Wikipedia

Wikipedia:Reference desk/Archives/Computing/2017 March 21

< Wikipedia:Reference desk‎ | Archives‎ | Computing
Computing desk
< March 20 << Feb | March | Apr >> March 22 >
Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.


March 21

edit

Problems with photo rotated

edit

I have problems with photos taken with my Nikon D7100 in portrait mode. When they are transferred to my Windows 10 computer, they are in portrait mode. However, there are problems with such photos. If I try to upload one to Commons with Vicuna Uploader, it locks up on those photos. Also, I sometimes use a program to reduce the size of JPEG files. These files come out un-rotated. I've written to the author of Vicunna but gotten no reply. I have gotten a response from the JPEG software and he is looking into it. But is there something I can to to avoid these problems? Bubba73 You talkin' to me? 02:41, 21 March 2017 (UTC)[reply]

Usually the camera stores the converted information from gravity sensor in the EXIF. Users should copy the JPG files from camera as file to the computer to prevent any losts of quality. If image processing software cuts or deletes the EXIF, the picture can not be longer to be rotated automatically. --Hans Haase (有问题吗) 22:39, 21 March 2017 (UTC)[reply]
I copy the photo files to my computer using Photoshop. The JPEG shrinker takes out the EXIF data, which explains why it is un-rotated. But I don't know about the problem with Vicuna, which locks up with rotated files that have the EXIF data. Bubba73 You talkin' to me? 22:48, 21 March 2017 (UTC)[reply]
See the first section of Wikipedia:Reference desk/Archives/Humanities/2015 August 6 and the WP:VP/T and WP:GL/P results for the image in question. I used a D3200 for it, and I'm guessing that your situation and mine are similar. Nyttend (talk) 00:07, 22 March 2017 (UTC)[reply]
The D7100 is very similar to the D3200, so it could be a similar problem. It says that the file needs rotating after being downloaded, but when I downloaded it, it shows up the correct way. Bubba73 You talkin' to me? 00:23, 22 March 2017 (UTC)[reply]
One workaround to correct the probs is to display each on the screen, take a screen shot, and paste it into your fave photo editor. However, that does limit the resolution to that which can be displayed on the screen. (I'm guessing if you have such an expensive camera that you at least have a 4K resolution screen, though, which should come close to the max resolution of your camera.) StuRat (talk) 05:07, 22 March 2017 (UTC)[reply]
I have just a regular screen, and doing that would take too much time and lose too much quality. Bubba73 You talkin' to me? 15:18, 23 March 2017 (UTC)[reply]

One think I just thought of is that I can have the picture rotated in the camera first. I'll see if that makes a difference. Bubba73 You talkin' to me? 15:53, 23 March 2017 (UTC)[reply]

Okay, I remade the image from the color information, and it seems to be loading correctly now. It looks like there was some metadata in the file that's getting picked up by the displaying software (the web browser, Firefox in my case) instructing it to rotate it 90 degrees clockwise. This is a little unusual in that it (the rotation) is not something I've ever seen before, and I've been working with digital graphics professionally and as a hobby for over two decades now. My advice going forward is for you to get a program like GIMP to rotate the images (it can do batch rotations if needed). If that looks like too much to handle, there are other, simpler options out there, as well. Don't rotate them using the uploading software, the camera or the Windows photo viewer. ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 16:34, 23 March 2017 (UTC)[reply]
I'm not rotating them in the camera. I import them with Photoshop, and then I've tried several ways to view them, and all of them rotate them (as they should be seen), except Internet Explorer. Maybe I have Photoshop set to rotate them - I'll check that. Bubba73 You talkin' to me? 17:53, 23 March 2017 (UTC)[reply]
I just found this page which documents the metadata rotation I referred to. this and this cover the topic, as well. ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 16:37, 23 March 2017 (UTC)[reply]

Well, I found an option in Photoshop, and the help says

Rotate JPEGs Using Orientation Metadata/Rotate TIFFs Using Orientation Metadata

When these options are selected, the orientation metadata of the image is updated to rotate the image. The image data is untouched; the rotation is specified by just changing the metadata. Rotating an image using its metadata is a faster process than rotating the image itself.

