Broadcast safe levels

Do you have questions about Desktop Video, Converters, Routers and Monitoring?
  • Author
  • Message
Offline

litote

  • Posts: 70
  • Joined: Tue Nov 27, 2018 1:16 pm
  • Real Name: Greg Hughes

Broadcast safe levels

PostFri Jan 04, 2019 9:47 am

This is a question I have put to the Davinci Resolve forum but am not getting any clear-cut answer so I thought some BMC user might be able to answer it:

Regarding broadcast safe luma levels for Rec 709, in Davinci Resolve, can someone confirm that Davinci automatically ensures these are legal even if the luma levels in the scope reach 0 and/or 1023 in the waveform monitor to the point of clipping? Or, should they be set between the range of 64-940 as I have seen in two tutorials?

Thanks
Offline

Dermot Shane

  • Posts: 2720
  • Joined: Tue Nov 11, 2014 6:48 pm
  • Location: Vancouver, Canada

Re: Broadcast safe levels

PostFri Jan 04, 2019 6:32 pm

Greg:
it seems your not happy with the answers here;
viewtopic.php?f=21&t=84165

but there's folks with deacdes of delivering high level content and folks with engineering degrees talking... l'd advise re-reading and ask about what you do not understand..

it's likely some camera centric folks have an engineering degree, but most likely not a ton of seat time delivering to high level online delivery, mainstream broadcasters, and studios

so i guess when you ask about creating legal masters you need to define a bit tighter what level of "legal" you are targeting... on that list Andrew delivers to EBU broadcasters, one set of standards to meet, Tom seems ot deliver to American broadcast, so another defination of "legal".. i mainly ship my master to Deluxe or Technicolor London for QC, and i'm meeting international delivery standards.. a somewhat wider web to get caught in

maybe tell us that version of "legal" your are concerned about?
Offline
User avatar

JPOwens

  • Posts: 1511
  • Joined: Fri Apr 12, 2013 8:04 pm
  • Location: Victoria, British Columbia, Canada

Re: Broadcast safe levels

PostSat Jan 05, 2019 7:27 am

litote wrote:even if the luma levels in the scope reach 0 and/or 1023 in the waveform monitor


The questioned ambiguity here is down the street of why I am not terribly fond of the internal scopes supplied by the application. In the end, "0"-1023 is an internal range, relative to Resolve's float processing. If you were using an external scope and monitoring the display-referenced output, this would be obvious -- and underscores why it is important to understand the mapped range that is assumed with various codecs and matrixed outputs. Yes, you can export full-range RGB -- usually by mistake because certain containers cannot flag the appropriate decode/mapping for software players. This is why there is yet another "gamma bug" thread every other day here on this forum. As of the adoption of 2.4 and the death of Final Cut, *there is no more gamma bug*. In 999,999 cases out of 1,000,000 it is because someone has messed up the range in export. This is a completely different issue than "broadcast legal" because 1023 would be perfectly legal if it was decoded correctly. It would be mapped to 940. In sympathy, this has been a problem for over a decade now, and confuses the hell out of users who up to this point haven't realized that digital media can even be expressed in different codec regimes and ranges. For them, RGB images exist between 0 and 255 or greater given 2* extra bit depths and that was all there was to it. Quite obviously, wrong. As you work further into cinematography and scene/display reference, --- that's not even for starters a fundamentally flawed concept, it's a fatally blind start, right over the cliff out of the gate.

Scope-reading. I guess it's just as deep a mystery as it ever was.

jPo, CSI
Offline

litote

  • Posts: 70
  • Joined: Tue Nov 27, 2018 1:16 pm
  • Real Name: Greg Hughes

Re: Broadcast safe levels

PostSat Jan 05, 2019 9:08 am

@Dermot Shane

Yes, that forum has since helped more to answer my questions to some degree, though went off on an irrelevant tangent.

I am ignorant about the various HD Rec 709 legal ranges between different broadcasters and bow to the knowledge of these forum posters, though I was aware that they vary. I would rephrase my question to: is there a conservative standard that would be accepted universally for luma levels, e.g. 64-940? If I am correct, this is more conservative than the corresponding level you see in the waveform monitor of 1-1023 (which I think translates to 0-100 IRE)? There is a Data Levels option of Video in Davinci which I am guessing corresponds to this range of 64-940.

From two replies in that other forum I am assured that by just setting Data Levels to Auto or Video in DELIVER-Render Settings there should be no problems, in their experience at least. They didn’t mention a particular broadcaster’s standard or very clearly explain why there is both an Auto and Video option and if there is any difference between them, or why you should leave it at default (Auto).

