Real-time analog to DV capture - what is the right device to use?

[Serge_KB]

Hello all,

Having analog video signal (composite PAL, with nonstandard resolution of roughly 800x600) as the input, I need to encode it to digital video without much compression in order to display it in real-time in my own software application.

The most crucial points are the following:
- the quality of the digital video should be as good as it is possible, so encoding should use smth like DV for compression (or no compression at all)
- the time lag should be as small as possible, small enough to be unnoticeable at all.

What is the right hardware for this task?
Does the Canopus ADVC300 converter, for example, introduce any noticeable time lag during encoding?

[edDV]

Hello all,

Having analog video signal (composite PAL, with nonstandard resolution of roughly 800x600) as the input, I need to encode it to digital video without much compression in order to display it in real-time in my own software application.

The most crucial points are the following:
- the quality of the digital video should be as good as it is possible, so encoding should use smth like DV for compression (or no compression at all)
- the time lag should be as small as possible, small enough to be unnoticeable at all.

What is the right hardware for this task?
Does the Canopus ADVC300 converter, for example, introduce any noticeable time lag during encoding?

Analog PAL should be 576 active scan lines (625 total). DV records 720 horizontal samples (nyquist bandwidth 6.75 MHz) with non square pixels. When stretched out to 16:9 aspect ratio, the square pixel display equivalent becomes1024x576.

The ADVC 100/110/300 models encode analog to DV in hardware with near zero delay.

[dvd3500]

Agree with the above...

[richared1979]

I have a very similar question. I use an ultrasound apparatus that produces an S-VHS stream (600x800 px). I need to capture the stream in a specialized software (Noraxon MyoResearch XP) that requires DV format. I do not need any image enhancing features, but I need zero delay (or, if necessary, a standardized/measurable delay), and as good output quality as possible.

Is the Canopus ADVC300 the right device, or are there other options? Many reviews of that device focus on video quality enhancements, but I don't need that. Suggestions?

[sanlyn]

The most crucial points are the following:
- the quality of the digital video should be as good as it is possible, so encoding should use smth like DV for compression (or no compression at all)

DV is always compressed.

[edDV]

...
I use an ultrasound apparatus that produces an S-VHS stream (600x800 px).

S-Video is analog Y/C (480 active lines NTSC, 576 active lines PAL). Analog video doesn't have pixels, just lines. Your machine may display 800x600 in VGA, but in the D/A process gets reduced to 480 or 576 scan lines for S-Video with ~5MHz analog bandwidth (resolves to about 400 "horizontal lines of resolution").

...
I need to capture the stream in a specialized software (Noraxon MyoResearch XP) that requires DV format.

So you are saying you want to capture analog NTSC/PAL with A/D to DV format. That is what the Canopus ADVC does. The resulting DV video is 720x480 or 720x576 but the effective resolution is limited to the source analog bandwidth. In other words, you won't be able to resovle much more than 400 luma changes across a line. The exact resolution depends on the analog output circuits in the device. The max Canopus ADVC Nyquist limit is 6.75 MHz or about 540 luma changes across a line.

[2Bdecided]

ADVC110 is preferable to ADVC300 for a clean source you don't want to "improve".

Cheers,
David.

[aedipuss]

i've not used a ADVC300, does it have adjustable dnr/tbc or a way to turn them off? not that it matters much now that they are not made anymore....

[vito321]

Is there a handheld or relatively small video capture device that can be used to convert the analogue output of an ultrasound machine to digital and record onto sd card with reasonable quality???

thanks

[edDV]

i've not used a ADVC300, does it have adjustable dnr/tbc or a way to turn them off? not that it matters much now that they are not made anymore....

You can turn all the "features" off. Then it is effectively an ADVC-100/110.

[aedipuss]

oops never mind mixed up mb/s and MB/s again.

[richared1979]

...
I use an ultrasound apparatus that produces an S-VHS stream (600x800 px).

S-Video is analog Y/C (480 active lines NTSC, 576 active lines PAL). Analog video doesn't have pixels, just lines. Your machine may display 800x600 in VGA, but in the D/A process gets reduced to 480 or 576 scan lines for S-Video with ~5MHz analog bandwidth (resolves to about 400 "horizontal lines of resolution").

