Film-Tech Forum ARCHIVE: Cinema room calibration test tone suggestions

This is topic Cinema room calibration test tone suggestions in forum Digital Cinema Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=16;t=002321

Posted by Anhtu Vu (Member # 7689) on 06-11-2015, 12:00 AM:

We have 2 cinema room that needs to be calibrated on a regular basis. One is using the CP650/QSC DCM10 and the other, a soon to be installed Qsys core 250i processor.

I use the test tone generated within the CP650 but it's a pain to use and i'm questioning its accuracy, are there any test tone generator or files i can ingest directly into the DCP server for this purpose ? thx

Posted by Carsten Kurz (Member # 5396) on 06-11-2015, 05:32 AM:

Which tools do you use for calibration? No big deal to put pink noise and sweeps into a DCP.

- Carsten

Posted by Steve Guttag (Member # 268) on 06-11-2015, 05:52 AM:

The CP650 pink noise is actually decent and trustworthy. MANY rooms have used it as "the" reference. My only fault with the PN in the CP650 is that it is quieter than previous Dolby processors (e.g CP500 or any that relied on the CAT85 generator) so rooms set with the CP650 WILL be louder than older Dolby processors. In the DCinema world, this is no longer of any consequence though.

Posted by Carsten Kurz (Member # 5396) on 06-11-2015, 06:52 AM:

This is not meant to calibrate a room in a technical sense, but it still fits the thread topic decently, so I'll link to it anyway...

http://www.freedcp.net/dcp/test/bewegte_bilder-tricks17-test_film-full_content-51-j2k.zip

It's a test-suite for image and sound, it has german, english, french comments, contains full and individual channel noise and sweeps, and some orchestral parts.

- Carsten

Posted by Carsten Kurz (Member # 5396) on 06-11-2015, 08:26 AM:

quote: Steve Guttag
My only fault with the PN in the CP650 is that it is quieter than previous Dolby processors (e.g CP500 or any that relied on the CAT85 generator) so rooms set with the CP650 WILL be louder than older Dolby processors.

Steve - is that a different issue than this one?

http://forum.qscservice.com/viewtopic.php?f=3&t=3728#p14604
(bottom post)

edit: Looks as if it is:

http://www.film-tech.com/cgi-bin/ubb/f1/t010229/p1.html

- Carsten

Posted by Harold Hallikainen (Member # 5405) on 06-11-2015, 08:57 AM:

Do the rooms need to be calibrated on a regular basis or checked on a regular basis? I don't see what can drift much in current electronics.

Harold

Posted by Sam D. Chavez (Member # 1841) on 06-11-2015, 09:34 AM:

More like verify the levels are correct as one never knows who did the job before you.

Posted by Marco Giustini (Member # 4544) on 06-11-2015, 03:43 PM:

Carsten,

No, that is just that a 650 is 'idiot proof' and will reduce the digital gain internally before the DSP so you can push ALL the EQ up and you will not clip the internal DSP. On a USL processor you can select how much dynamic you want, the Dolby is fixed.

The AP20 for example does not pad the input and as such the internal dynamic is zero. If you use any positive EQ, the internal DSP *will* clip when it receives a 0dB FS at that frequency. Dirac does not apply as it cuts only.

What Steve mentioned was that Dolby have changed the reference level over the years. Hence if you align a room using a CP500 your final levels will still read 85dBC but since the reference level is LOUDER, the film will play softer. To better explain, if you run a Dolby Digital Pink Noise print on a system aligned at 85dB on a CP500, it won't ready 85dBC in the auditorium. I believe it is approx 2dB difference (1.6?) so it should read 83.4dB. Same print, aligned on a CP650 will read 85dBC in the room.

I guess Steve says that it does not apply with DCinema anymore because you can now run reference pink noise from the server and align the LEVELS using that. That will take into account the whole chain.

Posted by Anhtu Vu (Member # 7689) on 06-11-2015, 03:55 PM:

quote: Steve Guttag
The CP650 pink noise is actually decent and trustworthy. MANY rooms have used it as "the" reference. My only fault with the PN in the CP650 is that it is quieter than previous Dolby processors

Well, this is kind of contradicting as a ''trustworthy'' PN should not vary from one processor the another. Personally, i've never trusted the test tone generated from the CP650, but then again, as a mixer i'm super picky.

