VideoHelp Forum




+ Reply to Thread
Page 3 of 3
FirstFirst 1 2 3
Results 61 to 85 of 85
  1. Member BrainStorm69's Avatar
    Join Date
    Aug 2002
    Location
    Texas, USA
    Search Comp PM
    Originally Posted by Wilbert
    Could you make two more caps at 368x480 and 400x480? The differences between the screenshots should be different with these sizes.
    If I have the time tonight, I'll see about it. But I'm going to be pretty busy tonight.

    Originally Posted by trevlac
    @BrainStorm69

    Nice post. Personally I think you should show 352 resized up to 704 via lanczos vs a 704 cap. The TMPGEnc resize is probably a bicubic which may both sharpen and blur.
    If you can tell me what to use to do it, I can try.

    Originally Posted by trevlac
    But this goes to my point about practical. Test your full process and see what works for you. If you are going to make 352x480 dvds, looks like a 352 cap is better.
    I'm confused by this statement except with respect to the Philips chip.

    Originally Posted by trevlac
    BTW: The BT878 352 looks so bad because there is a vertical filter. Try 360 vs 368 on the BT. BIG difference not due to 8 pixels.

    Also ... it is reasonable to cap at 368 with the BT and BTWincap drivers ... and just crop to 352. The AR will not be off as much as you would think. This is because the BTwincap (NTSC) caps about a 0.6 wider window than is used in the 352 size. This translates to about 4 pixels. So the AR is off by about 12 pixels not 16. VirtualVCR lets you crop as you cap.
    I'm aware of the Vertical filter and have capped at 368x480 cropped to 352x480 in the past with the AverTV Stereo. Right now I'm mainly using the card with the Philips chip. Interestingly, the 352x480 cap looks just a little bit (not much at all) better to my eyes with this card in this test.
    Quote Quote  
  2. @junkmalle

    Let me start by saying i enjoy this type of sharing of perspective and info. This is how I learned this stuff. I'm still learning a lot.


    Originally Posted by junkmalle
    Originally Posted by trevlac
    b) Samples(pixels) do not have to line up with peaks.
    Of course not, that was my whole point. I chose those two algnments as extreme examples. Repeat it with any any phase and aperture size you like. You will see the following. As you slide the wave sideways the picture oscillates between high contrast black and white pixels (my first example), through less and less contrast between the pixels, and finally at 90 degrees (1/4 wave) the contrast is completely gone (ie, the solid gray in my second example). It then starts to cycle back to more and more contrast between the pixels until you reach the next peak in contrast at 180 degrees. I leave it as an exercise for the reader to see what happens when the signal frequency differs by a small amount!
    I may still be missing your point. But I think you missed mine. Samples may not line up with the peaks, but they do not need to in order to calculate the peak values. As long as there are 2 samples for a full cycle, and they are not zero (then you need more than 2). If that is the case, only 1 sine wave will cross the points. The theory is based on the ability to break down a complex wave into sine and cosine waves.

    Your example of 2 grey pixels breaks the >2 rule. If the pixels are a little white and a little black then as long as your wave is the highest frequency, only 1 frequency sine wave can connect the dots.

    I also think in your example you are making a logic jump that 2 adjacent pixels with 'low contrast' can not represent analog peaks with higher contrast. This is a good and subtle point, and I will have to think of an example to show (or not) how low contrast pixels can translate into peaks in the analog. The question is, if pixels are point measures ... what is in between.


    Yes, this is the Nyquist theorem. But that doesn't mean that at 1.99x you have crap and and 2.01x you have a perfect representation of the original analog waveform. It is a rule of thumb that says the highest frequency that can appear is 1/2 the sample rate. It doesn't say that it takes a 2x sample rate to perfectly represent the waveform.
    True. And as far as I know ... reconstruction of the original waveform is much more important than small differences in the sample rate. This why I say not to resize too many times. Each time, the wave is reconstructed and then resampled.

