Talk:Serial digital interface

Requests for Citation
This Article contains this statement: "SDI and HD-SDI are currently only available in professional video equipment; various licensing agreements, restricting the use of unencrypted digital interfaces to professional equipment, prohibit their use in consumer equipment[citation needed]."

Someone has requested a citation for that statement.

SDI is a Standard of the SMPTE (Society of Motion Picture and Television Engineers). There are Patents on SDI and to conform with the Standard you need to obtain a License (similar to what is required to use MPEG) to use the Protocol within an electronic device (to make a Chip that understands the Protocol) and obtain permission to use SMPTE's name. SMPTE is not going to grant their permission to use this Standard within a "consumer Device" since the Standard calls for the use of unencrypted digital video (though encrypted digital video is also permitted the use of encrypted digital video within movie theaters or television stations is not done since they already know the encryption code; such codes are used at external access points and not within the buildings (much like you lock your front door but do not have locks on every door within your home)).

The reason SMPTE would not grant licenses is because they don't want someone to hook up a cheap video recorder to a feed and obtain a perfect copy of a motion picture or television signal. This does not mean you could not purchase (expensive) professional equipment nor does it mean that you could not purchase a cheap converter and hook it up to a cheap consumer device, it only means that you can not purchase a cheap consumer device with a SDI interface in it (EG: a $200 camcorder with SDI); such a device would only be useful for piracy. The alternate scenario would be to encrypt all signals within movie theaters and television stations or to search all employees to ensure they did not have a cheap digital camcorder (with an SDI Interface).

To ask for a citation in this case is similar to asking for proof that people do not have locks on all their interior doors, they do not in almost all cases; some people do not even lock their front doors (in small towns).

There is a Patent here that discusses the logic and need to restrict access to unencrypted digital video here: http://webcache.googleusercontent.com/search?q=cache:xICuJpMOHSUJ:www.freepatentsonline.com/y2011/0032981.html+serial+digital+interface+encrypted&cd=20&hl=en&ct=clnk&gl=ca&client=firefox-a or http://www.freepatentsonline.com/y2011/0032981.html

The idea is to make piracy more difficult, but not impossible. The need for a "consumer device" to have an SDI Interface is is unlikely since the consumer can simply record the degraded signal if they can claim some legal right to do so (and SMPTE does not want you to record an un-degraded signal unless you are in a position where it would be legal to do so; IE: you are a professional (a producer of the digital images) and own the rights to your own recordings).

It is no different than trying to buy a piece of wire shaped like a lock pick, you will find it very difficult to purchase even though it is legal (everywhere?) to pick the lock on your own home. If the police see you picking the lock on your own home they would likely come onto your property and ask what you are doing. It is no different than using a coat hanger to open your own car and the police asking for proof you own the vehicle (do you need a citation for that too). The small difference is that SDI is patented and SMPTE has rights to their own name. They are not going to put SDI on "consumer equipment" that would then be most useful for piracy and not particularly useful (to the consumer) for any other purpose (since the consumer most likely does not have SDI Cameras, if they did then they could pirate with that - they start at $2-3 thousand bucks). 70.71.103.239 (talk) 03:15, 12 September 2011 (UTC)


 * All the request for citation asks for is a link to some kind of documentary evidence for the assertion in the sentence. If these 'various licensing agreements' exist, perhaps there's a record of this somewhere? If they don't exist, the sentence should be rewritten. Spoonriver (talk) 07:58, 3 September 2012 (UTC)

HDSDI security cameras are available for less than $100USD. Licensing of the technology is between SMPTE and the chip makers. HDSDI serializers/deserializers and line drivers/receivers can be readily purchased. SMPTE does not limit use of unencrypted interfaces such as HDSDI to protect copyright, that is done by government. — Preceding unsigned comment added by Whitcwa (talk • contribs) 14:22, 5 June 2014 (UTC)

Photo
The photograph is of a thin lead with a 50 ohm BNC connector, but in my experience SDI is usually sent down a thicker lead with a 75 ohm BNC. Kpooji74 (talk) 17:06, 29 July 2009 (UTC)