This is the only specific instruction that was posted on how to set a legal luma level in Davinci, so the other posts were of little help. If I must manually set the legal values for different broadcaster standards for luma levels, it was not answered as to how to actually set them in Resolve.

Further complicating the issue is that I am aware there is another way of setting limits, via a LUT.

Having since contacted the professional colorist whose course I referred to, he suggests to bring the output footage that you apply this measure to back in to Resolve to check if the levels in the scopes are now within acceptable limits, which suggests that he does not have as much confidence in this auto-legalizing function as these two other posters.
Offline

Howard Roll

  • Posts: 2506
  • Joined: Fri Jun 03, 2016 7:50 am

Re: Broadcast safe levels

PostSat Jan 05, 2019 9:43 am

litote wrote:This is a question I have put to the Davinci Resolve forum but am not getting any clear-cut answer so I thought some BMC user might be able to answer it:

Regarding broadcast safe luma levels for Rec 709, in Davinci Resolve, can someone confirm that Davinci automatically ensures these are legal even if the luma levels in the scope reach 0 and/or 1023 in the waveform monitor to the point of clipping? Or, should they be set between the range of 64-940 as I have seen in two tutorials?

Thanks

For video level clips the 0 to 1023 range on the Resolve scopes is actually 64-940, it's a Resolve idiosyncrasy.

If your export is set to video level and you don't check preserve super whites your exports will be luma safe.

You can test this yourself in about 5 minutes. Crank the highlights on a clip and export it both ways. The version exported with super whites will have data above resolves 1023 ceiling, the unchecked export will not.

Scout's Honor.
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostSat Jan 05, 2019 11:41 am

litote wrote:From two replies in that other forum I am assured that by just setting Data Levels to Auto or Video in DELIVER-Render Settings there should be no problems, in their experience at least. They didn’t mention a particular broadcaster’s standard or very clearly explain why there is both an Auto and Video option and if there is any difference between them, or why you should leave it at default (Auto).

This is the only specific instruction that was posted on how to set a legal luma level in Davinci, so the other posts were of little help. If I must manually set the legal values for different broadcaster standards for luma levels, it was not answered as to how to actually set them in Resolve.

Further complicating the issue is that I am aware there is another way of setting limits, via a LUT.


If you concerned only about luma then there is really 1 standard (at leats for modern/current digital world). Luma must be in legal levels (64-940 for 10bit, 16-235 for 8bit) and that's about it. Broadcasters may have different threshold how much it can go below/above it. Most use EBU R103 standard which allows for luma to go few % outside perfect levels. Current spec is here:
https://tech.ebu.ch/docs/r/r103.pdf
As you see luma can be within 20-984, so there is a fair amount of margin.
There should be also more than 1% of the whole frame with bad pixels to be treated as invalid signal. All those thresholds are there to cover lossy compression overshoots (they do happen even for intermediate codecs). If you encode something to ProRes or DNxHR and look at high contrast areas you will find blacks going <64.

As I said- you really making it more complicated then it's. Don't every tick "preserve super blacks/whites" and your files will never fall outside those "perfect" levels and fail for luma errors on QC. It's not so "easy" with some other NLEs which operate in YUV. Then you can easier go out of safe.

If you read this document carefully then you will find this:

"When television signals are manipulated in YUV form, it is possible to produce "illegal" combinations that, when de-matrixed, would produce R, G or B signals outside the range 0% to 100%."

Resolve operates in RGB and then data is converted to YUV on export, so by staying away from super blacks/white setting you should really never fall into problems (if later QC is done properly). You are protected by pure math :)
If you don't believe use do what Howard suggested.
Offline

litote

  • Posts: 70
  • Joined: Tue Nov 27, 2018 1:16 pm
  • Real Name: Greg Hughes

Re: Broadcast safe levels

PostSat Jan 05, 2019 11:57 am

Thank you all for your helpful replies.

I have only been able to come as far as I have in my humble understanding of Resolve and other areas thanks in large part to individuals like yourselves who take time out of their day to help a stranger. It is very much appreciated.
Offline

Dermot Shane

  • Posts: 2720
  • Joined: Tue Nov 11, 2014 6:48 pm
  • Location: Vancouver, Canada

Re: Broadcast safe levels

PostSat Jan 05, 2019 6:27 pm

litote wrote:I would rephrase my question to: is there a conservative standard that would be accepted universally for luma levels, e.g. 64-940? If I am correct, this is more conservative than the corresponding level you see in the waveform monitor of 1-1023 (which I think translates to 0-100 IRE)? There is a Data Levels option of Video in Davinci which I am guessing corresponds to this range of 64-940.