Anyways, has anyone heard of this ?

https://www.sounddogs.com/suppliers/2746.htm

Posted by Carsten Kurz (Member # 5396) on 06-11-2015, 03:56 PM:

Oh well. So much for 'Fader 7'...

- Carsten

Posted by Marco Giustini (Member # 4544) on 06-11-2015, 04:41 PM:

There is nothing wrong with the 650's PN and it's very consistent between... 650's. I believe 750's will have the same level as well.

You will find that AP20s and QSC and Dolby all have slighty different pink noise. That is why using a CPL is the final level.

Posted by Carsten Kurz (Member # 5396) on 06-11-2015, 06:47 PM:

Anhtu - you choose your test tones after your tools (or measurement methods). Aside from real music and soundtrack examples, the only test-tones I ever needed were pink noise and sweeps and the occasional sine. You get them everywhere, and even with free tools like Audacity you can create them (and analyse them) to your specific needs. If you 'buy' something, you need to make sure it has the proper digital level. If you're able to verify this, you can create it yourself just as well. A DCP can be a source for 16 channels of 24Bit/96KHz test signals.

I don't think you should distrust the CP650 internal generator. Actually you can use it as a reference for other software generated patterns.

As one of your rooms does not use a CP650, a DCP is the easiest and most practical thing.

- Carsten

Posted by Anhtu Vu (Member # 7689) on 06-11-2015, 09:47 PM:

I've tested the CP750 before settling for the QSC and, unless there was something wrong with the demo, its PN level was different than the CP650.

Posted by Steve Guttag (Member # 268) on 06-12-2015, 06:07 AM:

Starting with the CP650, Dolby lowered the internal level of the pink noise generator. as such, you will turn the gains up higher for it to reach 85dBc. This is in comparison to the CAT85 PN generator card (or Component Engineering's SG-1). I'm pretty sure the CP750 has followed suit. Thus a -20db reference signal into a CP650 (or CP750) should be at 85dBC but would be lower in legacy products. So again, what is different is the Pink Noise level in the CP650 and later versus legacy products (of which, now the CP650 is also legacy). Naturally, if you always use your own external generator, you will have consistent results through all processors (like we did in the Dolby analog days).

I don't know many people that don't use a processor's internal PN for tuning. When I did a DTS/Dolby/SDDS shootout many moons ago, that is when I discovered that SDDS cheated their outputs by using an HF deficient PN such that you would tend to tune an SDDS a bit "bright."

I say it doesn't much matter now since going forward, not too many people are going to put in legacy products (or have some CP65s and some CP650s to have the reference discrepancy show up.

I do find it interesting that Dolby dropped the AES level down 10 dB (haven't verified it) on the CAT778. I've used the card an never had a level issue with it. One would think if you are going to design a product like the DCM10D and that the CP650 is a likely candidate, that you would have a CP650 setting that would apply suitable gain to the input stage. Then again, with the CAT778, that is supposed to be used by better than the average person that should understand digital clipping and not tune a room such that the EQ stage is being used as a gain stage too. There shouldn't need to be a pad in there.

The CP650 DOES have two choices on gain stages...that is there is "normal" and "noise floor optimization." Which like the JSD series protects you or not from getting into trouble. By going to noise floor optimization, you lower the noise floor by using more of the "bits." for signal. I wonder if on the CAT778 if you went to that mode if you wouldn't get the gain back that was taken away.

Also one shouldn't forget on the DCA amplifiers...the Dataport inputs have resistor pads on them such that if you are using the Dataport inputs, you are supposed to raise the inputs to maximum. Due to this, I've seen DCMs clip out before the amplifier. It is one of the things that drove me away from the DCMs and the whole Dataport thing.

You are going to have more difference nowadays on how the surrounds are set up.

