What is black stretch?

[Kevin Red]

I just read a little something about black stretch here but couldnt find anything else about it. Is it on the HC1?

[Greg Boston]

Don't know Kevin if it's on the camera you mentioned, but black stretch is pretty much just that. It makes the roll-off at the bottom of the gamma curve more gradual which will reduce contrast a little bit, but will maintain detail in darker areas. The master pedestal determines where that roll off starts at the bottom.

-gb-

[Alexander Karol]

I don't think the HC1 has black strech. I think it's only on its pro brother, the A1U. People actually bash the HC1 a little here and there for its low light performance.

[Paul Rickford]

Correct-There is no black stretch on the HC1 and the low light is not good enough (in my opinion) for a HD cam, I know it's only quoted at 7lux but it goes way to grainy to quickly with the auto gain and ends up looking like standard dv -not what you want in a HDV production! A urban myth has built up about how powerful CMOS chips can handel low light etc., The poor HC1 just can't live up to it and I suspect the Build up out there now that the A1 being so much better is going to be the next myth!
Paul

[Laurence Kingston]

Yeah, it's just the A1 that has the black stretch feature.

As far as the "urban myth" of the powerful CMOS chip, my understanding is that it actually is not totally myth. Here's what is going on: the light sensors, be they CCD or CMOS can pick up a certain dynamic range. The DV or MPEG encoders can handle a smaller dynamic range as they encode the image on to tape. When the light levels go down, the dynamic range of image goes down.

Both the C1 and A1 have the same CMOS senosor, but in the case of the A1, the lower dynamic range of the picture is being expanded and the blacks are stretched down to black before it gets encoded on to the tape. The CMOS is more sensitive than standard CCDs and has the in between values available to stretch before encoding and this is what makes the difference in low light performance.

I have an A1 on order which should be here on Thursday. What I'm expecting (compared to my "see in the dark VX200") is for low light performance to be quite good in moderate levels, but fall of sharply once the light falls beyond a certain level. In other words, in an average indoor room, it should be close to VX2000 performance and well above H1 performance, but if the light fell further, the low light performance would drop sharply and be nowhere near as good as my VX2000.

Because of this I will probably keep my VX2000 for low light situations.

[Paul Rickford]

Laurence- since my last post I have got hold of a A1 and have done a few tests between the cams, sadly in my personal opinion (or happy if you already own the HC1!) I could for the most part not tell them apart for image quality. The black stretch does open up the shadows a tad, but works best in a well exposed contrasty scene. In low light with the gain kicked in you just get grey instead of black grain. I think the CMOS and EIP is already working to the max on the HC1 and all the A1 is doing is giving you control within the same paramiters. Its still a great cam and worth the extra for XLR and Sony 2 year silver cover, but i'm betting you will want to keep the VX.

Cheers

[Laurence Kingston]

I get mine on Thursday. Your post makes a lot of sense. I just love my VX2000 in low light and I'm sure I will be keeping it. I'm still going to be happy with the A1 I'm sure. On my VX, I have always been unhappy with the audio situation. I have a Rode Videomic, a Studio 1 beltpack XLR converter and an ME66 on a boom. It sounds great, but one thing I can't do is use a boom mic with a camera mic for backup. I also can't conveniently use a shoe mount light when I'm using the Rode Videomic.

Another thing that nobody mentions is the time code features. A friend of mine just got back from an audio gig on a low budget movie. They had rented a portable hard disc recorder with timecode and it worked extremely well with the camera. With a lanc connection between the camera and the audio recorder, the audio box went into record every time the camera did, and at editing, all the footage and audio is expected to just line up by itself in the editor. Too cool if you ask me. Anyway, the A1, PD150, PD170, Z1 and PDX10 all do this, but none of the consumer versions of these cameras do. At some point, a relatively cheap flash ram recorder with timecode is sure to become available and when it does, timecode will be a big deal to all of us I'm sure!

How bad is the A1/HC1 in low light?

[Laurence Kingston]

Well I got the HVR-A1 yesterday. What a great camera! Yeah the zoom is too fast and the black stretch feature is a minimal improvement in low light, but the picture is just incredible.

[Michael Liebergot]

What are your oppinions on the manual (Exposure, white balance etc.) controls on the A1.
Do you think that they are good enough to use in the field for run and gun shooting, like event videography?
Also does the A1 have a push auto button, for quick focusing like my VX2100 does. Speaking of Auto Focus, how would you rate the A1's?

Thanks,
Michael

[Alexander Karol]

Guys, this thread has completely deviated from its original topic. Please start new threads for such inquiries.

[Douglas Spotted Eagle]

Originally Posted by Alexander Karol

Guys, this thread has completely deviated from its original topic. Please start new threads for such inquiries.

it hasn't deviated from topic yet, or I'd split it off to start a new thread with it. Just be sure we keep this on topic, we'll be fine

[Dave Haynie]

Originally Posted by Paul Rickford

A urban myth has built up about how powerful CMOS chips can handel low light etc., The poor HC1 just can't live up to it and I suspect the Build up out there now that the A1 being so much better is going to be the next myth!
Paul

It's not a myth... more of a goal. Any imaging chip has the same basic issues: there's a sensor site of some kind (CMOS, CCD, something else) that receives a certain amout of light. That light is based on the lens aperature, the pixel size (which is turn derives from the sensor size vs. resolution). When you're done, that's the light you get, period.

Next is how you deal with that light. At some point, the analog value recorded at that sensor site has to turn into a quantized digital value. So it's fed through a variable gain amplifier and on into an ADC (analog to digital converter). The setting of that amp is based on either you autoexposure logic or the bias you dialed in.

The noise issue is the same issue you always have in systems: where's your noise floor versus your signal? If your noise floor intrudes on the signal, you see noise (which looks curiously like grain from the film world). Camcorders with good low light performance have typically done it by using larger CCDs: thus, more light per pixel, thus, a lower bias for the same effective EV level.

Due to their integration and parallel nature (versus the serial CCD), CMOS sensors can potentially address this in another way: lower system noise. There's good theory behind this, but it's been a learning process -- the first CMOS sensors were cheap but noisier than CCDs. Today, the recent run of modern 35mm Digital SLRs is seeing this realized: clean images from CMOS sensors up to ISO 800 or even ISO 1600, much better than film in the same situation.

These Sonys are the first serious CMOS camcorders. I think their primary goal was "at least as good as CCD"... we'll have to wait and see if they can do that better. From all accounts (don't have my A1 yet, but it's in the works) this is the technology to back, over time.