about analog pass-through capturing method...

[edDV]

Well capture quality and a TBC are two different issues.

They go hand in hand but yes it is true that some digital camcorders have built-in TBC (or at least TBC like) enhancement of analog input.

However DV AVI is still DV AVI and not necessarily the best choice for VHS captures. This has been covered before.

edDV probably has a link.

It all has to do with the 4:1: color sampling issue BS.

- John "FulciLives" Coleman

I don't think there is a 4:1:1 problem for VHS capture. I've been arguing the opposite. The issue is how MPeg2 encoders process the 4:1:1 into 4:2:0 (i.e. the interpolation used). It isn't a big deal anyway.

It was discussed here
https://forum.videohelp.com/viewtopic.php?t=299086&postdays=0&postorder=asc&start=30

[FulciLives]

I forgot the person asking this was in a PAL video country so yeah perhaps DV AVI is the best option afterall considering he either keeps his current DV cam or wants to purchase a Canopus ADVC-110 or Datavideo DAC-100 etc.

If he must sell his DV cam then a HARDWARE MPEG-1/MPEG-2 capture card/device will be cheaper than a Canopus ADVC-110 or DataVideo DAC-100.

- John "FulciLives" Coleman

[kamaleon]

Yes indeed. Anyway i won't be buying or selling my cam in the next month (well at least i don't think i will), so that'll give some more time to hang out with you lot an learn some useful stuff

Anyway, about my question on the external box that could eventually grab in mjpeg/hffuyuv, any ideas anyone?

[edDV]

Yes indeed. Anyway i won't be buying or selling my cam in the next month (well at least i don't think i will), so that'll give some more time to hang out with you lot an learn some useful stuff

Anyway, about my question on the external box that could eventually grab in mjpeg/hffuyuv, any ideas anyone?

Uncompressed capture through a USB2 port isn't going to work except for a very tiny picture. For instance, 10bit 720x576 4:2:2 SMPTE-259M as used for Serial Digital Interface SDI is 270 Mb/s sustained. This includes audio. 8bit would be ~80% of that or ~216Mb/s.

Working from the other direction uncompressed 4:1:1 would be ~125Mb/s (no audio).

These rates are too fast for sustained USB2. Some form of upstream hardware compression is needed to bring down the data rate. External hardware devices exist for MPeg2, DV and MPeg4. I'm not aware of devices (other than internal PCI cards or another computer) that would perform MJPEG or hufyuv compression in an external device.

[kamaleon]

When you say "uncompressed capture" do you mean uncompressed rgb? sorry but i'm a bit of a newbie and i lack some understanding of basic principles on analog and digital video.
I've recorded a tv show with my hauppauge wintv-usb2 as "uncompressed rgb" @352x288 1411kbps.

[edDV]

When you say "uncompressed capture" do you mean uncompressed rgb? sorry but i'm a bit of a newbie and i lack some understanding of basic principles on analog and digital video.
I've recorded a tv show with my hauppauge wintv-usb2 as "uncompressed rgb" @352x288 1411kbps.

RGB or more likely YUV 4:2:2 or 4:1:1
http://en.wikipedia.org/wiki/Luminance_(video)
http://en.wikipedia.org/wiki/CCIR_601

Most USB2 ports, unless purpose built cards top out between 24 and 80 Mbps sustained and consume considerable CPU cycles. In rare cases 160-240 Mbps sustained can be reached but not with typical computer ports.
http://en.wikipedia.org/wiki/USB#Speeds

[edDV]

For 352x288 = 101,376 pixels

@8bit RGB = 304,128 bytes

x25 frames/s = 7,603,200 bytes/s or 60.825 Mb/s

Correct?

[jagabo]

For 352x288 = 101,376 pixels

@8bit RGB = 304,128 bytes

x25 frames/s = 7,603,200 bytes/s or 60.825 Mb/s

Correct?

And in all likelyhood YUY2 is being sent down the USB cable and the computer is converting to RGB. So the data rate would be 2/3 that.

[edDV]

For 352x288 = 101,376 pixels

@8bit RGB = 304,128 bytes

x25 frames/s = 7,603,200 bytes/s or 60.825 Mb/s

Correct?

And in all likelyhood YUY2 is being sent down the USB cable and the computer is converting to RGB. So the data rate would be 2/3 that.

Yes and 720x576/25 would be 4x that (162.2Mb/s with no audio) so you see why compression at the capturing/sending device is useful and needed. An MPeg2 hardware encoder would send 720x576/25 @ 6-9 Mb/s (15Mb/s tops) including audio.