Posted by Carsten Kurz (Member # 5396) on 06-12-2015, 07:19 AM:

Well, the CP650 has the difference between analog and digital outputs. Then with internal crossovers... The CP750 has only analog outputs. So it's not easy to actually verify the pn level that is generated internally. You have to trust it. Of course you can compare different CPs, but which is right and which is wrong...
As with color and brightness shooting, it is always best to have the signal come from a DCP, as it will run through exactly the same signal pipeline as final content. More and more servers offer mix/level functions now, so it get's more important to be able to rule that out as well. It's also quite easy to tap into AES signals between server and processor to get to the ground of things if needed.

- Carsten

Posted by Steve Guttag (Member # 268) on 06-12-2015, 08:01 AM:

If you have your own PN generator (analog out) you can feed the multichannel analog in on the CP750...it is a 300mV input like all other cinema processors. The compare that level to the internal PN to see if the level is different.

Posted by Carsten Kurz (Member # 5396) on 06-12-2015, 08:08 AM:

I would trust the internal generator anyway, even if I KNOW it could be 3dB off, as long as I am aiming at cinema calibration, for simple practical reasons.

If I'm after <3dB in combination of an analog pink-noise generator fed into analog inputs of a mixed-signal device, I wouldn't trust it anyway - not the generator, not the input, not the metering, not the processing in between.

Our fader is mostly set to 5.5 and adjusted as needed. What do I need 3dB absolute noise level accuracy for?

- Carsten

Posted by Philip Wittlief (Member # 1724) on 06-12-2015, 01:26 PM:

quote: Steve Guttag
The CP650 DOES have two choices on gain stages...that is there is "normal" and "noise floor optimization." Which like the JSD series protects you or not from getting into trouble. By going to noise floor optimization, you lower the noise floor by using more of the "bits." for signal. I wonder if on the CAT778 if you went to that mode if you wouldn't get the gain back that was taken away.

There is a Unity Gain setting available when using the CAT778 that keeps your level up as long as there is not any boosting going on in the CP650 that would cause digital clipping.

Posted by Anhtu Vu (Member # 7689) on 06-12-2015, 02:12 PM:

On a side note…what do you guys use for loudness reading aside from the cheap Radio Shack loudness meter ?

Posted by Marco Giustini (Member # 4544) on 06-12-2015, 02:14 PM:

quote: Steve Guttag
that is when I discovered that SDDS cheated their outputs by using an HF deficient PN such that you would tend to tune an SDDS a bit "bright."

I have an 3000 in my garage (salvaged from the skip). I remember reading the HF issue years ago here on FT. I tested the PN of my unit and it is ruler flat. Maybe they 'fixed it' with a software?

I also don't like the Dataport thing and I seriously don't understand why the level has to be padded down. I've seen DCM's clipping too.

Carsten, your point is solid. That said, for conversation's sake, to verify the levels you just need to use a Pink Noise CPL which must be in today engineer's toolbag as the tweaker used to be in yesterday engineer's. [Smile]

Posted by Carsten Kurz (Member # 5396) on 06-12-2015, 07:13 PM:

quote: Anhtu Vu
On a side note…what do you guys use for loudness reading aside from the cheap Radio Shack loudness meter ?

These are quite popular in europe:

http://www.nti-audio.com/en/home.aspx

They also have one model with a tailored cinema software:

http://www.nti-audio.com/en/solutions/audio-acoustic-installations/cinema.aspx

- Carsten

Posted by Steve Guttag (Member # 268) on 06-12-2015, 11:13 PM:

The SDDS I measured was an DFP-D2000, not the 3000.

Posted by Harold Hallikainen (Member # 5405) on 06-13-2015, 12:04 PM:

SMPTE is about to publish a pink noise standard. It precisely specifies bandwidth, RMS level, crest factor, spectral flatness, etc. The standard will include DCPs with the standard pink noise. The RMS level of the standard has been set to match the Dolby standard used in studios. An adjustment to the RMS is made (so it's a little over -20dBFS) to adjust for the wider bandwidth of the new noise and the use of an RMS meter instead of an averaging meter. You should be able to play this DCP and compare it with your internal pink noise generator. We are implementing the new algorithm in products as we revisit the DSP code. We've already got it on one OEM product.

