Correct. OBS taps into the OS preview, or the graphics output screen -- not the direct driver for the card. It was designed as a streaming screen recording software, not capturing.

Screen (or preview) recording can have problems.

You could actually already do this with VirtualDub, and even with Windows XP, using Microsoft drivers. VirtualDub natively has the feature as well, screen recording, but it's rarely useful or mentioned. It never worked well, aside from a quasi AIW setup, since the AIW was actually piped through the graphics as well, so it worked fine.

Why then does it show the attached capture cards to choose from for video capture? Given that every video editor under the sun can capture from the cards, why not OBS, when it clearly also has that feature/selection?

I've not looked at OBS since last year, did something change? It does appear so. However, is it actually taking the card output, connecting to the driver, or just interfacing with the output preview?

And it now allows lossless recording?

It has a capture mode similar to other capture utilities and supports the decklink card, I havent had any issues with it.

I went through the vdub guide posted on this site in a fair amount of detail, but could not find anything about turning off monitoring. Could you please clarify how this is done?

The guide is here: http://www.digitalfaq.com/forum/video-capture/7427-capturing-virtualdub-settings.html

Yes it has a set of red/white RCA monitor outputs.

Hmmm I wouldn't know how to determine this technically, all I can say is that it seems so.

Yes, in Settings / Output / Recording / Recording Quality when "Lossless Quality" is selected the capture is written to a AVI file - see attached screenshot.

OBS does capture from the same stuff as virtualdub can, though as it's designed around mixing stuff on a video surface rather than outputting the data from the card directly, you have to be very careful with the video settings to avoid it messing with the video.

Those settings are not correct. Is it PAL or NTSC?
PAL should be 720x576 and 25fps,
NTSC should be 720x480 and 29.97... fps (NOT 30)
any other frame rate will result in dropped or inserted frames to compensate.
Color space should be ITU-R BT.601, not BT.709 (that is for HD video.).
Decoded format: Should be some variant of 4:2:2 YUV, not 4:2:0 (which will mess with the color since the video is interlaced)
Important that these things are correct both in OBS settings and for the video codec used, it's not guaranteed to match.

And this is why!

An easy tell is always digital-only artifacts, such as screen tearing.

I get rather tired of hearing comments like "it seemed fine to me", but it wasn't picked over in detail, to actually see if there are intermittent errors. Too many people act as if certain errors will be immediately observable, and jump out of the screen and slap them in the face. Errors can be random, and pernicious. You'll only see them when you see them. And it's often long after the capture was done, sometimes years later. (A lot of people capture 100 family videos, and don't even watch 1 of them until all 100 are done. Bad form, sloppy science.) After you notice, especially if it's cyclical, you'll be annoyed by it. This is why so many people are recapturing video. The 2000s was failed first attempt, 2010s/2020s is 2nd try. Happens all the time. Those base-on-nothing "fine" comments mislead newbies far too often, usually a case of newbies misleading newbies (blind leading blind).

Yes, I do admit to be a clueless soul trying to figure this stuff out, however

means that as clueless as I am, I actually viewed the captured footage carefully but saw no signs of anomalies (however I have no other way of checking than visual review).

BTW, could somebody please help with this question?

Go to audio tab, uncheck audio playback.

Sorry to persist on this, but I did already have that unchecked. I have attached a file showing screenshots of all audio related settings.

Any other ideas? (I can post a very short sample of avi is that helps)

vdub may not work properly with a pro capture device, I personally never got vdub to capture properly but MediaExpress works flawlessly, We are talking 270Mb/s in 10bit, 8bit slightly lower, but still a more lean app like MediaExpress or AmarecTV is the way to go, vdub is filled with plugins, filters, options and settings and even with audio monitoring turned off it could choke on 270Mb/s.

Thanks for your reply.

After playing around with vdub and AmarecTV I decided to stick with MediaExpress. The only remaining issue I had was that when I imported the resulting lossless AVI file into Vegas Pro, it would not recognize the audio track. To get around this I came up with the following ffmpeg conversion syntax based on googling/reading of the ffmpeg documentation:

ffmpeg -i input.avi -acodec pcm_s16le -vcodec copy output.avi

This will essentially copy the lossless video without modification and convert the audio to a standard that Vegas Pro will recognize.

I thus end up with a lossless AVI file (PAL) which exactly matches hodgey's specs:

720x576
25fps,
Decoded format: 4:2:2 YUV

So I think I am good to go.

I thank everyone on this forum for their input in the process and their patience with my beginner questions - it has been a steep but rewarding learning curve.

Dennis

Yes, I've raised this few times too when LS and others repeat that it's only screen capturing software. I don't especially like OBS, but it does direct capture from hardware, I've even linked the source code showing this. OBS has been able to do this for a very long time.

Personally, I don't use it, I don't like it for video capture, but it can make a perfectly good job of it if you dial in the right settings. It's certainly not its forté, but it's demonstrably incorrect to just say it's only capable of screen recording.

If individuals like it and get the results they like, then crack on in my opinion.

There are a lot of purity tests set by some individuals in this community. It's quite amusing to watch sometimes.

I'm tryin' OBS right now, seems like it can "tap" into the ouput of my hdmi card. The encoding options are limited though (NVenc or x264), and aac 320 for audio, not quite enough. No deinterlacing filter that i can see aswell or infos on dropped frames. I don't think it's adapted for our purpose

I can't speak as to compatability with your card.