A DV encoder passes 720x576/25 DV format over IEEE-1394 at 25Mb/s for video + a bit more for audio, metadata and transmission overhead.

[kamaleon]

Ok, i'll get back to this later on as i need some time to digest all this info

Anyway, regarding my original question... the analog-to-digital converter used on camcorders, is it the same processor that processes the image from the lens (the chip that can actually "film" or is it just used to convert the analog inputs if there are any?

[jagabo]

Anyway, regarding my original question... the analog-to-digital converter used on camcorders, is it the same processor that processes the image from the lens (the chip that can actually "film" or is it just used to convert the analog inputs if there are any?

No, the information coming from the CCD is already digital. The A/D for converting analog video will be a different chip (or at least a different section of the chip). The DV compression will probably be done by the same chip.

[kamaleon]

No, the information coming from the CCD is already digital.

Right on. Does this apply to all digital camcorders? (like the cmos ones, hi8, etc).

The A/D for converting analog video will be a different chip (or at least a different section of the chip). The DV compression will probably be done by the same chip.

Are these chips we're talking about the same that we can find in tv tuner/capture cards?

[edDV]

This little graphic shows a highly simplified view for DV 16:9 mode.



A single optical/CCD assembly divides the pixels into R,G and B. The CCD sensor(s) exports RGB to the processor. The processor filters the image and then sizes to 720x576. Then the RGB is matrixed to Y, Cb, Cr components. Next this "raw" video is compressed to DV format. DV encoding compresses each frame about 5x but each frame retains integrity.

In 16:9 mode, the 16:9 image is horizontally squeezed and recorded as 4:3. The player detects the 16:9 flag and widens the picture for display.

The A/D path would digitize the incoming analog video to Y, Cb, Cr. This would then be sent to the DV format encoder as an alternate to the CCD.

[kamaleon]

Great stuff mate, thanks

In 16:9 mode, the 16:9 image is horizontally squeezed and recorded as 4:3. The player detects the 16:9 flag and widens the picture for display.

Is this the reason why HD 16:9 camcorders produce 1440x1080i (rather than 1920x1080i)?

The A/D path would digitize the incoming analog video to Y, Cb, Cr. This would then be sent to the DV format encoder as an alternate to the CCD.

Suppose we can admit the same happens in mpeg2 or avchd camcorders right?
So i take from your words that the DV format encoder is not the same chip as the A/D one then.(?)
I repeat my previous question, the D/A converters, are they the of the same kind (if not the same) one can find in capture cards?

[JohnnyMalaria]

This little graphic shows a highly simplified view for DV 16:9 mode.

That graphic is incorrect. The CCD section shows that the full area of the chip is used for 4:3 and that 16:9 is obtained by simply ignoring the upper and lower regions. Such a camcorder does not and cannot record true 16:9 and will never generate a 16:9 flag in the DV signal (assuming that the width of the CCD is equivalent to 720 pixels). (I suppose I should never say never...)

Camcorders that do true 16:9 do so by using CCDs that are wider. These DO "squeeze" the image and the examples in the graphic apply.

Two of my camcorders have native 16:9 capability. One of them uses the extra CCD area usually used for the electronic image stabiliser. In 16:9 mode, this is disabled. My other one has an optical image stabiliser and so isn't affected.

[JohnnyMalaria]

Ok, i'll get back to this later on as i need some time to digest all this info

Anyway, regarding my original question... the analog-to-digital converter used on camcorders, is it the same processor that processes the image from the lens (the chip that can actually "film" or is it just used to convert the analog inputs if there are any?

Yes and no.

When people refer to analog-to-digital conversion with a DV camcorder, they are really refering to analog-to-digital conversion followed by encoding of the digital data into the DV format.

Any camcorder that has the passthrough function will use an analog-to-digital converter for the incoming analog video signal. When using the camera section, there are separate analog-to-digital converters to convert the raw image sensor information.

External video signals are in "Y/C" - basically a black-and-white signal and a separate (for S-video) colour signal. The Y part is converted to a digital signal. The C part is converted to two digital signals called Cr and Cb.

The raw signals from the image sensors are in Red/Green/Blue format. These are converted to digital and then mathematically converted to Y, Cr and Cb.

The DV encoder electronics will accept digital Y, Cr and Cb data and encode them into the DV format.