Having since contacted the professional colorist whose course I referred to, he suggests to bring the output footage that you apply this measure to back in to Resolve to check if the levels in the scopes are now within acceptable limits, which suggests that he does not have as much confidence in this auto-legalizing function as these two other posters.

@ Greg
first i think you are far ahead of alot of folks, by asking these questions.. smart move

-second on the levels thing, i would get a rejection out of hand if i handed a film over to QC with video levels, instant fail, and i would get the same responce if i handed a TV movie over to QC with data levels

so circle back to what is the deliverable? read the tech spec's carefully!

- like your professional colorist tutor, i would NOT be even a tiny bit close to trusting to Resolve to automagicly create a legal master, no how, no way.. and mentioned that in the Resolve forum thread ;-)

but my version of "legal" and someone else's can be very diffrent, and the stakes are pretty high in my world, the daily intrest charges on the money used to make a film are significant, the distb will not release their MG until 45 days after the QC report says "pass".. a one day delay in passing QC can easly cost a producer far more than a day in a gradeing suite getting it right the first time and passing what ever outdated stupid requirments the disb forces one to meet....
Offline
User avatar

JPOwens

  • Posts: 1511
  • Joined: Fri Apr 12, 2013 8:04 pm
  • Location: Victoria, British Columbia, Canada

Re: Broadcast safe levels

PostSat Jan 05, 2019 8:23 pm

Andrew Kolakowski wrote:"When television signals are manipulated in YUV form, it is possible to produce "illegal" combinations that, when de-matrixed, would produce R, G or B signals outside the range 0% to 100%."


Yes, that is accurate -- part of what I was alluding to previously in this and other threads.

To get to the scope-reading heart of the matter, the notions of 0-1023 and 64-940 are actually irrelevant with respect to measuring output values, except within your relative ranges.

It is as irrelevant now as the concept of "Pedestal" is in digital video. And I see newbs cranking up the black to 7.5 IRE. NO! Analog composite RS-170A. Get over it.

So if the scope actually showed IRE units, then it's relevant. That really only comes out the back end of an I/O device; the "forced" notion of a Decklink or UltraStudio port. Then the meniscus scale of 0-100 IRE is what you would be measuring. And the various tolerances that the CCIR deems acceptable. These depend absolutely on matrixing, as above. If you can show that Y+C never exceeds the outer limits of -20 to +120 IRE, then you are good to go. Very few modern software scopes can do this. Resolve can't show it.

As Marc *im-politic*-ally put it, the old-timers see it as plainly as the piano keys on the end of a runway when you've got an approach to judge. Haven't collapsed a nosegear yet. Do remain seated with your seatbelts fastened until we reach the gate.

jPo, CSI
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostSat Jan 05, 2019 10:26 pm

"Recommendation
The EBU, considering that,
 video levels have traditionally been measured with devices that display a trace, such as a traditional waveform monitor,
that readings in mV no longer give relevant information in digital signal infrastructures,
 television systems now include high dynamic range and wide colour space images as well as
standard dynamic range and colour space images in the same digital container,
 that a certain tolerance can be allowed in digital signal levels,"

Time for industry to actually try to follow EBU recommendation. Maybe BM can start by removing legacy Broadcast Safe option which is more misleading than useful.
Offline

Peter Cave

  • Posts: 3768
  • Joined: Thu Aug 23, 2012 6:45 am
  • Location: Melbourne, Australia

Re: Broadcast safe levels

PostTue Jan 08, 2019 5:00 am

We still must deliver to outdated delivery specs that include measurements in IRE, mV and bits. I would prefer to keep the legacy legalise options as they can be ignored for most delivery of masters. Perhaps relabelling the Broadcast Safe option to make it clearer to end users what it's function is?
Resolve 18.6.5 Mac OSX 14.4 Sonoma
Mac Studio Max 32GB
Offline
User avatar

Marc Wielage

  • Posts: 10901
  • Joined: Fri Oct 18, 2013 2:46 am
  • Location: Hollywood, USA

Re: Broadcast safe levels

PostTue Jan 08, 2019 5:40 am

To go back to something Joe P mentioned earlier, I would recommend that anybody who's interested in the problem of "what's legal and what's not," read the (free) Tektronix document "Using Waveform Monitors as Artistic Tools in Color Grading":

https://www.tek.com/document/primer/gui ... resolution

You can make a good case that being conservative on reading scope values will generally help, especially if you're trying to get just under the maximum range accepted by the distributor.
marc wielage, csi • VP/color & workflow • chroma | hollywood
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostMon Nov 21, 2022 6:56 pm