From an audio perspective 320 AAC is complete overkill for any domestic videotape though, don't be put off at that, you can comfortably capture all information at far lower rates than that. 320 is usually reserved for other purposes, AAC 128 will comfortably digitise VHS audio.

128 AAC is too low for this purpose.

Let's go back to AC3 for a moment.
- 192 was the minimum, noticeably compressed (unless your speakers and hearing was lousy).
- 224 was the standard low for high quality source, no noises, mild compression detected in some sources.
- 256 was the standard high for most decent or quality audio, no noticeable compression
- 384 was required for audio with noises, such as hiss, ie most consumer analog videotapes, even with audio NR/restore/processing.

Those are stereo values. 6-channel doesn't equate, due to more channel, but also less actual audio in the channel. But still, baseline for 6-channel was 384, up to 512.

Now comes AAC.

It essentially moved the numbers down by 2 spots.

So, in AAC:
- 96 is the low minimum, noticeably compressed (unless your speakers and hearing is lousy)
- 128 was the high minimum, less noticeable, but still noticeable at times
- 160 was the standard low for high quality, some sources can cause oddities and loss
- 192 is the standard high for high quality -- or low for low quality, with oddities/loss
- 224 or even 256 is usually needed to properly process noisy audio from video tapes, mostly so you can distinguish the audio noise from the audio high.

With AAC, NR/restore is suggested to remove the noise as much as possible, but it still cannot be 100% filters out, and you don't want noise to blend with the highs.

The compression issue with AAC is not the same as AC3. With AC3, there was the traditional "digital sound" to the audio, as values reduced. But AAC is more like H.264. PEG-2 gave you blocks and mosquito noise, while H.264 just blurs and mushes. AAC with a low bitrate starts to mush the audio, and it can gets harder and harder to distinguish sounds. If you have XM radio, the worst MLB off-game audio is a good example of this. (Note that the MLB is really low bitrate stream, so it's far below what you'd get from AAC 128.) So 128 doesn't "sound digital", but it definitely mushes frequencies together.

AAC is far more complex than AC3, because it has multiple profiles, and actually is not constrained to bitrates like 128, but could be 125, 130, 132, etc. But that's another issues as well, in the decoder expectations.

Using Fraunhofer is probably one of the most important aspects.

I could go on, but out of time.

FYI, this isn't my opinion. This is the sort of collected data that I had access to when I worked for studios. But some of this is so commonly known that "the scene" also uses this in widely accepted audio specs by contributing members.

NOTE: You should not be capturing audio directly to AAC.

What audio MediaExpress spits out? Mine outputs PCM 48Khz 24Bit, If this is indeed your output I don't see why Vegas would not recognize it.

On a side note, If you already know how to use script commends why don't you just use ffmpeg and QTGMC to do pretty much anything needs to be done post capture? NLE is a pain in the a$$.

Mine spits out the same audio = PCM 48Khz 24Bit, however as mentioned when added to the Vegas Pro timeline, only the video track shows up. After converting the audio to something like PCM 48Khz 16Bit, the audio track also appears - not sure why, maybe a missing codec on my PC?.

Agree that it can be a bit painful, my use is very limited, and if not needed I will just use ffmpeg. Here are a couple of situations where it comes in handy:

1) Really bad video footage with noisy transitions and dropped frames (where the capture stops). This ends up creating a number of AVI files. I put these on the timeline, trim off the noise and render as a single mp4.

2) Precious but real crappy footage recorded on cameras which are faulty, e.g. missing a color, I can sometimes improve on these using some of the FX filters, Neat video etc. (although the basic tenet of my coversion is to intervene as little as possible with effects/editing /filtering etc.)

-- merged --

In the interest of "vindicating" OBS's reputation as a "screen recorder only", I spent some time time trying to replicate the lossless capture achieved using the Blackmagic Media Express software (v. 3.8) using OBS Studio (v.27.2.4).

OBS has an advanced output option which allows for specifying custom ffmpeg style arguments which provide much more control over the capture parameters than the standard options. I believe the results came quite close - I have documented the comparison results in the attached PDF in detail.

I am not saying it is recommended to use OBS over Media Express - it is quite true as pointed out by several forum members that OBS has many more options/moving parts and it would be quite easy to mess up the transfer if any of the key option selections are incorrect. Media Express on the other hand has very limited options, and thus less room for error.

But I thought it would be an interesting exercise nonetheless.

A synchronized side-by-side full viewing of both clips using Kinovea did not reveal any obvious visual differences, but then again it is early days for me and I might have missed something a trained eye would detect.

I would be interested in hearing opinions on whether the resulting video characteristics are comparable, and/or any thoughts on how they might differ significantly - in short would OBS configured in this manner be an acceptable replacement?

When you capture multiple segments with MediaExpress the easiest, quickest and most lossless way to trim them or merge them is vdub2, You don't even have to select the files one by one, MExp names files in incremental numbers, So you only need to load the first file to vdub2 and then go to file/append video segment and vdub load all the files in the same file name matrix. They will all be loaded in the timeline as one large file and from there do your editing and output the file without transcoding, You can also output to HuffYUV for almost third the size with lossless compression. You just have to make a folder for each tape or capture to avoid automatically loading files from previous captures.