    The question does come up then "In practice how much higher do you need?". To given the reconstruction room for error and such. I certainlly don't think it is 200%. If VHS can contain a max of 320x480 you don't need 640x480 unless the reconstruction method is really poor. I threw out 10%. 352 gives 10% above 320. Besides, 320 is probably never reached. 352 gives 15% room over 300.


    In any case Trevlac, thanks for the perfectly civil discussion. I've seen so many threads like this that break down into huge flame wars!
    The pleasure is mine. I really post this stuff because it helps me shake out what I truely don't know. If I can prove and explain it ... I have a good handle. I'm certainly no expert ...

    BTW: Your arguments fall in line with what my friend Arachnotron says on the topic. He ran a bunch of tests with patterns and made measures on scopes. It seems if you sample < 4x you get a pattern in the samples that is at a different frequency than the frequency of the test pattern. But it does not make any sense that you have to sample at 4x or DVD could not contain above ~ 4.5MHz. And the test DVDs clearly do. I think the answer to this lies in what pixels really are and how the analog is reconstucted.

    I'd point you to our discussions over at doom9, but a lot of untrue statements were made as we were exploring the topic. (Especially by me ) So I don't want to confuse my current points and understanding with those.
    Quote Quote  
  3. Member BrainStorm69's Avatar
    Join Date
    Aug 2002
    Location
    Texas, USA
    Search Comp PM
    I've just realized that my post with the caps faulty for the true purpose of this thread -- resolution for VHS capture. I was somewhat caught up last night just seeing the actual difference in the cards at the resolutions, so I capped straight from DVD. If I can get the time tonight, I'll record the DVD menu to VHS and repeat the tests with caps from the VHS tape.

    It should be very relevent to this thread to compare the VHS caps to each other as well as the DVD caps.

    Sorry if I mislead anyone.
    Quote Quote  
  4. Originally Posted by BrainStorm69
    Originally Posted by trevlac
    @BrainStorm69

    Nice post. Personally I think you should show 352 resized up to 704 via lanczos vs a 704 cap. The TMPGEnc resize is probably a bicubic which may both sharpen and blur.
    If you can tell me what to use to do it, I can try.
    With the ATI and Philips cap at 704 and 352. Resize the 352s back up to 704 using the Lanczos option in the built in VirtualDub resize filter. Compare the BMP results. If there are differences, make some compressed 704 images for us to see and make sure the compression doesn't change things too much. If there are no differences ... just tell us. You can also cut out little parts to post. Like the words.

    Your DVD menu is a nice 'real world' source. The resolution pic played from DVD will certainlly have differences. I doubt there are differences in any VHS source. Even the test pic recorded to VHS via DVD. Edit:{Except for the BT vertical problem }

    Originally Posted by BrainStorm69
    Originally Posted by trevlac
    But this goes to my point about practical. Test your full process and see what works for you. If you are going to make 352x480 dvds, looks like a 352 cap is better.
    I'm confused by this statement except with respect to the Philips chip.
    I don't see much difference in anything but the BT878. So for ATI and Philips, I'd do 352 caps so TMPGEnc does not have to resize.


    Originally Posted by BrainStorm69
    Interestingly, the 352x480 cap looks just a little bit (not much at all) better to my eyes with this card in this test.
    Me too. I would say this is because Either the Philips or TMPGEnc or a combo of both is sharpening the picture. You can do the same thing with some DVD players and the Sharpness control on a TV.

    But most pleasing is very important ... for example. I have a $1800 professional DV capture device. It captures more detail in the resolution test pic than anything else I have ever tested. It also has a problem where the colors are clamped 25% too low. The pictures look washed out and dull.

    My point here is that resolution is only 1 factor. Personally, I think proper color is more important after a certain point. I want a vibrant and rich picture. Luma resolution only gets me so far.