Thanks for all the information! I am just scanning my files in QCTools to chekc for illegal broadcast color values. I have edited in final cut and have problems especially with drone footage that has sub zero color value spikes. I tried the broadcast safe filter in resolve and gamut mapping but still QCTools gives me a very small amount of spikes beyond legal range. With all the other filters its the same in AVID, Premiere and FCPX - even the BCC Broadcast Safe Effect. So after reading the discussion I am not sure if those values can be accepted when they cannot be filtered.

I guess this passage in the QCTools manual for the BRNG = Broadcast Range Filter corresponds to the 1% rule of the EBU-Guidelines: "Typically anything with a value over 0.01 will read as an artifact."

The spikes have a maximum of 0,008, when I push the button on the right to "hide spikes" everything turns green. Is this accepted? Or do I have to worry about those spikes?
Attachments
Screenshot 2022-11-21 at 18.46.13.png
Screenshot 2022-11-21 at 18.46.13.png (12 KiB) Viewed 8831 times
Screenshot 2022-11-21 at 19.51.03.png
Screenshot 2022-11-21 at 19.51.03.png (38.41 KiB) Viewed 8831 times
Offline

SamBroggs

  • Posts: 75
  • Joined: Thu Aug 31, 2017 7:40 pm

Re: Broadcast safe levels

PostMon Nov 21, 2022 8:50 pm

I have to deliver for German broadcast so it’s ARD ZDF EBU specs.
mbp late 2012, 16gb ram, 2,6ghz, geforce GT650M, davinci resolve 15
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostMon Nov 21, 2022 9:07 pm

Resolve color safe filter is useless for modern delivery as it's based on analogue normalisation which basically no one uses today and definitely not ARD. You need EBU R-103 legalisation.

Files exported directly from Resolve (you can't just export without any processing as Resolve won't go over RGB) should be EBU-R103 safe (don't later try to do any manipulation in NLEs which operate in YUV).
If QC tells you that you're not compliant then ask about thresholds. Some places have 100% limit for the signal without any thresholds, which is basically in compressed (counts for any lossy compression including ProRes, DNxHR, etc.) world impossible to achieve. You would have to artificially limit signal to some unknown, way <100 levels and hope (as there is still no guarantee) that codec still doesn't produce single pixel overshoot which goes >100%. Any proper delivery spec should follow R-103 spec or be at least realistic (bit smaller or higher thresholds), but never expect 100% perfect pixels as it's practically impossible for typical delivery (specially broadcast which uses relatively highly compressed formats).

Files going to ARD/ZDF need to follow precisely their presets. Resolve doesn't offer it, AME does.
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostTue Nov 22, 2022 9:17 am

Thanks a lot Andrew!

So you think values of 0,008 should be fine for QC?

This film has been edited in FCPX, does the broadcast range filter in AME work 100% reliable so I can convert it directly there or do I have to roundtrip to resolve and then again to AME? I think to much re-encoding is not beneficial for QC.
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostTue Nov 22, 2022 9:32 am

Don’t know as I don’t use those QC tools. If they detect out of gamut errors according to R-103 spec then fine. AME filter works fine and it does small compression, not just hard clipping ( don’t use high values for compression though). No need to go over Resolve.
Higher quality master the better and you always need to start with better quality than final delivery ( or export directly to desired format if possible).
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostTue Nov 22, 2022 9:42 am

I read description of BRNG filter and it’s not good it just checks video levels. Gamut detection is different story.
Offline

Steve Fishwick

  • Posts: 1006
  • Joined: Wed Mar 11, 2015 11:35 am
  • Location: United Kingdom

Re: Broadcast safe levels

PostTue Nov 22, 2022 10:06 am

A lot of misinformation in this thread I am afraid. Millivolts are still used in just about every finishing suite here, that I have worked on, for all the major broadcasters and Telestream's Vidchecker (the de facto QC tool here), very rarely shows out of gamut RGB fails, more usually only PSE and blanking, when using a safe limiter clamped to 700 milivots, at least on Avid Symphony. All scopes here, everywhere I see, are set for millivolts. Though it doesn't actually matter if they are set for millivolts, IRE or %, they are all working within the digital realm now. It would only matter if we still used analogue scopes, which we don't.