Sorry, I should rephrase... I have an ultrasound apparatus that internally stores AVI videos of 600x800 px. I need to record the video stream on a computer. This necessitates the use of the S-VHS output on the ultrasound machine, which causes the image to be reduced to (480 or) 576 scan lines. Is that correct? I'm pretty sure this system gives 576 scan lines, at least that's what we got on the computer earlier on.

...
I need to capture the stream in a specialized software (Noraxon MyoResearch XP) that requires DV format.

So you are saying you want to capture analog NTSC/PAL with A/D to DV format. That is what the Canopus ADVC does. The resulting DV video is 720x480 or 720x576 but the effective resolution is limited to the source analog bandwidth. In other words, you won't be able to resovle much more than 400 luma changes across a line. The exact resolution depends on the analog output circuits in the device. The max Canopus ADVC Nyquist limit is 6.75 MHz or about 540 luma changes across a line.[/QUOTE]

I think I understood that... At least I understood that the Canopus ADVC (probably 110) should do what I need.

Thanks everyone!

[Cornucopia]

If your apparatus internally stores the files as 800x600 AVI, why not let us know the model #, etc? Maybe we could give you tips on how to DIRECTLY access those original AVIs (as opposed to D->A->D conversion, resizing, and it's accompanying generational quality losses)...
Yes, you might still have to convert (from some less-used codec to a more common one), but you'd save on the other types of losses.

Shortest path, and all that. Nearly ALWAYS better.

Scott

[mamoltub]

Hi Scott,
(I posted as richared1979 the other day, now I'm able to access my own user)

Thanks for your interest, I appreciate it a lot. The ultrasound device is a Philips HD11 XE. The AVI (or DICOM) files that are recorded in the ultrasound device can be exported by USB after the data aquisition is over. When I do so, I get high quality video.

However, I need to synchronise the ultrasound video with some other data (forces, EMG, joint angles, etc). The trigger system for the ultrasound device is very impractical, as it eats too much of the recording time. (The ultrasound device is limited to recording 399 frames per trial.) This means that using the trigger to synchronise (or rather time-match) video recording on the ultrasound device and other data recording on the PC is not an option. Instead we capture the ultrasound videos in the same computer software that records the other data.

I guess that it might be possible to replace some of the hardware in the ultrasound device, so that the video can be streamed in its original format (without converting to an analog signal?). Unfortunately I'm not very skilled at hardware questions, and our Philips contact has not been able to answer our questions.

If you have any suggestions on how to improve this setup, I'd be very thankful.

[Cornucopia]

I find this quite interesting, but it looks like an online user manual for that equipment is NOWHERE to be found on the internet (without resorting to unsavory means)! And I tried...

It could be that your device could be retrofitted with a much larger storage capacity, such that you could take longer segments than just 399 frames (~= 12 seconds?). Although, sometimes those kinds of things are locked-in for other reasons.

You WANT to get as clean a signal as possible, and it's clear from the MyoResearch literature that DV works well with their stuff, so it would make sense to follow edDV's advice and go with a ADVC-type box (Analog -> Digital using DV codec). The choices are limited since DV is somewhat passé, but you should still be able to find ADVC stuff on ebay, etc. if nowhere else.

If S-Video is the BEST that your Phillips machine can stream out in realtime, then that's what you'll have to go with. Make sure the ADVC box has S-Video inputs then. And, it will be 720x576 as suspected. Hopefully, that will be good enough for your needs.

I don't know if this is possible, but have you considered doing BOTH the live export stream & the triggered file storing? Then, if you have time, you could match/synchronize the two in an NLE, and basically "replace" the streamed+converted shots with the available (and higher quality) DicomAVI shots?

HTH,

Scott

[mamoltub]

Thanks again for your advice, Scott. I was able to find an ADVC 55, which does the same job as my previous setup, with much less cables and settings. For some reason the videos are about half the size, though to the eye, I don't see much of a difference except slightly darker images.

The ultrasound machine does have a DVI port, for connecting a second monitor. We are looking into grabbing that signal and converting it to a data stream. But that's probably another thread.

[Cornucopia]

Right off the top of my head, I'd suggest: DVI out -> DVI-to-HDMI adapter -> HDMI cable -> Blackmagic Intensity Pro HDMI capture card.

Should be much cleaner/clearer and possibly higher rez than going Dig -> Analog -> Dig again...

Scott