Bug #70603 for Image-ExifTool: Negative gain settings on Sony HXR-NX5U reported incorrectly

Tue Aug 30 11:29:41 2011 mistydemeo [...] gmail.com - Ticket created

Subject:	Negative gain settings on Sony HXR-NX5U reported incorrectly
Date:	Tue, 30 Aug 2011 10:29:33 -0500
To:	bug-Image-ExifTool [...] rt.cpan.org
From:	Misty De Meo <mistydemeo [...] gmail.com>

The Sony HXR-NX5U has negative gain settings which, unusually, appear to be real settings and not postprocessing effects. ExifTool incorrectly reports gain settings below 0 dB; for example, -6 dB will be reported as 42 dB. Judging from the manual, it's possible that Exif is being incorrectly recorded, or not recorded at all, for negative gain images. From the manual, p. 30: "When you record high definition image quality (HD) movies with the gain set to [-6dB] and play them back with the data code, the gain value is displayed as [---]." Tested using Image-ExifTool-8.63 on Perl 5.10.0 as shipped with Mac OS X 10.6.8.

Tue Aug 30 11:56:53 2011 EXIFTOOL [...] cpan.org - Correspondence added

Thanks for this report. Would it be possible for you to send me a full set of videos, one for each gain setting? Please make them very short (1 sec or less). If you can do this, I am sure I can sort this out. My email is phil at owl.phy.queensu.ca Thanks. - Phil

Tue Aug 30 11:56:54 2011 The RT System itself - Status changed from 'new' to 'open'

Tue Aug 30 11:56:54 2011 EXIFTOOL [...] cpan.org - Given to EXIFTOOL

Tue Aug 30 11:59:17 2011 EXIFTOOL [...] cpan.org - Correspondence added

In my previous response I was assuming that there are a maximum of 16 gain settings (since there are only 4 bits allocated to this information in the H264 stream). If there are more than this, then don't send a full set of videos, just a selection across the possible range of values should be sufficient. - Phil

Tue Aug 30 12:15:18 2011 mistydemeo [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #70603] Negative gain settings on Sony HXR-NX5U reported incorrectly
Date:	Tue, 30 Aug 2011 11:15:07 -0500
To:	bug-Image-ExifTool [...] rt.cpan.org
From:	Misty De Meo <mistydemeo [...] gmail.com>

Sure, I should be able to provide you samples. I don't have access to the camera right now, but I'll record samples and send them to your e-mail when I get the chance. Thanks! - Misty

Thu Sep 15 10:43:06 2011 EXIFTOOL [...] cpan.org - Correspondence added

On Tue Aug 30 12:15:18 2011, mistydemeo@gmail.com wrote: Show quoted text

> Sure, I should be able to provide you samples. I don't have access to > the camera right now, but I'll record samples and send them to your > e-mail when I get the chance. Thanks!

I'm in no hurry, but I don't want you to forget about this. - Phil

Thu Sep 22 11:38:22 2011 EXIFTOOL [...] cpan.org - Correspondence added 30 min

Hi Misty, Thanks for the samples. It seems that all values seem to be converted correctly except for the -6 dB value, which is certainly an anomaly. I checked, and only the 4 bits I was decoding in the MDPM information change between your samples. Here is the breakdown: 00000.MTS 0x0f = 15 (minus 1 times 3 = 42 dB) 00001.MTS 0x00 = 0 (minus 1 times 3 = -3 dB) 00002.MTS 0x01 = 1 (minus 1 times 3 = 0 dB) 00003.MTS 0x02 = 2 (minus 1 times 3 = 3 dB) 00004.MTS 0x03 = 3 (minus 1 times 3 = 6 dB) 00005.MTS 0x04 = 4 (minus 1 times 3 = 9 dB) 00006.MTS 0x05 = 5 (minus 1 times 3 = 12 dB) 00007.MTS 0x06 = 6 (minus 1 times 3 = 15 dB) 00008.MTS 0x07 = 7 (minus 1 times 3 = 18 dB) 00009.MTS 0x08 = 8 (minus 1 times 3 = 21 dB) The value of 0x0f is anomalous. If I interpret the 4 bits as a signed integer, then it would indeed represent -6 dB, but with this decoding 0x08 would represent -21 dB (which isn't the case). Too bad we don't know what the values 0x09 through 0x0e are used for. It could be either 0x0f is either -6 dB or just used for any gain which can't be represented by the 4-bit encoding. For your camera it would of course be safe to call it -6 dB, but I worry about other cameras. Unfortunately as far as I can tell there is no free documentation for this. For now, I will just decode 0x0f as -6 dB, but it is possible that the decoding could change to indicate something like "Outside the range -3 to 39 dB" in a future update. Please let me know if this doesn't make sense to you, or if you have a better suggestion. - Phil

Thu Sep 22 11:47:46 2011 mistydemeo [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #70603] Negative gain settings on Sony HXR-NX5U reported incorrectly
Date:	Thu, 22 Sep 2011 10:47:34 -0500
To:	bug-Image-ExifTool [...] rt.cpan.org
From:	Misty De Meo <mistydemeo [...] gmail.com>

Thanks for checking so quickly, Phil. Hm, that -6 dB value is very curious. I'll get a recording at the "hyper gain" setting to see how that compares; it'll be interesting to see if it uses the same 0x0f, or if it has a separate value. Misty

Thu Sep 22 12:02:02 2011 mistydemeo [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #70603] Negative gain settings on Sony HXR-NX5U reported incorrectly
Date:	Thu, 22 Sep 2011 11:01:50 -0500
To:	bug-Image-ExifTool [...] rt.cpan.org
From:	Misty De Meo <mistydemeo [...] gmail.com>

Luckily, the camera was free so I was able to get a hyper gain sample. I've e-mailed you the file, but the Exiftool output does read "42dB". It looks like 0x0f is a generic "not within the standard/valid range" value. Misty

Thu Sep 22 12:10:22 2011 EXIFTOOL [...] cpan.org - Correspondence added

I got the sample, thanks. Yes, unfortunately the "hyper gain" setting records exactly the same information as when the gain is set to -6 dB. So I'm thinking of decoding this value as "Out of range" if that sounds OK with you. - Phil

Thu Sep 22 12:13:20 2011 mistydemeo [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #70603] Negative gain settings on Sony HXR-NX5U reported incorrectly
Date:	Thu, 22 Sep 2011 11:13:10 -0500
To:	bug-Image-ExifTool [...] rt.cpan.org
From:	Misty De Meo <mistydemeo [...] gmail.com>

That sounds good to me. Thanks! Misty

Thu Sep 22 12:30:40 2011 EXIFTOOL [...] cpan.org - Correspondence added

Done and done. This patch will appear in ExifTool 8.65. Thanks for all your help. - Phil

Sat Sep 24 11:32:41 2011 EXIFTOOL [...] cpan.org - Status changed from 'open' to 'resolved'

Sat Sep 24 11:32:41 2011 EXIFTOOL [...] cpan.org - Fixed in 8.65 added