The document referred too, which is the same, as the DPP here in the UK, merely says:

Experience has shown that colour gamut "legalisers" should be used with caution as they may create
artefacts in the picture that are more disturbing than the gamut errors they are attempting to
correct. It is advisable not to “legalise” video signals before all signal processing has been carried
out.


Which merely means if you only clamp a signal, without grading it, with reference to the scopes first, you will be clipping it badly and therefore those artefacts are obviously more noticeable than just being legal, which is the final stage. I grade with the top safe limiter switched off first, bring the levels within scope and grade shot to shot, like most every finishing editor and then switch the safe limiter track on at the end. Safe limiters, particularly software ones are never 100% catchalls but if you've graded a show, which is obviously is normal, there shouldn't be a problem.
Offline
User avatar

Marc Wielage

  • Posts: 10901
  • Joined: Fri Oct 18, 2013 2:46 am
  • Location: Hollywood, USA

Re: Broadcast safe levels

PostTue Nov 22, 2022 10:35 am

Andrew Kolakowski wrote:Resolve color safe filter is useless for modern delivery as it's based on analogue normalisation which basically no one uses today and definitely not ARD. You need EBU R-103 legalisation.

Assuming the problem is illegal chroma excursions, one thing we do on selected shots is to use an OFX Color Space Transform node at the end (on a shot by shot basis) with Saturation compress to bring it down by maybe 3 or 4 or 5%. I don't use it on every shot, but if we have a case with (say) bright police car lights or intense neon, then we bring that out at the very end to limit the excursions for Rec709. Sometimes we'll include a soft highlight clip as well right after that. We set up the node structure in a way that if we need an HDR pass, we can bypass those nodes and let the sat hit a lot harder.

I would like to see a true "Broadcast Legalizer" as an OFX Plug-in or as a DCTL, but until then, this comes pretty close. Again, I'm not a fan of stomping on the signal all the time, but this will work for problematic scenes.
marc wielage, csi • VP/color & workflow • chroma | hollywood
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostTue Nov 22, 2022 11:04 am

Thanks everyone. Maybe I will have to treat those shots individually. Its mostly drone shots from a mini 2. For some strange reason it constantly creates negative values in the RGB parade. There is no way to change picture profile settinges in the mini 2..
Attachments
Drone Spikes subzero.jpg
Drone Spikes subzero.jpg (87.6 KiB) Viewed 8676 times
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostTue Nov 22, 2022 11:06 am

And I received a reply from QCTools, they are very helpful, I hope the they will continue their great work.

"I have had confirmation that we do check for the ARD_ZDF parameters but we don’t have a dedicated template for it at the moment.

However, this is on the roadmap and will be released as a template early next year."

In the meantime I will be looking for som QC Software that is really pay per use. Software like Pulsar wants me to pay a minimum of 100$ deposit every six months. So for me who has only a few films every year to scan its a bit on the expensive side. Does anyone know a software thats REALLY pay per use without any regular minimum deposits?
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostTue Nov 22, 2022 11:33 am

You can use Telestream cloud (https://www.telestream.net/telestream-c ... ontrol.htm) which charges per minute of checked content (no other fees), but...

Look at ARD/ZDF spec- it's public. Typically gamut and loudness are most complex checks, so once you have this covered (AME will do both) rest is just making sure other bits of the spec look fine- levels, blanking, safe ares, etc. This can be achieved by human+free tools checks.

If you create master from scratch then you can play with Resolve and do "creative" fix, but for existing/bulk masters you can just rely on legalisers, so more drastic (but used a lot) approach.
Last edited by Andrew Kolakowski on Tue Nov 22, 2022 12:24 pm, edited 1 time in total.
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostTue Nov 22, 2022 11:54 am

Petehikers wrote:Thanks everyone. Maybe I will have to treat those shots individually. Its mostly drone shots from a mini 2. For some strange reason it constantly creates negative values in the RGB parade. There is no way to change picture profile settinges in the mini 2..


Those (I think) are overshoots from compression (maybe heavily saturated footage?). There are YUV combinations in your signal which create negative RGB values after conversion to RGB. Resolve should take care of it.
This is exactly sort of thing which could cause a problem if you do editing with tool which operates in YUV, eg. Edius. YUV processing is not bad thing, but it can have these problems, so you have gamut legalisers, which will slightly compress or hard clip them. If you use such a source in Resolve then after exporting final master you should not have those problems as Resolve operates in RGB and then on export converts to YUV. This means that reverse YUV->RGB can't produce bad values as long as you did not process file further after Resolve export. There will be small overshoots, but those come after compression and are unpredictable, but covered by R103 thresholds.