 * I came over to this discussion page specifically to leave a note to this effect.   Nominal cable diameters are normally 4-5 times the size of the cable show in the photo.  In addition, the strain relief is normally much better. I'm happy to take a photo of one of the HDSDI cables I purchased, but I'm not sure how to upload such things to wikipedia, nor how copyright works when submitting photos of manufactured products. 3ricj (talk) 01:24, 1 February 2010 (UTC)

I just placed a photo of a more solid, true 75 ohm connector with an 8 mm cable on its back. Robijn (talk) 12:05, 12 December 2018 (UTC)

How is AMD DirectGMA related ?
I'm unsure of the purpose of the AMD-related diagram on the [|RelatedInterfaces]. It isn't related to SDI itself, nor is it mentioned in the text. It almost sounds like advertising for AMD technologies. 194.78.75.226 (talk) 09:42, 29 February 2016 (UTC)

Audio bits
Digital Betacam has 20 bit audio. This article puts the bit depth of SDI embedded audio at 16 bit. Is SDI embedded audio 20 bit capable? I can't seem to find the answer.


 * SD-SDI can carry up to 16 channels of embedded audio. The sample rate will be 48 kHz and quantizing resolution of 20 and 24 bits are supported.  HD-SDI is the same except that the audio channels are always carried at 24 bits.  Of course not all of these bits are necessarily meaningful.  The original capture of the material may be at resolutions as low as 16 bits.  That will be as much information is trully present downstream, regardless of the bandwidth of the channel.  Signals at different bitdepth are always carried such that the most significant bits are aligned with each other.  In other words, higher bit depth will provide more granularity (and thus a lower noise floor), but does not make the signal louder.  user: tvhead

Nuts
All these names for the same thing are driving me nuts. Is this the same as CCIR 602 and D1 video?

Rewrite
I rewrote the article, including lots of additional information, and removing lots of factual errors. I also removed the cleanup tag. Hope you all like it. --EngineerScotty 08:36, 5 March 2006 (UTC)

Confusion over Colorimetry / Encoding
There probably needs to be a paragraph pointing out that the values of Y, Cb and Cr are different for HD (1080i and 720p), versus SD (576i and 480i) (also ED 576p and 480p), because of the differences between Recommendations 709 and 601.

Also there's a lot of computer editing and format conversion software that doesn't recognize the overshoot margins inbuilt into the digital gamut of the ITU-R/SMPTE standards.

Then a better explanation of narrow-range and full-range of video data. How the Video payload may use 4 to 1,019, but both YCbCr or RGB need to be limited to 64~940??? And how 12bit adds 2 LSBs but scaled in a similar fashion to their 10-bit counterparts ???

Then if is always linearly sampled video data, its not actually the same as gamma/PQ encoded with fixed achromatic point format like BT.2100 table9 ???

a link to a redirect
if you click "SMPTE 292M" found under "1.1 Standards" you will be redirected th this page

an idea to remove/change it? —The preceding unsigned comment was added by 80.61.121.14 (talk) 16:00, 5 February 2007 (UTC).

Why mention G.703 at all ?
As far as I know, a G.703 line -- e.g. a telecom carrier-provided T1 (~1.5Mbit/s) or E1 (~2Mbit/s) line -- is never used to carry a SDI, let alone an HD-SDI signal. The available bandwidths are orders of magnitude too small. The only similarity between SDI and G.703 is that they are serial digital transmission channels typically carried over coaxial cable, but as the bandwidth, technologies, equipment and application domains so different, I think that any mention of a IT/telecom standard like G.703 here is irrelevant or misleading in a video context (which "SDI" generally implies) and should be removed. Besides, if serial data transmission over coax lines is to be mentioned, standards like Ethernet over thin coax (10BASE2) could be included too. 213.224.83.4


 * G.703 also defines DS3/E3 standards. These are, occasionally, used for video. —Preceding unsigned comment added by 78.154.109.115 (talk) 13:47, 22 November 2010 (UTC)

serial / parallel
Great page, but one point of confusion: This is the Serial Data Interface, but the text says "In SD and ED applications, the parallel data format is defined to 10 bits wide, whereas in HD applications, it is 20 bits wide, divided into two parallel 10-bit datastreams (known as Y and C)." So is it serial or is it parallel? Presumably it gets serialized before being sent on the one coax cable, but this reader for one is confused. Rockfox212 22:33, 21 October 2007 (UTC)