    PS: It's nice of you to do these tests. Hopefully they may help some people. I certainly don't know how all this stuff works. I've been learning for a while and find even the worthless (to me) details interesting. For example, I don't have a philips chip card. But your process does make that philips 352 look good.
    Quote Quote  
  5. Video Restorer lordsmurf's Avatar
    Join Date
    Jun 2003
    Location
    dFAQ.us/lordsmurf
    Search Comp PM
    If you're worried about VHS detail, buy one of these:
    http://www.signvideo.com/dr1000_image-enhancer_video-processor.htm

    And if you're worried about color, buy one of these:
    http://www.elitevideo.com/index.asp?PageAction=VIEWPROD&ProdID=8
    Want my help? Ask here! (not via PM!)
    FAQs: Best Blank DiscsBest TBCsBest VCRs for captureRestore VHS
    Quote Quote  
  6. I don't see much difference in anything but the BT878. So for ATI and Philips, I'd do 352 caps so TMPGEnc does not have to resize.
    You don't see any difference in 704 and 352 for the ati screenshots? I suggest you take a closer look at them
    Quote Quote  
  7. Originally Posted by lordsmurf
    If you're worried about VHS detail, buy one of these:
    http://www.signvideo.com/dr1000_image-enhancer_video-processor.htm

    And if you're worried about color, buy one of these:
    http://www.elitevideo.com/index.asp?PageAction=VIEWPROD&ProdID=8
    I thought about getting one of those Elite BVP4's. You can pick them up one ebay for a little over $150 used. The only reason I would want one is to use it as a proc amp in analog. If the signal is way out of wack ... there is a problem with fixing it in digital. Namely that you create gaps and get banding if you don't have much to work with.

    Other than that, I'm not sure they are worth the money and desk space, because you get much more control in digital. I don't have a BVP4 and have read good things. But I have also read people saying they are not worth the money ($700 new I think). I don't really know, but I'd think I could get similar results in VirtualDub.

    PS: I sent my Laird back to the manufacturer and asked them to fix the color problem.
    Quote Quote  
  8. Originally Posted by Wilbert
    I don't see much difference in anything but the BT878. So for ATI and Philips, I'd do 352 caps so TMPGEnc does not have to resize.
    You don't see any difference in 704 and 352 for the ati screenshots? I suggest you take a closer look at them
    OK Ivo I put them side by side in the browser still don't see much difference. They are both 352 so any difference should be due to the TMPGEnc (bicubic ?) resize vs the ATI (bilinear?). Cut out a piece and show me where to look.

    PS: I think at 704 I would see a difference.
    Quote Quote  
  9. Video Restorer lordsmurf's Avatar
    Join Date
    Jun 2003
    Location
    dFAQ.us/lordsmurf
    Search Comp PM
    Having used one, I'd actually have to disagree. I think the hardware controls (for color correction and sharpness) work far better than the digital ones.

    Once you go digital with analog source, your options are limited. While still in analog land, you've got a bit more headroom.

    I will say the "res boost" isn't all it's cracked up to be, but these units have managed to give the appearance of higher res, if not outright supplying higher res.

    I'm a fan.
    Want my help? Ask here! (not via PM!)
    FAQs: Best Blank DiscsBest TBCsBest VCRs for captureRestore VHS
    Quote Quote  
  10. @junkmalle

    Here is the source: http://www.bores.com/courses/intro/basics/1_antia.htm

    Here is the picture of interest:


    The top signal has 50 cycles. The next item has 100 samples. The 3rd line shows the samples 'reconstructed' using straight lines. The 4th shows reconstruction using a sinc.

    I believe your exmple is a discusion of line #3. This assumes samples are connected by straight lines. As you 'slide' the wave across the samplign points you have high contrast if your samples fall on the peaks and 2 almost grey pixels if you go to the other extream.

    Going along with this, one has to 'believe' that 2 grey pixels next to each other actually can represent a peak. The only way I can make this jump is to view pixels as measures that need to be taken in groups of 12 or more. So the 2 don't show a peak. You need a bunch around them to calculate it.

    Frankly ... I'm still working on this in my mind. The reconstruction has a great deal to do with the results ... and I don't know how real devices handle this (my TV/DVD Player / And PC Monitor - Graphic Card).