It's enough to put bad file through Resolve and it will "fix" it (by clipping I assume). You just have to add some processing to force Resolve go over RGB (enough to put eg. color space conversion filter without any setup).
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostTue Nov 22, 2022 12:32 pm

Thanks! Seems that grading in FCPX was not the best idea. So if AME takes care of all of this then I think I could skip going over Resolve.

By the way: I found out that if you control the Drone with the app Litchi you can dial down contrast, saturation and sharpening a little. I hope its not another added processing step after capture..
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostTue Nov 22, 2022 12:44 pm

AME will fix it and it won’t be really visible.
Offline

Steve Fishwick

  • Posts: 1006
  • Joined: Wed Mar 11, 2015 11:35 am
  • Location: United Kingdom

Re: Broadcast safe levels

PostTue Nov 22, 2022 12:54 pm

Marc Wielage wrote:Again, I'm not a fan of stomping on the signal all the time, but this will work for problematic scenes.


In practice Marc, no signal now is completely stomped on. This was only necessary in the days of older transmitters and CRT analogue TVs. Soft limiters particularly let through quite a bit of RGB gamut overshoot and there is a % allowed either side. For broadcast here (and international versions I deliver a lot) a safe limiter is always applied on the final output. If you didn't apply one, they would in post, before the final AS-11 or Prores HQ, respectively. Though of course we do so (in broadcast) because of these old lowest common denominator legacies whereas even more flexibility, would not be the end of the world now.
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostTue Nov 22, 2022 12:55 pm

Andrew Kolakowski wrote:You can use Telestream cloud (https://www.telestream.net/telestream-c ... ontrol.htm) which charges per minute of checked content (no other fees), but...

Look at ARD/ZDF spec- it's public. Typically gamut and loudness are most complex checks, so once you have this covered (AME will do both) rest is just making sure other bits of the spec look fine- levels, blanking, safe ares, etc. This can be achieved by human+free tools checks.

If you create master from scratch then you can play with Resolve and do "creative" fix, but for existing/bulk masters you can just rely on legalisers, so more drastic (but used a lot) approach.


Thanks! I think all the other checks can be done by QCTool as well. I just wanted to have peace of mind to be sure the AME filter takes care of everything. So just one scan for the final output to see some green lights and be happy. :)
Offline

Steve Fishwick

  • Posts: 1006
  • Joined: Wed Mar 11, 2015 11:35 am
  • Location: United Kingdom

Re: Broadcast safe levels

PostTue Nov 22, 2022 1:59 pm

Just a word of caution to anyone reading into delivery specs, re setting safe limiter limits and from practical experience delivering for broadcast, they usually say something like:

RGB components must be between -5 % and 105% (-35 and 735mV)
and
Luminance (Y) must be between -1% and 103% (-7mV and 721mV)

If you set the software limiter to these broader limits you will almost certainly get an RGB QC fail, because of the overshoots inherent in software limiters, I mentioned. We always set the limiters to 700mV precisely. It will still give you broadly these overshoots and then it very rarely ever fails, at least through Vidchecker, in this respect and there has never been an issue with full QC, eyeball or otherwise, using this method, either.
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostTue Nov 22, 2022 4:38 pm

i just installed ScopeBox and it works great. However it finds LOTS Luma and Chroma excursions even with FCPX broadcast filter set to a higher level. Transcoding in AME and importing back made it even worse.
The thing that helped is importing in Resolve, lifting blacks for the whole timeline for a very small increment and adding a color space transform with saturation compression and custom luma max. Its strange as the values even in the luma and RGB TimeTrace look fine most of the time but there are still massive alerts..
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostTue Nov 22, 2022 5:11 pm

Your thinking that it has to be clean of any warnings is wrong.
Any slightest compression does overshoots and scopes can always show warnings.
This is why R103 spec exists and it defines thresholds within which overshoots are allowed to happen. Proper deliver spec should always use thresholds and then you don’t have to go crazy :)
Don’t try hard to overdo it.
If you do grading you have quite a lot of creative control in Resolve and your export will be always fine anyway.
Issue is when you receive files etc. and just need to deliver them to certain spec. Then you just put them through legaliser and job done. No one is going to try fix it ‘creatively’ as there is no time for it.

If you put your file through AME, R103 preset will clear it. You don’t check that in scopes but in QC tools which operate based on R103 spec, Vidchecker, Baton etc. This is how broadcasters check it.