Not all applications recognize orientation metadata. If you plan to import your images into such applications, leave these options deselected.

so I've turned that off to see what happens next time I import files. Bubba73 You talkin' to me? 20:04, 23 March 2017 (UTC)[reply]

Seagate ST4000DX001 HDD dying or not?

edit

One of my HDDs (where my games are stored) is sort of (maybe?) dying* according to some programs. No real important data stored on it, fortunately.

0.1% damaged blocks (3 red ones) according to HDTune, "Caution" according to Crystaldisk (uncorrectable sector count: 100) -- but SeaTools disagrees, as well as Speccy. (S.M.A.R.T. = Good.) Also did a CHKDSK which found nothing.

http://imgur.com/8LH5SN4 http://imgur.com/V71c0NW

Matt714 (talk) 04:48, 21 March 2017 (UTC)[reply]

Most defects going more worse. The reasons of an defect are different. It may be the structure of the platter. Sometimes still the signal amplifier on the arm of the actuator has a problem. Sometimes, just the contacts to the controller PCB are corroded. But do not try to fix even this simple failure when needing a reliable mass storage. Steve Gibson argued, some OEMs would not copy the map of bad sector which are marked as bad at formatting. When the stored data has the expected level of magnetism when reading, no error is detected. But if the data of the physical bad sector is being replaced, the error occurs when an other magnetic level could not be stored at that point. It also requires to verify the stored information or checksum it. --Hans Haase (有问题吗) 08:41, 21 March 2017 (UTC)[reply]
  • Crystaldisk is not a generally trustworthy application: Most CNET reviews are 1-star, due to included spyware. I'm not surprised it told you that it found problems, as untrustworthy diagnostic software will always find problems. HDTune is reliable enough, according to Tom's Hardware and CNET reviews. 0.1% damaged blocks is not that unusual, especially on a disk where applications are stored. SeaTools is the only application you mentioned which I've used, and I have no complaints. It does it's job with a relatively small footprint, doesn't waffle about with proprietary measurements and is easy to use. I would trust it. ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 13:47, 21 March 2017 (UTC)[reply]

The comments above aside, I don't see that the tools actually disagree on anything significant. They key point seems to be that the disk still passes SMART but has developed 100 bad sectors. I'm fairly sure if you look at HDtune's SMART details, you fill find it also has 100 bad sectors. How long has the disk had these bad sectors?

Contrary to what ᛗᛁᛟᛚᚾᛁᚱPants says, I would say 100 bad sectors is actually not that common on hard disks even with use. Yes it may only be 0.1% but we're not talking about an SSD here and this obviously developed since the drive had left the factory otherwise they wouldn't show up as bad sectors. Sure it happens but it's enough to start to cause concerns hence why different tools have different opinions. (Seatools relies mostly on SMART or any errors during self tests. It's really primarily about RMAs and secondly providing utilities to do stuff than about detailed health information.)

If you use a tool to write on these sectors, they should be re-allocated (the disk should have more than 100 spare sectors) and will show up in the reallocated sector count instead. Definitely 100 bad sectors is often enough to ask for an RMA (especially on a new disk), although you may just get a refurbished drive that also has 100 bad sectors or more but these don't show up anywhere since like with a new drive it's only internal to the drive so may not be the best idea unless the drive is very new and you're dealing with the retailer so have a hope of getting a new drive.

I'd strongly suggest you keep an eye on it with a tool that actually tells you how many bad sectors (both reallocated and non-reallocated) show up in SMART. If these continue to increase, then there's a good sign you drive is going to die sooner rather than later. Speedfan's online submission thing is useful provided you keep a copy of the URL (make sure you have the right URL, if the URL is http://www.hddstatus.com/hdrepanalysis.php you don't). But you could also just copy the readings from HD tune or anything else. On the other hand, if it's stable long term with 100 bad sectors then the drive is probably not that much worse than one of the same vintage etc without any. One thing you didn't mention, are these 3 blocks in HDtune actually consecutive or in different places? If they're in 3 different places it's probably also slightly more concerning.