Typically, the same chip that does the DV encoding will also decode DV data into Y, Cr and Cb. The Y and C are calculated and an analog voltage created that is proportional to the digital values. Finally, the analog signals are further processed to generate analog signals that TVs can understand (addition of the timing signals etc).

[kamaleon]

I'm glad i've joined this board

[edDV]

This little graphic shows a highly simplified view for DV 16:9 mode.

That graphic is incorrect. The CCD section shows that the full area of the chip is used for 4:3 and that 16:9 is obtained by simply ignoring the upper and lower regions. Such a camcorder does not and cannot record true 16:9 and will never generate a 16:9 flag in the DV signal (assuming that the width of the CCD is equivalent to 720 pixels). (I suppose I should never say never...)

Camcorders that do true 16:9 do so by using CCDs that are wider. These DO "squeeze" the image and the examples in the graphic apply.

Two of my camcorders have native 16:9 capability. One of them uses the extra CCD area usually used for the electronic image stabiliser. In 16:9 mode, this is disabled. My other one has an optical image stabiliser and so isn't affected.

Agreed I didn't want to go into too much detail on 16:9 when the main point was the need to process the CCD RGB into YCbCr and then encode to DV format.

The full graphic is here. Later generations use more pixels on the CCD and down filter for both 4:3 and 16:9, but single CCD models still need to divide the CCD into R,G and B so overall resolution is lower than 3xCCD.



Better camcorders use "optical" stabilization to avoid the CCD issues you describe.

[JohnnyMalaria]

Better camcorders use "optical" stabilization to avoid the CCD issues you describe.

Funnily enough, one of my camcorders is the veritable PDX10

I've even got an electronic version of the service manual for it. If you ever need to convince yourself of the price tag some of these machines carry, the service manual will do it! The amount of electronics stuffed inside something that small is nothing short of astonishing.

[kamaleon]

Indeed

[kamaleon]

Back again folks, been doing some research about my wintv-usb2. I can't seem to be able to tell which chip it uses. apparently it's got both a Texas Instruments TVP5150A and an eMPIA Tech EM2880.

Would you know which chip is the one that's responsable for the A/D conversion? (which would be the relatives to the famous BT8X8 or SAAXXX)? i'm trying to determine exactly how it works...

Thanks!

[edDV]

Back again folks, been doing some research about my wintv-usb2. I can't seem to be able to tell which chip it uses. apparently it's got both a Texas Instruments TVP5150A and an eMPIA Tech EM2880.

Would you know which chip is the one that's responsable for the A/D conversion? (which would be the relatives to the famous BT8X8 or SAAXXX)? i'm trying to determine exactly how it works...

Thanks!

Well the TVP5150A is the composite/S-Video input, comb filter and NTSC/PAL decoder.
http://focus.ti.com/docs/prod/folders/print/tvp5150a.html

The EM2880 is the video streamer over USB2 block.
http://www.empiatech.com.tw/pro_em2880.htm

According to wiki, later models of the PVR USB2 used
* Video Decoder: Conexant CX2584x
* MPEG Encoder: Conexant CX23416
http://en.wikipedia.org/wiki/Hauppauge_Computer_Works

[kamaleon]

Great reply edv, cheers for that one

So the empia is the relative to the bt878, right? i'm sorry, but i'm not familiar with the right terms exactly such as video streamer over usb2 block

[SingSing]

Analog capture is a mixed signal operation. The quality of the signal is depends on the analog amplifier and filter, and signal conditioning. The Chip set is responsible for resolution, synchronization, and algorithm. If you are concerning about visual quality, knowning the chipset alone does not help.

Computer graphics is also a mixed signal operation. That is the reason, that with the same video chip set, a better video card produces crisp, and vivid picture, with better amplifier, and signal conditioning, and cost more.

[kamaleon]

Analog capture is a mixed signal operation. The quality of the signal is depends on the analog amplifier and filter, and signal conditioning. The Chip set is responsible for resolution, synchronization, and algorithm. If you are concerning about visual quality, knowning the chipset alone does not help.

Thanks SingSing, i am concerned on compatibility issues with software too. The only reference i have is "conexant" or "philips SAAXXX" etc that's why i'm trying to know which chip mine has... empia or texas instruments...

[jagabo]

The TI and Connexant chips are video digitizers:

http://focus.ti.com/docs/prod/folders/print/tvp5150a.html

http://www.conexant.com/products/entry.jsp?id=120

The Empia EM2880 is a digital video processing and USB chip:

http://www.empiatech.com.tw/pro_em2880.htm

None of this has much to do with compatability as far as capture software is concerned. The capture software doesn't see the video digiting chip. It only sees the MPEG data coming across the USB cable. The same is true for DV and firewire.