If for some reason file still fails then use stricter 103% preset or 100% one. I delivered 1000s of files worldwide and never had issues. Had case with delivery spec with no thresholds, but I ‘fixed’ it over 1 email :D

There use to be nice Edius plugin which not only shown warnings, areas of bad pixels, but also had an extra flag showing if given frame passes R103 check or not. It was made by people for the people with practical features. So many tools have plenty features which are not practical and useless at the end.
Offline
User avatar

Uli Plank

  • Posts: 21278
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: Broadcast safe levels

PostWed Nov 23, 2022 2:23 am

I can highly recommend OmniScope: https://timeinpixels.com/nobe-omniscope/
No, an iGPU is not enough, and you can't use HEVC 10 bit 4:2:2 in the free version.

Studio 18.6.5, MacOS 13.6.5
MacBook M1 Pro, 16 GPU cores, 32 GB RAM and iPhone 15 Pro
Speed Editor, UltraStudio Monitor 3G, iMac 2017, eGPU
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostWed Nov 23, 2022 8:05 am

Uli Plank wrote:I can highly recommend OmniScope: https://timeinpixels.com/nobe-omniscope/

Thanks! Is it better than scopebox and why?
Offline
User avatar

Uli Plank

  • Posts: 21278
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: Broadcast safe levels

PostWed Nov 23, 2022 8:19 am

The Pro version can make detailed error logs. The last time I've tried ScopeBox it didn't. Or does it now?
No, an iGPU is not enough, and you can't use HEVC 10 bit 4:2:2 in the free version.

Studio 18.6.5, MacOS 13.6.5
MacBook M1 Pro, 16 GPU cores, 32 GB RAM and iPhone 15 Pro
Speed Editor, UltraStudio Monitor 3G, iMac 2017, eGPU
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostWed Nov 23, 2022 11:11 am

OmniScope checks for gamut errors only in Pro version, but yet again- it has nothing to do with R103 spec in terms of thresholds. It helps, but it's not final check.
You just normalise file, check with Telestream Qualify and done.
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostWed Nov 23, 2022 2:18 pm

Thanks a lot! I am still waiting for an answer from Telestream, it seems like I cannot just register and start using it.
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostWed Nov 23, 2022 2:40 pm

Uli Plank wrote:The Pro version can make detailed error logs. The last time I've tried ScopeBox it didn't. Or does it now?

It does
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostWed Nov 23, 2022 3:13 pm

Petehikers wrote:Thanks a lot! I am still waiting for an answer from Telestream, it seems like I cannot just register and start using it.


Why not ?
You shouldn't even have to talk to them. You register, add card details and done.
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostWed Nov 23, 2022 3:21 pm

Hmm.. people dont seem too optimistic about adobes limiter, maybe i will try i avid as well..
https://community.adobe.com/t5/premiere ... p/12151346
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostWed Nov 23, 2022 3:37 pm

I know gamut limiter works (or at least did last time I used it) fine. No idea what people do :)
It's a gamut limiter, wich is not exactly the same thing as YUV levels limiter.
And last time- after compression some spikes (seen on scopes) will be always there.
Seems like those people don't exactly know what it does and what should be the end result.
I don't have much more to say. You are master of your masters :)

Unless you can do proper check against R103 spec you are bit "blind in your fixes".
Offline
User avatar

Marc Wielage

  • Posts: 10901
  • Joined: Fri Oct 18, 2013 2:46 am
  • Location: Hollywood, USA

Re: Broadcast safe levels

PostThu Nov 24, 2022 6:27 am

Steve Fishwick wrote:In practice Marc, no signal now is completely stomped on.

Perhaps we see things differently.
marc wielage, csi • VP/color & workflow • chroma | hollywood
Offline

Steve Fishwick

  • Posts: 1006
  • Joined: Wed Mar 11, 2015 11:35 am
  • Location: United Kingdom

Re: Broadcast safe levels

PostThu Nov 24, 2022 7:44 am

Marc Wielage wrote:Perhaps we see things differently.