    I've talked to quite a few people about this and been given tons of grief. A reasonable approach to the practical answer is to just test and look at your results. But I also like to know the theory so I can use it as a guide.

    --------------
    Here is a discussion about this on doom9. I'd start on page 5 and go to Arachno's 3rd post down (on nyquist).

    http://forum.doom9.org/showthread.php?s=&threadid=67287&perpage=20&highlight=post%20al...g&pagenumber=5

    As I said before ... there is a lot of bunk in there ... On the 6 tap 32 phase resizer that the BT uses ... I did finally figure out what that means. I wrote a vdub filter to play with the idea. It's basically like a Lanczos3 resize. Xcept I don't know what filter taps they use.
    Quote Quote  
  11. Originally Posted by lordsmurf
    Having used one, I'd actually have to disagree. I think the hardware controls (for color correction and sharpness) work far better than the digital ones.

    Once you go digital with analog source, your options are limited. While still in analog land, you've got a bit more headroom.

    I will say the "res boost" isn't all it's cracked up to be, but these units have managed to give the appearance of higher res, if not outright supplying higher res.

    I'm a fan.
    OK I'll take your word on it. And I agree on the 'headroom' thing, it's just the ease of use/pratical difference I wonder about.

    Most of the stuff I play with is already mpeg2 or DV. It was hard to justify the Laird for DV in alone. I use it for DV out back to my monitor when adjusting and editing.

    Like I said. I've been looking at them. I'd pay $150 for the newer + model just to see. Another item I find interesting is one of those Hotronic pro level TBCs. They of course have all the proc amp controls too. Not sure if they effect digital or analog.
    Quote Quote  
  12. Member BrainStorm69's Avatar
    Join Date
    Aug 2002
    Location
    Texas, USA
    Search Comp PM
    OK, here they are one more time, this time the same DVD menu recorded to VHS on my Mitsubishi HS-U781.

    ATI 352x480


    ATI 704x480


    Conexant/BT878 352x480


    Conexant/BT878 704x480


    Philips SAA7130 352x480


    Philips SAA7130 704x480


    EDIT: For those that want to see the BT878 without the vertical filter...here it is at 368x480 cropped during capture to 352x480. Notice that the BTWinCap drivers are artificially sharpening the picture at the default setting of 50 sharpness, the "auto" box unchecked

    Conexant/BT878 368x480
    Quote Quote  
  13. Member NamPla's Avatar
    Join Date
    Jun 2004
    Location
    Whoop Whoop
    Search Comp PM
    How do I know what chip my card uses? I got a Winfast 340 (I think it's a GeforceFX?), and I "think" it uses a Phillips chip.

    If that's so, then can I safely assume my card is a good "all-rounder" and can capture VHS/TV equally well at 352/704/720?

    Judging from BrainStorm69's examples, there's no real difference at all between the 352 and 704 captures (with the Phillips-based card) - which agrees with my own testing.

    Thanks. (Interesting topic!).
    Quote Quote  
  14. Hmm, a whole bunch of posts disappeared
    Quote Quote  
  15. Video Restorer lordsmurf's Avatar
    Join Date
    Jun 2003
    Location
    dFAQ.us/lordsmurf
    Search Comp PM
    I don't even remember what they were. I think the bulk of it is left, as long as the images folder gets back online. The forum is fubar right now. I'm not even gonna use it until it's back to normall 100%.
    Want my help? Ask here! (not via PM!)
    FAQs: Best Blank DiscsBest TBCsBest VCRs for captureRestore VHS
    Quote Quote  
  16. Member BrainStorm69's Avatar
    Join Date
    Aug 2002
    Location
    Texas, USA
    Search Comp PM
    I hope the pics haven't been lost. I deleted them all from my hard drive.
    Quote Quote  
  17. I hope the pics haven't been lost. I deleted them all from my hard drive.
    Well, now 9-13 and I just got back around to reading this thread again and the pics ain't showing for me now!