[kamaleon]

i use some french software like k!tv and meuhmeuhtv that say they'll be compatible to this or that chip, like SAAXXX or BT8X8... I'm sorry but you guys seem to really know your mustard but i find it hard to keep up just trying to find something i can relate to.

Anyway, sorry about my ignorance but the difference between a "video digitizer" and a "digital video processing" is something i still have to learn.

[jagabo]

The video digitizer converts analog video into digital data. The digital video processing chip manipulates that digital data. In the case of the Empia chip it can adjust the color, brightness, contrast, etc, then compresses to MPEG, and finally streams the MPEG data to the USB port.

Software designed to work with SAAXXX or BT8X8 chips cannot access the video digitizing chip with an MPEG encoding device like the PVR USB2, even if it is the same chip.

[kamaleon]

The video digitizer converts analog video into digital data. The digital video processing chip manipulates that digital data. In the case of the Empia chip it can adjust the color, brightness, contrast, etc, then compresses to MPEG, and finally streams the MPEG data to the USB port.

Ah! now i understand much better, thanks

Software designed to work with SAAXXX or BT8X8 chips cannot access the video digitizing chip with an MPEG encoding device like the PVR USB2, even if it is the same chip.

Yes, i realize this, however i'm not personnally concerned by it. I do not use an mpeg encoding device. These were just examples.

[grantaccess]

Hi all. I want to thank you all for the info you're sharing - I learned so much I have a headache. In fact I registered because of this thread.

Absolutely.

http://www.datavideo.us/products/dac_100_main_page.htm

http://www.canopus.com/products/ADVC110/index.php

But it might be cheaper to keep your existing camcorder to use for conversion if you get a new one. (I assume you are planning to sell your existing one?)

I cannot speak to the ADVC110 but I own the Datavideo DAC-100. The DAC-100 gives very nice results when your analog source is good. If however your source is a magnetic tape that is stretched or degraded you will not be happy with the results (unless, as people have pointed out, you employ a Time Based Corrector). When the DAC-100 cannot properly detect a sync pulse it will stop sending any data stream to the computer. This causes the captured video to be seamless but incomplete. This makes it very difficult to quickly look through the captured video for problems. Another issue is that any audio data, that could have been captured while the video sync pulse was problematic, is lost. I have not tried either product but I know that both Canopus and Datavideo offer converters that have integrated TBCs (the next step up in both product lines.)

The Sony DV Passthrough from my camcorder gives me far the best results because it has a stabilizing effect on the the incoming analog (VHS) signal.
(It seems to use some form of simple TBC for the analog input)

Many Sony Camcorders do in fact incorporate a TBC which I believe is operational in passthrough mode. I have performed captures on source material with sync problems on two models of Sony camcorder in pass-through and with the DAC-100 mentioned above. I do not have the model numbers handy but I can get back to you with them if you like. The older Sony performed very similarily to the Datavideo DAC-100. Both had problems reading the sync pulse and both halted the data stream while the sync pulse couldn't be Detected. The newer Sony performed light years ahead of the other two devices. The analog source material had to be very bad before it was unable to convert. I will have to check whether the stream was lost or not (we set aside the more problematic source material and did not get a chance to return to it.) While checking into why one model performed so much better I found in the specs that some Sony camcorders have Time Based Corrector listed in the specs. I looked up the manual for The_Doman's Sony (http://www.henrys.com/manuals/sony/DCR-TRV120-320.pdf) and found that a TBC is mentioned in the playback options for correcting jitter. I assumed that the TBC was employed to make some consumer function of the camera possible (and marketable) and that it's functionality in the pass-through convertion was just a bonus for those of us who convert older source material.

One very good function of the DAC-100 is the fact that it streams both analog and DV simultaneously. Depending upon which program you are using to capture the video you will likely not be able to monitor the audio in realtime(post digitization.) I ended up using the DAC-100 to send video to a Television for monitoring the transfer after it was digitized by the Camcorder. I would connect the analog source to the newer Sony Camcorder using the analog inputs. Connect the Sony to the DAC-100 via firewire. Then connect the DAC-100 to the computer via firewire, and to a television via it's analog outputs. This way the person digitizing can monitor the digitized audio and video on a television which is much less fatiguing than on the computer or on the camcorder's little monitor.

hope some of that helps (or at least makes sense)