Harold

Posted by Steve Guttag (Member # 268) on 06-13-2015, 03:46 PM:

So what about the surrounds? DCI specifies them as -20dBFS on the recording the same as all other channels. However, in set up it is normally 82dBc/side in a 5.1 system so they SUM to 85dBc for the entire surround channel. So really, surrounds are effectively capped at 103dBc at 0dBFS.

But what happens with a 7.1 mix? Do you go with 82dBc/quadrant. So when a 5.1 mix is played on that system will it suddenly go up to 85dBc/side and 88dBc if all four channels are playing at once? Is the processor supposed to know that since you are a 7.1 theatre to pad the surrounds by 3dB when in non 7.1 mode/format?

How does the JSD60 and JSD100 handle it? Do they apply any gain/reduction based on channel configuration of the format versus the system? I ask because as of current software, I know some that don't and I doubt many that are rolling their own DSP based cinema systems are thinking about that.

Posted by Marco Giustini (Member # 4544) on 06-13-2015, 06:29 PM:

Good point.

I believe any Dolby product does know, the AP20 does not. Not sure about the USL processors though.

Somehow - acoustic reasons I guess - the real world number is more 2dB rather than 3dB. At least that was the case where I tried.

Posted by Carsten Kurz (Member # 5396) on 06-13-2015, 07:38 PM:

Marco - the AP20 does 5.1/7.1 surround level adjustments. Aside from the automatic functions, you could, of course, always implement it as well as individual calibrations, as every AP20 format can have it's own calibration anyway.

The AP20 will also do this for alternative surround sources like HDMI.

- Carsten

Posted by Harold Hallikainen (Member # 5405) on 06-13-2015, 11:50 PM:

The JSD-100 drops side surrounds and rear surrounds by 3dB if you're in a 7.1 auditorium and play 5.1 content (surround content driving both side and rear surround). The JSD-60 input mixer can be set to whatever you want, but the manual suggests a 3dB drop in this circumstance (mix coefficients of 71%). There was an interesting discussion about this on the pink noise SMPTE group recently. The discussion that IF you were to fade from one speaker to another, how far down should the speakers go when you're half way (equal power in each speaker). It was suggested that if you drive the speakers with equal power and are on the centerline between the speakers, the pressures add, so you're up 6dB. But, if you are not on the centerline, powers add, so you are up 3dB. As a compromise, a lot of panners (ie, Pro Tools) drop 4.5dB when you are at the center of the pan.

Harold

Posted by Marco Giustini (Member # 4544) on 06-14-2015, 05:56 AM:

quote: Carsten Kurz
the AP20 does 5.1/7.1 surround level adjustments

Automatically, when a 5.1 format is selected on a 7.1 system? That was not the case when I tried, I had to dial a -2dB manually. I was checking with PN from the server.

Posted by Steve Guttag (Member # 268) on 06-14-2015, 10:41 AM:

See? I like how this little nuance is not so cut and dried amongst the various implementations. The QSC DCPs don't do it either but you can burn through one of your EQ presets to create it. I hear that on an upcoming firmware, they are going to have the automatic level setting for 7.1.

The problem with surrounds is people like em loud. I can't tell you how many times I hear people spout their evaluation of a movie soundtrack based on surround activity. Go over to Cinema Treasures for the Mid-Atlantic area Theatres. So if you ignore setting the levels right, then 5.1 gets a boost of 3dB and those types of people love it...forget about that is not what the guy mixing the movie thought was right.

I also think an opportunity was missed to leave mono surrounds back to the film era. With DCP it was established that 5.1 was to be the MINIMUM sound format to be supported. As such, why the 82dBC/side? That was to accommodate mono surrounds of the film era such that a split surround mix and a mono-surround movie would play identical in level. This was especially true in its origination on 70mm where format 43 used the mono surround track below 500Hz to mix with stereo surrounds on tracks 2 and 4 above 500Hz...track 6 was used for surrounds on both format 43 (5.1) and format 42 (4.2). Yes, in its inception, format 42 had stereo bass since it was often playing out of LC/RC speakers. Even Surround-EX had to bow down to the mono surround level. A proper EX decoder would take -20dBFS signal from Ls/Rs channels and only play 85dBc from the entire rear channel despite having a 6dB increase in source level. As such, if you had a mono surround theatre and played PN from "S" channel, it was 85dBC, if you had 5.1 and you played PN from both surrounds you still had 85dBc total (82dBc/side) or if you had an EX theatre, you still had 85dBc but just from the rear speakers. If you turned off one of the surround PN sources, then it would be 82dBc in EITHER of the 5.1 or EX theatres.

