Real vs market ISO - page 3 - General Discussion and Assistance - CHDK Forum supplierdeeply

Real vs market ISO

  • 110 Replies
  • 22703 Views
*

Offline blackhole

  • *****
  • 844
  • A590IS 101b
    • Planetary astrophotography
Re: Real vs market ISO
« Reply #20 on: 07 / July / 2013, 12:12:33 »
Advertisements
a530 propset1 and a590 propset2

Code: [Select]
platform,build,build time
a530 100a,1.2.0-2804-0,May 27 2013 15:24:05
iso,sv96,sv96_m
80,370,449
100,411,480
200,507,576
400,603,672
800,699,768

platform,build,build time
a590 101b,1.2.0-2940-0,Jul  7 2013 10:06:40
iso,sv96,sv96_m
80,371,449
100,411,480
200,507,576
400,603,672
800,699,768
1600,795,864

Re: Real vs market ISO
« Reply #21 on: 07 / July / 2013, 12:13:15 »
Details in the attached spreadsheet.
All those values look really good - especially the iso_m/iso_r.  As reyap pointed out,  ISO80 is usually a special case and with that in mind, a simple multiplication to convert iso_m to iso_r in the UI is quite possible.
Ported :   A1200    SD940   G10    Powershot N    G16

*

Online reyalp

  • ******
  • 13433
Re: Real vs market ISO
« Reply #22 on: 07 / July / 2013, 14:30:52 »
If people can post the original CSVs from the script as an attachment (zip if there's more than two), it will be easier for me to work with them all in a consistent way.

edit: I'm not asking for people to repost what has already been posted, just for any additional posts.

A few things to keep in mind:
1) If the camera does not have manual ISO in P mode, the script won't produce usable values.
2) The script doesn't know the range of valid ISOs for a given camera. You should input these in the script parameters. Values beyond the valid range may produce odd results.
3) It's possible additional delay is required after get_shooting goes true, shooting_set_sv96 has this lovely thing:
Code: [Select]
            while ((shooting_is_flash_ready()!=1) || (focus_busy));

You can test whether the script is reading values correctly by
1) enable propcase display, with the page that shows SV_MARKET
2) select a mode in the canon firmware that lets you use manual ISO
3) select an ISO mode, press and hold shoot_half until SV_MARKET updates
4) compare SV_MARKET value with what the script put in the log.
5) for a sanity check =3.125*(2(SV_MARKET/96)) should equal (or be very close to) the selected ISO

Thanks everyone for testing, I'll post again after I've looked through the data so far.
« Last Edit: 07 / July / 2013, 14:43:17 by reyalp »
Don't forget what the H stands for.

Re: Real vs market ISO
« Reply #23 on: 07 / July / 2013, 14:45:46 »
I'll have to try a different approach with the G10 and SD940.  Here's the originals for the A1200 and SX50.   The A1200 does not have a menu entry for 3200 but seems to go there okay - the SX50 does have a Canon entry for 6400.
Ported :   A1200    SD940   G10    Powershot N    G16


*

Online reyalp

  • ******
  • 13433
Re: Real vs market ISO
« Reply #24 on: 07 / July / 2013, 15:15:35 »
The A1200 does not have a menu entry for 3200 but seems to go there okay - the SX50 does have a Canon entry for 6400.
Yeah, on D10, setting ISO_MODE to 3200 appears to produce the same values as using the special "ISO 3200" mode.

Another note is that you can set ISO_MODE to values that don't appear in the menu, e.g. ISO_MODE 90 produces values APEX96 values between 80 and 100. I'm going to make a variant of the script to see what happens in the transition between the ~1.7 factor at ISO 80 and ~1.6 that is present for all the ones >= 100

It looks like the g10 and sd940 aren't updating the ISO at all. I've attached a new version of the script that adds a 1 second delay after halfshoot goes true. If flash is enabled while running the script, you might want to turn it off.

I've also attached a new spreadsheet with all the data so far. I re-arranged the columns so you can just paste in the log output and extend the formulas in the remaining columns.
Don't forget what the H stands for.

*

Online reyalp

  • ******
  • 13433
Re: Real vs market ISO
« Reply #25 on: 07 / July / 2013, 15:25:50 »
On the results so far:

It looks like the very old cameras (through a430) have a different pattern, where "real" ISO is slightly higher than "market", but the difference is quite small (e.g. 115 vs 100), and don't treat the lowest values specially.

The remaining cameras (ixus 65 through sx50) appear to be very consistent market = 1.76*real for ISO 80, and 1.65*real for the remaining values.

If these are really the only two groups, this is pretty promising, because the number of early cameras is quite small.
Don't forget what the H stands for.

*

Offline dvip

  • ****
  • 451
Re: Real vs market ISO
« Reply #26 on: 07 / July / 2013, 15:44:27 »
Here are the A590IS (100b) and the SX40HS (100i).
A590IS, minimum iso: 80
SX40HS, minimum iso: 100

Both the SX40HS and the SX50HS have the same sensor
but the SX50HS uses iso 80 as a minimum.

*

Online reyalp

  • ******
  • 13433
Re: Real vs market ISO
« Reply #27 on: 07 / July / 2013, 15:49:04 »
One more script update. This version adds and option to test the sub 100 values in 1 unit increments. I don't think we need to do this for every camera.

On d10 and a540 (updated spreadsheet attached) it looks like it transitions to 1.6-ish immediately after ISO 80, and doesn't actually produce 1 ISO increments in the APEX96 values.
« Last Edit: 07 / July / 2013, 16:12:55 by reyalp »
Don't forget what the H stands for.


*

Offline dvip

  • ****
  • 451
Re: Real vs market ISO
« Reply #28 on: 07 / July / 2013, 16:11:05 »
@reyalp, I get the same values for both cameras (A590IS/SX40HS) with the last isotest.lua script.

*

Online reyalp

  • ******
  • 13433
Re: Real vs market ISO
« Reply #29 on: 07 / July / 2013, 16:26:26 »
@reyalp, I get the same values for both cameras (A590IS/SX40HS) with the last isotest.lua script.
I'm not sure what you mean by "same" ? The only difference should be that it produces more lines in the CSV if you check the "test MIN-100" option, and the starting ISO is < 100.

Anyway, as I said I don't think it's necessary to test the 80-100 range every time, I would be very surprised if the the modern cameras behaved differently. It looks like only 80 is special, so cameras that start at 100 can use a simple conversion factor.

Regarding the SX40, it's interesting that you get an SV96 of 401 for ISO MODE 80, which would correspond to an "market" ISO of ~92 using the normal conversion factor. The "market" value is the same as for ISO 80 on other cameras, suggesting that the canon firmware just does a blind conversion with checking the limits.
Don't forget what the H stands for.

 

Related Topics