I too noticed this rather confusing mistake, in SD and ED it is indeed a serial data format (i.e. one word at a time in a series). The offending sentence has been corrected to reflect this. 93.174.217.140 (talk) 17:58, 9 May 2012 (UTC)

Supported video formats
If you are going to clarify that NTSC and PAL are used incorrectly to describe their highly compatible formats (although PAL really isn't... B/G/D/I/K are), you should also correctly define them. NTSC is technically a video system definition and does not include color (NTSC-II and NTSC-III do), PAL is technically a colour encoding method, not a video system. --tonsofpcs (Talk) 18:28, 29 March 2008 (UTC)

Requested move
I know people like to capitalize things that are known by an acronym, or that appear in the title of a standard, but in wikipedia this is not the normal convention. Plenty of sources write it as generic "serial digital interface (SDI)", so we should, too. Agreed? Dicklyon (talk) 23:44, 21 September 2008 (UTC)
 * Support - acronyms for general usage items don't have to be capitalized when spelled out and I have seen the words in lower-case form in manuals of broadcast gear. Krocheck (talk) 05:52, 25 September 2008 (UTC)

New SMPTE document numbering style
Apparently SMPTE has changed the way they refer to all of their standards: https://www.smpte.org/standards/find-standard/document-numbering

Should we start an initiative to change all the references to all the standards in all SMPTE articles? — Preceding unsigned comment added by Binba (talk • contribs) 08:05, 22 April 2014 (UTC)


 * Perhaps. But the page you link to says existing documents will only be updated when they're revised so we could arguably do the WP revisions as the individual standards are revised. ~KvnG 13:59, 3 May 2014 (UTC)

6G-SDI and 12G-SDI
1. It looks like SMPTE ST 425:2011 is a proposed standard that includes 6G-SDI and 12G-SDI, covering UHD and 4K. However I found almost no mentioning of it anywhere, and I'm not a SMPTE member. Is it a proposed, draft, unpublished, or discarded standard?

2. Given the vacuum in the 4K field, it's worth mentioning SMPTE project group 32NF-70 which is developing those standards. https://kws.smpte.org/kws/public/projects/project/details?project_id=180 Again, I have no further information, so I do wonder about its mentioned "completion date" of 30 June 2015, which seems far behind the market developments.

3. It's important to use the article to clarify that current 6G-SDI (and 12G-SDI) equipment in the market uses proprietary implementations and is not a SMPTE standard. Binba (talk) 08:14, 22 April 2014 (UTC)

Packets versus words?
Each word is stated to be 10 bits long, but how may words are in a packet? Or is the term packet here being loosely used as another term for the data word? If each packet is one word, then the packet size is 10 bits. If each packet contains multiple words, then the packet packet size is a multiple of 10 bits. This needs to be really cleared up, since the term packet is used several times in the article (refering to packets before and after V-sync, and H-sync, and also those that contain video data). — Preceding unsigned comment added by Benhut1 (talk • contribs) 07:29, 30 November 2016 (UTC)

One-way or Two-way data?
This is perhaps obvious to most but not to me.

Is this strictly a one-way data stream?

Does video signal only move from a source to a victim... or is there ANY data, packets, signals, confirmation, or anything that comes from the victim to the source? Can the source sense anything about the victim?

I suggest adding a simple sentence to detail this fact. — Preceding unsigned comment added by Sweerek (talk • contribs) 13:47, 25 February 2021 (UTC)

self-synchronizing and self-clocking
The article says NRZI and also self-synchronizing and self-clocking. NRZI is normally not self clocking, though with enough transitions a PLL can lock onto the data stream. Even so, it should not be considered self clocking like Manchester code. This is the reason that 1600 BPI 9 track PE tapes are so much more reliable than 800 BPI NRZI tapes. Gah4 (talk) 10:18, 19 April 2024 (UTC)