    Anyway it looks like after all the reading and discussing that went on here while i been gone seems to lean towards 352X480 now with good capture device and software ect... though lower quality capture devices may not do well at that setting because of resizing and filters, right?

    I never saw the pics, but reading the comments about them it seems VHS capture at 352 was equal to the same capture at 720 for quality, correct?

    Although I have not been around much for a few weeks, I did do some testing on this myself. No pics to offer though.
    My captures at 352X480 from a VHS-C camcorder at about 5000BPS look as good as the ones I did from the exact same tape at 720X480 and 8000BPS. Niether I nor anyone else can SEE the difference when viewing them on the PC. After burning to DVD and watching on the PC still the same! I did not burn the 720X480 though, but that should not matter.

    ALSO, I just snagged a GREAT deal on some older camcorders! 1 VHS tape camcorder, works good but bad battery. Used external P/S for it.
    When recording to the tape the quality is lower it seems than when capturing directly to the PC. What I have done is connect the camcorder line outs to my ATI AIW Radeon 7500 and capture directly from the camcorder WITHOUT recording to tape. 352X480 still looks the same as 720X480 capturing directly from the line out! Most likely dirty heads, but when recording to the brand new VHS tape the quality of the tape is lower than my direct captures.

    If people want to save power they should learn what an OFF button is for and forget all this power saving crap! My problem is that I can turn on the camcorder and it will run about 15 minutes without recording (sends signal through the line out durring this time ) but then the darn thing goes into standby power saving mode and shuts off the viewfinder and of course stops sending any signal also! Dam I hate power saving crap!

    So what I found for ME is that 352X480 captures are equal quality with my equipment as 720X480 captures (cannot SEE any difference). AND either my camcorder has dirty heads OR capturing analog signal direct from the camcorder is better quality than recording to tape first then capturing the tape! Actually I THINK both are true! A tape most likely suffers a signal loss durring both record and playback compared to no siganl loss with a direct capture. Well other than cable and conection losses but you have those both ways.

    Can't complain too much if I need to clean the heads on a $10 camcorder
    My other $10 camcorder I am not sure about yet, it is a Panasonic camera
    that connects to a portable VCR. Never saw something like this before, either a more profesional model or really old type like when they first started making them? It has no way to record to tape or power itself! It connects to a VCR with a 10 pin connection.
    I did find the portable Panasonic VCR ($5) in with some other stuff, Also an RCA VCR that is strange. The RCA looked like a normall VCR but older topload and had the camera connecter on the side too ($5 also) so I bought it too. What I found is that one is 2 piece. Portable VCR and docking station! I noticed the bottom was loose when I picked it up, figured it had broke and someone took it apart for some reason but got it for parts anyway. When I got home I found instead the Tv tuner and clock (left side) is part of the docking station and all the VCR controls are on the portable VCR. The bottom was not loose, it was of course the bottom of the VCR and top of the docking station instead.

    SO far I did not like the picture though but haven't had time to mess with it much yet. Except for batteries everything worked, but then the RCA portable went dead, Hopefully just a loose power connection.

    So anyway, looks like I'll be doing alot more recording/capturing/burning/playing again soon. Anyone want to see the deer in my yard??
    overloaded_ide

    Spambot FOOD
    Anti-Spam
    Quote Quote  
  18. We haven't had one of these threads for a while... Reminds me of the good old days of VCDHelp...

    Simply put, unless there is a PROBLEM with capturing at a higher resolution (e.g., HDD constraints, insufficient bandwidth --> dropped frames, etc.), capturing at the higher resolution (and then doing software resizing) will not look any WORSE than capturing at a lower resolution.

    If you are using a resolution higher than the actual amount of video information (i.e., you are oversampling), the only effect on the quality is either NONE or BETTER (through a better SN ratio... i.e., less noise).

    The arguement AGAINST using a higher/highest framesize is really one of efficiency. Generally, you will need more resources, the process will be more difficult and the process will take longer. Depending on your setup, there may be minimal quality advantage.