We don't have that legacy in DCinema...5.1 is where we are starting. So 5.1 should have been the thing everything else had to be able to be compatible with, not 4.1 or 4.0 from film.

This is the stuff about DCinema that makes me roll my eyes a bit. [Roll Eyes]

It would have been clear that it was on the sound processor's to put in the necessary pad when playing a 5.1 track in a 7.1 surround configuration. If you want to add a fancy feature above the standard, then it is incumbent upon you to make sure your product is fully backward compatible.

Posted by Marco Giustini (Member # 4544) on 06-14-2015, 02:41 PM:

Automatic 7.1 levels is essentials, but so is the ability to tweak that automation. I usually check the 5.1 levels on the CP750 when doing levels and sometimes they're not perfect but you cannot adjust the amount of padding used by the processor.

Posted by Steve Guttag (Member # 268) on 06-14-2015, 05:11 PM:

How do you measure it?

Posted by Marco Giustini (Member # 4544) on 06-14-2015, 05:30 PM:

Pink noise CPL from server.

Posted by Steve Guttag (Member # 268) on 06-14-2015, 05:40 PM:

What meter?

Posted by Marco Giustini (Member # 4544) on 06-14-2015, 05:44 PM:

ops, sorry. WinRTA.

Posted by Carsten Kurz (Member # 5396) on 06-15-2015, 04:06 PM:

quote: Marco Giustini
Automatically, when a 5.1 format is selected on a 7.1 system? That was not the case when I tried, I had to dial a -2dB manually. I was checking with PN from the server.

You need to enable this in format options. Don't know since which firmware revision it is there, it was there when we installed our AP20 two years ago.

- Carsten

Posted by Marco Giustini (Member # 4544) on 06-16-2015, 04:19 PM:

There is a "surround offset" which can be manually set.

But last time I checked, when running 5.1 levels were not automatically adjusted.

Posted by Carsten Kurz (Member # 5396) on 06-16-2015, 07:36 PM:

It's under format and decoder options (also for HDMI 5.1/6.1/7.1) surround handling. I went into it today.

A 7.1 capable CP without that function is pretty useless. But as I said, the AP20 has enough format specific config options to adjust it also on the levels page.

- Carsten

Posted by Marco Giustini (Member # 4544) on 06-17-2015, 02:55 AM:

Unless you mean the "5.1 DCP on 7.1 system", which does what it says but does not pad the level, can you post a screenshot?

I remember the similar option in the HDMI decoder option, never tested that because the lack of a 5.1/7.1 pink noise DVD/BD.

Posted by Carsten Kurz (Member # 5396) on 06-17-2015, 10:46 AM:

No, they are not 'automatically' adjusted, that is true. 5.1 and 7.1 are two separate formats on the AP20, unlike on the Dolby CP750, where 7.1 is an add-on function on top of a 5.1 setup.

If you set the '5.1 on a 7.1 setup' format option on the AP20, it shows the signal routing and gives a hint that surround levels should be padded by 2dB, and if you do this, it is automatically applied to all surround channels.

-Carsten

Posted by Marco Giustini (Member # 4544) on 06-17-2015, 03:28 PM:

Yep. Better approach than QSC but I feel it should be automatic with the possibility to tweak the level manually if needed.

ALSO, they really need to make sure the pink noise levels are NOT affected by the "surround offset"... [Frown]

Posted by Carsten Kurz (Member # 5396) on 06-18-2015, 02:38 PM:

Here is some interesting read regarding pink noise in the cinema environment:

http://www.ohl.to/about-audio/how-pink-is-pink-noise

http://www.meyersound.com/pdf/cinema_technical_papers/cinema_calibration_tech_report.pdf

- Carsten

Posted by Marco Giustini (Member # 4544) on 06-18-2015, 03:23 PM:

Oh cheers, I was actually looking for that! I remember reading it somewhere but could not remember where! [Smile]

Posted by Carsten Kurz (Member # 5396) on 06-18-2015, 03:56 PM:

SMPTE is working on an official cinema pink noise pattern and reference DCP, so, once this is done, we will have an official standard independent from specific cinema processors or external generators.