It's not a different view, Marc, just perhaps we work in different fields, yours being a higher one :) . I work almost exclusively in broadcast and I was alluding really to the difference between modern software limiters as opposed to older hardware ones. I work to standards that are really a legacy, interlaced and RGB limited etc. that should not really be necessary now and a safe limiter on the top track is mandatory, for us. In that respect I agree totally with you but, as you know only too well, we hopefully use the waveform, whilst grading, to ensure that limiter is not stomping on the final output in an ugly way.
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostThu Nov 24, 2022 10:22 am

For sure it has to be graded and it has to look good even with the limiter applied.
And for me I can say it does and in the scopes everything looks fine - even with limiters.
Its only about those excursions I want to filter.
As I see it, FCPX Broadcast filter does not RGB limit or limit chroma excursions.
Adobes Broadcast Legal Effect limits RGB to 0-255 even in the strictest setting, so I guess I have to go through avid or resolve if I want to be RGB-safe, right?
The Avid-Filter filters RGB-Levels and as you told me Resolve does it automatically - so going from resolve or AVID to AME for my final delivery should be fine, but going from FCPX to AME should not work. Please correct me if I am wrong?? I am curious about your opinion.
Attachments
Screenshot 2022-11-24 at 11.11.44.png
Screenshot 2022-11-24 at 11.11.44.png (48.6 KiB) Viewed 8413 times
Offline

Steve Fishwick

  • Posts: 1006
  • Joined: Wed Mar 11, 2015 11:35 am
  • Location: United Kingdom

Re: Broadcast safe levels

PostThu Nov 24, 2022 10:37 am

Petehikers wrote:The Avid-Filter filters RGB-Levels and as you told me Resolve does it automatically - so going from resolve or AVID to AME for my final delivery should be fine, but going from FCPX to AME should not work. Please correct me if I am wrong?? I am curious about your opinion.


We still use Avid Symphony for final output, Peter, even if it is graded in Resolve. Mainly because of AS-11, necessary in the UK, though I'm not sure this is the case in the rest of Europe. Resolve now has the Mainconcept plug-in and I have it but I have not seen it used yet. Resolve has a legal/full level setting but that is not necessarily the same thing as a software limiter. The limiter in Avid has settings for luminance and RGB. Boris Continuum plug in also has a limiter effect too, which could be used in Resolve. Any final QC tool you use must include PSE (for epilepsy) too, more commonly known as Harding Test.
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostThu Nov 24, 2022 10:46 am

PSE is about purely UK thing. I think only Japan requires it as well. Not sure who else.

AME filter may be limiting levels as well (don't remember), so no need for some convoluted workflows.
It's also possible that once you within gamut limits you may be actually within levels as well as one assures other.
This is from guy who know math:
If the resulting RGB is in-range, it is not possible for Y to be out of range, because Y = a * R + b * G + c * B. Since a, b, c are between 0 and 1, Y must be less than max(R, G, B) and greater than min(R, G, B). The converse is not true.

There is nothing about UV though.
Last edited by Andrew Kolakowski on Thu Nov 24, 2022 11:16 am, edited 1 time in total.
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostThu Nov 24, 2022 11:13 am

Ok, I exported a picture out of broadcast range and applied the AME filter 100 ire (strictest setting).
Only Luma is filtered, RGB values are illegal.
So I guess I have to go through Avid Media Composer as well.
Does it output correct files in terms of metadata for ARD_ZDF_HDF01 for example?
Or do I have to put it through AME AFTER exporting from Avid again?
Attachments
Screenshot 2022-11-24 at 12.07.37.png
Screenshot 2022-11-24 at 12.07.37.png (248.06 KiB) Viewed 8400 times
Screenshot 2022-11-24 at 12.07.46.png
Screenshot 2022-11-24 at 12.07.46.png (221.45 KiB) Viewed 8400 times
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostThu Nov 24, 2022 11:18 am

Post this sample.
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostThu Nov 24, 2022 11:29 am

Here you are - the original without filter.
Attachments
Screenshot 2022-11-24 at 12.28.58.png
Screenshot 2022-11-24 at 12.28.58.png (527.94 KiB) Viewed 8394 times
Offline

Andrew Kolakowski

  • Posts: 9207
  • Joined: Tue Sep 11, 2012 10:20 am
  • Location: Poland

Re: Broadcast safe levels

PostThu Nov 24, 2022 11:39 am

I mean bad file.

I tested AME filter and it cleaned below out of gamut errors.
I'm not convinced about your checking procedure.

Image

Works fine.

Then if you set thresholds to 0% (so try to measure against perfect signal) it fails as it's ProResHQ export which is enough to introduce small overshoots.
Last edited by Andrew Kolakowski on Thu Nov 24, 2022 11:51 am, edited 2 times in total.
Offline

Petehikers

  • Posts: 60
  • Joined: Thu Jun 10, 2021 12:56 pm
  • Real Name: Pete hikers

Re: Broadcast safe levels

PostThu Nov 24, 2022 11:50 am

What file? It was a screenshot. Do you want the original video file?
Next

Return to Post Production

Who is online

Users browsing this forum: No registered users and 33 guests