bikinkawboy Loc: north central Missouri

Maybe some of you folks can answer this: as digital photography has progressed over the years, high iso image quality has improved dramatically. Has it been a result of improved sensor technology, improved software or some of each? I’m just curious. Thanks!

bsprague Loc: Lacey, WA, USA

With my gear and software, it is both.

pmorin Loc: Huntington Beach, Palm Springs

It has been my observation that all electronics need programming of some sort to work properly. You need improvement in each aspect of the process.

dustie Loc: Nose to the grindstone

Some of each.

Orphoto Loc: Oregon

If you attribute everything that happens in the camera to the hardware column, then i would say 75% hardware, 25%software.

Longshadow Loc: Audubon, PA, United States

Both.
New hardware requires new supporting software.

CHG_CANON Loc: the Windy City

Both, although the absolute superior results begin with superior exposure technique performed by the photographer (human) using their camera, then processing the results.

This link looks at some older images, now almost 4 years ago, for ISO-5000 and links to other groups of images at ISO-4000 and ISO-3200. Those were the limits of the cameras involved at that time. New models being sold now are offer even high ISO possibilities.

Shooting at high ISOs (ISO-5000)

Architect1776 Loc: In my mind

Both, from consumer observation.
I could not give tech details though.

burkphoto Loc: High Point, NC

YES!

EVERYTHING has improved since the first promising digital sensors were created in the labs about 40 years ago.

The whole industry has grown up. It has seen exponential changes in computing power, sensor density, sensitivity, and performance. The firmware and software have all been refined iteratively. So we all benefit from this collective, continuous improvement.

ELNikkor

One reason an upgrade every 10-12 years might be beneficial, even if the old equipment still works.

burkphoto Loc: High Point, NC

In the computer industry, five to seven years is probably more appropriate. Operating systems and software advance quickly and require more advanced hardware to run efficiently.

My current camera still works fine, but it is a ten year old design I bought eight years ago. I'll be upgrading soon. The recent advances have been too substantial to ignore.

Architect1776 Loc: In my mind

That's my schedule.
Each new model is just a slight upgrade usually. Some makers could just do a firmware update but make a new model with and older sensor but new firmware
So after 10-12 years you actually see a difference.

R.G. Loc: Scotland

If a camera comes out that doesn't have a brand new sensor but does have a new processor, it will be a significant upgrade because of the superior in-camera processing that the newer processor will allow. OTOH sensor design is constantly being refined and at the present time a new sensor also usually means a significant upgrade. That will be especially true when stacked photodiode sensors start to become available in new cameras. Global shutters increase a camera's usefulness but stacked photodiodes are a significant improvement as far as image quality is concerned.

larryepage Loc: North Texas area

I worked for Texas Instruments for a number of years in the last century. My group built several families of TTL logic building block circuits. There were hard limits on how fast those circuits could "switch" from one state to another, which put hard limits on how fast systems built with them could operate. Each new family we introduced over the years introduced changes that allowed for faster operation of the individual circuits and the syatems they were used to build. Those changes were 100% hardware.

During that time, I had a friend who worked for Western Electric. He was a really cool guy from Switzerland who was full of all kinds of stories. But one of the coolest things about him was that he had a patent related to sending data over a twisted pair telephone line. One of the assertions (with technical support) in that patent was that such data could never be sent any faster than 1200 baud (events per second) via twisted pair. You can imagine his despair when Hayes introduced their 2400 baud SmartModem which operated over a standard telephone line. That was done mostly with software.

Just so you know, USB communication, which is faster than several million baud, is also via twisted pair. That's accomplished with a combination of hardware and software. Some of each of which is really, really close to magic.

So sometimes it is hardware, sometimes software. Usually the requirement is for innovative hardware which can serve as a foundation for innovative software.

Dikdik Loc: Winnipeg, Canada

both, but mostly the improvement in hardware.