TBCs in the deck and the external full frame TBC working together to clean and stabilize the image. That's what most of the BM (and even AJA) gear look for when digitizing the signal, especially if going the SDI route.

I love using pro hardware, even back during the days where I couldn't even afford the stuff back in the 90's when this stuff was mainstream. But the desire was there. Nowadays, I'm able to afford more pro gear (via ebay in some cases) to where I can do my video the best way I can. With this type of hobby (or business for some who do it for a living) its important to recognize the material one is working with, and having the appropriate hard/software to get the job done right. It just depends on how deep in the rabbit hole you want to go once you get into this thing.

Not all capture cards capture exactly the same. You have one that captures perfectly, while you may have another that has a quirk to it, and one that requires alot of adjustments to get it right.

For me personally, I have alot of VHS tapes to work on, some from my late mother that I wish to keep. And once I am done with transferring and archiving the material, I will move on more towards HD workflows.

But one thing I do know, there is a wealth of information to be learned here, whether one just have a few months under their belt or those who have been in it for decades. The common goal most of us have is to preserve what we have on the legacy formats for the future as long as we're able. Trust me, in probably 10 or 20 years, SD capture hardware or the video decks may no longer exist, and you see how tough it is now to get them serviced with most of the techs retired or passed on.