    Regards.
    Michael Tam
    w: Morsels of Evidence
    Quote Quote  
  19. I still have one of the pictures I posted. It's a comparison of my Hauppauge WinTV PVR-250 capturing a resolution test pattern (referenced earlier in this thread) from VHS tape at 720x480 and 352x480. The 720x480 capture does show a little more detail:



    Enlargements and reductions were performed with VirtualDubMPEG with the Lanczos3 filter. This JPEG file is very close to the original BMP extractions. For those of you who don't know, the PVR-250 captures directly to MPEG only.

    Note that the visible alternating lines in the "300 line" area of the 720x480 capture is a moire pattern, not accurate detail from the original image.
    Quote Quote  
  20. If you are using a resolution higher than the actual amount of video information (i.e., you are oversampling), the only effect on the quality is either NONE or BETTER (through a better SN ratio... i.e., less noise).
    If you would have seen the screenshots above, you will see this statement is false for bt8x8 (sharp border at 368 ntsc / 384 pal) and ati chips (no sharp border). The specific reason can be found elsewhere in this thread.

    Something more general, explaining why it doesn't contradict with Nyquist. Quoting someone from Ars

    "More than that, Nyquist specifies the minimum sample rate required, under ideal conditions, to mathematically reproduce the original bandwidth limited signal. Since we have to physically reproduce the signal (under non-ideal conditions), we are pushed further from Nyquist's minimum. Things like non-ideal front-end anti-aliasing filters end up increasing the required frequency bandwidth"

    I hope BrainStorm can redo the screenshots Please
    Quote Quote  
  21. I stand by what I stated before as a truism...

    The caveat I gave was "unless there is a problem". If the driver handles scaling badly or inconsistently, you may well find "optimum" capturing resolutions.

    Assuming everything else is the same, there shouldn't be any worsening of "quality" in capturing at a higher resolution. There just may not be any benefit.

    As an audio analogy, it would be like capturing AM radio at 44.1 kHz sampling rate vs. 22.015 kHz. At 44.1 kHz, it won't sound any worse (and may possibly be better).

    Regards.
    Michael Tam
    w: Morsels of Evidence
    Quote Quote  
  22. Member BrainStorm69's Avatar
    Join Date
    Aug 2002
    Location
    Texas, USA
    Search Comp PM
    [quote="Wilbert"]
    I hope BrainStorm can redo the screenshots Please
    I'll see what I can do as soon as I have the time (I'm pretty busy right now). I've also uninstalled the AverTV Stereo, so I'll have to reinstall it.
    Quote Quote  
  23. This thread has probably been the absolute best information I have found on the net. so far. As some of you know, as a few of you have helped me quite a bit, I am looking to convert VHS tapes(which are already not the best quality..multiple generation concert videos) I have to digital format. From what i've gathered and understand I should capture at one of the following:

    Using ATI's MMC:

    mpeg at 352x480 using huffy encoder with 288 line field threshold, predict gradient RGB compression, Predict median YUY2 compression, or maybe a resolution a notch higher than burn it to svcd.

    OR mpeg4 at 640x480 encode interlaced, VBR max bitrate 8, target bit rate 8, motion estimation quality 100, audio encode 296kb. Then use TMPGenc to bring to to svcd format.

    Now, after that is done..either should be ready to go to svcd, right?

    If I wanted to go to xvid or divx, I could do that right off the bat by choosing AVI then the xvid codec, right?

    Again, all the help has been appreciated!
    Quote Quote  
  24. Member BrainStorm69's Avatar
    Join Date
    Aug 2002
    Location
    Texas, USA
    Search Comp PM
    Since this thread got brought up again, and I have my AverMedia AverTV Stereo (BT878) installed at the moment, I decided to re-do the VHS caps that got lost when Baldrick switched servers.
    Quote Quote  
  25. The Old One SatStorm's Avatar
    Join Date
    Aug 2000
    Location
    Hellas (Greece), E.U.
    Search Comp PM
    I suggest a new post if possible! A link from here to the new post may help those that seek info about those subjects
    La Linea by Osvaldo Cavandoli
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!