- Carsten

Posted by Marco Giustini (Member # 4544) on 06-18-2015, 04:17 PM:

It's just taken 20 years! [Smile]

Posted by Steve Guttag (Member # 268) on 06-19-2015, 06:09 AM:

We've had reference Pink Noise in both magnetic and optical form. However, the playback medium has been the limitation and said reference film has been used to just calibrate the playback device (optical or magnetic head).

So those of you that want to tune with a DCP reference...how long is this test material? Who wants to either have it looping or keep pressing play again after it runs out? And after that, to what benefit? How different is the various pink noises going to be as compared to the speakers, rooms, microphones and the various analyzers used to measure it?

The only value I could see is after you've tuned a room, to use it as a reference level source to fine tune the levels. That and also come up with an offset to the internal PN of your favorite processor.

At least with the color targets for calibrating the projector, one can pause the show and freeze the target for as long as it takes to make a stable measurement.

Posted by Carsten Kurz (Member # 5396) on 06-19-2015, 02:29 PM:

Didn't you bring up the issue with different PN ref levels in different cinema processors? There's your benefit. Not mentioning possible server-side mixdown/level options (SMPTE DCPs, etc.), so the benefit is there as there is a benefit in playing your colors from the server instead of the projector. Who cares about the size of an audio-only (blackj image) long DCP? And who cares about looping? A Doremi will do it as a simple Cinelister-Option, not even a need for a looping SPL.

Aside from the fact that THIS SMPTE reference-noise will be built into all future CPs as a logical consequence, instead of proprietary patterns used so far.

- Carsten

Posted by Marco Giustini (Member # 4544) on 06-19-2015, 03:02 PM:

it'll be a good thing but I feel the PN level issue is in the past now. The differences now are subtle and as mentioned it won't matter if you end up playing your film at 7.0 or 6.7 - not to mention that nobody will do that anyway [Smile]

Posted by Carsten Kurz (Member # 5396) on 06-19-2015, 04:21 PM:

When you look at the variations in commercial pink noise sources demonstrated in the links I posted, the variations certainly are not small.

- Carsten

Posted by Marco Giustini (Member # 4544) on 06-19-2015, 05:15 PM:

In my world I care about Dolby and Datasat and the difference is small. I do notice that when I run PN from the server.

As said, it'll be a good thing.

Posted by Carsten Kurz (Member # 5396) on 06-19-2015, 05:21 PM:

The problem starts when you do not trust the PN from your CP (which started this thread) and think you need an expensive accurate external source, and your own metering to confirm this ;-)

- Carsten

Posted by Steve Guttag (Member # 268) on 06-20-2015, 07:28 AM:

As I said...it might make the case for a final tweak of the levels at the very end and if you have a preferred sound processor, odds are you can determine what the offset is when tuning that processor (e.g. add 2dB to the level when tuning).

I don't dispute the value having reference PN to test against...it will also give all processor manufacturers a reference to test their internals against when designing product too. One would think that after the reference is made, then future designs will be made to match it. Of course, there are forces within the SMPTE and audio community in general to change how we measure and tune sound too.

Posted by Carsten Kurz (Member # 5396) on 06-20-2015, 08:26 AM:

That's right. Once we do have SMPTE DCinema PN patterns, we may soon be urged to not use it anymore because of modernized calibration methods...

- Carsten

Posted by Harold Hallikainen (Member # 5405) on 06-20-2015, 11:22 AM:

The SMPTE modern calibration group is using pink noise. They are developing a transfer function for the sound system (including auditorium) by correlating the signal driving the sound system with what the microphones pick up. I think that any wideband signal would work.

Harold
(member SMPTE groups on modern calibration, pink noise, stereo subtitles, immersive sound, and probably a few more)