P.S. I've used Seatools many times before, and also HDtune. Seatools is definitely useful for many things but given its design you have to take care with it. To give an example I have a Seagate disk with IIRC 500+ bad sectors, slowly increasing. When I zero or write the whole disk, it reallocated what it could. But actually some of the bad sectors were reassessed and seemed fine so weren't reallocated. But either the sectors or other ones would reappear and as also said, the number was slowly increasing. Clearly not a healthy disk. A lot of the time though, it would pass Seatools.

P.P.S. I'm assuming here that the bad sectors are just ones that are pending write for reallocation rather than one that can't be reallocated because the drive is out of spare sectors. You should look carefully though. This is one area where the Speedfan's online submission thing is useful (which uses HDDStatus.com so you could probably find other tools) since it provides some explaination of the SMART variables of concern to those who don't understand them. You don't of course need online submission for that, but it doesn't sound like any of you tools have been very good in that regard and HDDstatus.com does IMO do a decent job and I find them useful for other reasons. (Well I know Seatools and HDtune are only really useful for raw data.)

If it is in fact 100 sector that can't be reallocated due the drive being out of spare sectors, I'd consider this a very big sign the drive is probably dying. I believe though that such a drive should fail at least one Seatools test, either SMART, long drive self test or zero fill (well the last is not technically a test but it can still fail it) when this happens though.

Nil Einne (talk) 01:19, 22 March 2017 (UTC)[reply]

Thanks all for your answers. Just re-passed the SeaTools diagnostic tests and it passed S.M.A.R.T and Short Generic, but failed Short DST. I will contact Seagate for a RMA. Matt714 (talk) 19:39, 22 March 2017 (UTC)[reply]

Lowercase name entry in emails leads to spam filtering?

edit

When I send an email, the recipient usually gets a name to pop up instead of the email address, because the sender also has a name field in their settings where they can enter a first and last name. I am wondering if there is an effect on spam filters if the first and last name field are in all lower case. (Clarification: I am not talking about the email address).2601:281:8400:B440:D520:1FA3:BFF:733 (talk) 22:42, 21 March 2017 (UTC)[reply]

How are you sending your email? Through Outlook/Windows Mail/etc., or are you sending it straight from the server, or some other way? I wonder if the address book's settings may affect the mail's outgoing appearance and thus what the receiving spam filter sees. Nyttend (talk) 00:04, 22 March 2017 (UTC)[reply]

For clarify are you saying that the to: or from: or other address fields of the email actually has a name in it with all lower case? There are no first and last name fields that are a standard part of email headers. It is possible to include a display name in the various address fields, but fortunately this display name doesn't deal with things like first names and last names [1] which are concepts which don't apply to all names. Email clients (both web and standalone) may themselves deal with first names and last names if they desire, and they could also have internal names which aren't added to the outgoing email or coming from the incoming email.

Anyway if you are referring to the display name then yes, it's possible spam filters will be affected by this in various ways, including depend on casethe case usage in the display name field. For example, some spam filters use some degree of machine learning and if they found that an all lower case display name is more likely to indicate that an email is spam or not spam, it may adjust it's spam score based on this. Even manually programmed rules could have this if the designer, felt it was useful. I don't deal with the nitty gritty of spam filters so I have no idea if any known ones do such things but I'd note a number of large email service providers use largely secret algorithms probably using some degree of machine learning so we have no way of knowing for sure even if none which do so are known. In fact with machine learning it can be difficult for even people who control the thing to know depending on the complexity of what it does (if it always modifies a score you can easily test this, however if it only changes the score in some cases then this may be difficult to detect).

Nil Einne (talk) 12:01, 22 March 2017 (UTC) 10:21, 23 March 2017 (UTC)[reply]

Retrieved from "https://en.wiki.x.io/w/index.php?title=Wikipedia:Reference_desk/Archives/Computing/2017_March_21&oldid=1088105317"