Validity of X-curve For Cinema Sound

[Andysu]

I like to have a rant against dolby labs, rubbish dsu in the afternoon.

 

[Frontino]

You are correct! When I started doing made for HT mixes in 2000, Dolby recommended we do them without using the X curve. I think the X-curve should have been abandoned decades ago, it is a relic of the past.

Is the X curve an actual first arrival response or a target on RTA to achieve a flat response?

 

[audio2920]

Dolby measure it using RTA / pink. And yes the goal is flat response from the speakers. (I.e. In a lot of cases, once a speaker system is calibrated to be "flat" in a large room, it'll be fairly close to the x-curve, but there are a ton of variables that can throw the correlation out if you're unlucky, obviously)

 

[Frontino]

I asked that question because I realized that calibrating Dolby's noise at 85 dBC with the standard X curve actually gives ~106 dB peak in the midband (~109 true peak) instead of ~105 dB without X curve.

So I'm in doubt if I should compensate to align the midband, and thus the entire bandwidth, to 86 dBC or keep it at 85.

 

[Frontino]

^ where is pictures of your room and speaker

I can't post pictures because I won't be home until Christmas, but I won't deny that it's very small. It's basically a near field setup with a Klipsch THX Ultra 2 system.

The last feature I played at 85 was The Green Knight and it was amazing. Indeed, when I saw Skywalker Sound in the ending credits, I understood why it sounded so good.

 

[audio2920]

academy award best achievement in sound . i believe they win if mixed at certain dB x curve scale level ? that is what read around in white paper years ago .

i can test this many ways which i not seen a soundmixer filmmixer even providing a single video ? wow they must be really that important then ?

Whatever......

So I'm in doubt if I should compensate to align the midband, and thus the entire bandwidth, to 86 dBC or keep it at 85.

I wouldn't obsess over this too much; there are indeed several ways of looking at it. But, assuming you're happy with the general sound/FR you have, I'd personally just knock the level down a tiny bit because you're likely in a room where the direct:indirect ratio is a bit higher than in a large cinema...?

I'd keep in mind a couple of things though:

(A) the X-cruve/RTA method was really just a quick and dirty way for cinemas to get their replay systems "flat" in an analogue world. You shouldn't really need it in a small room, IMHO.

(B) something I always go on about; the loudness standards (or lack thereof) used for consumer distribution formats mean the mix may or may not be designed to play at "85" anyway. It is more than possible that, unless you've got an actual cinema exhibition version of the film, the recorded level of the mix was not referenced to 85, so it may be blisteringly loud at 85.

 

[Frontino]

I don't use the X curve or a modified one or any sort of EQ. I just play Dolby's noise filtered from 500-2K adjusted to -20, encoded in both Dolby TrueHD and DTS MA, and aim it to 85 dBC Leq (82 for surr.) with a SPL meter put on a tripod.

I exclusively play back movies at THE cinema reference level. I tried lowering the amp, but I always felt I was loosing the impact of peak levels that only true reference gives.
If a soundtrack has dialogues that were obviously mixed below reference (hot), I simply keep listening until the movie ends, then I open the tray and throw the disc in the trash can.

 

[audio2920]

Ah ok! So are you saying that because of the roll off in X-curve the midband level is actually louder to make up for the reduced level at HF and LF?

I think the thing is, the X-curve doesn't deliberately reduce HF content. For sure, it's only an approximation, so there's definitely scope to be wrong and cause incorrect negative EQ at HF, and thus require more global gain to hit 85 on a meter. But not intentionally.

I tried lowering the amp, but I always felt I was loosing the impact of peak levels that only true reference gives.

Well, that's great. I mean, if that's the level you're happy with then job done, right? I agree that there's excitement about listening loud that just disappears when you reduce it by even a dB or two. (Probably distortion and hearing damage, hahaha)

If a soundtrack has dialogues that were obviously mixed below reference (hot), I simply keep listening until the movie ends, then I open the tray and throw the disc in the trash can.

Very few discs I own are at 85. Although to be honest I don't watch tons at home these days. Do you throw away a lot?

 

[Frontino]

Not a lot. I try to be very selective with my spending, but sometimes a dud slips through.

 

[Frontino]

Very few discs I own are at 85.

Do you think Tenet is one of those, untouched, direct port of the theatrical mix? If you're not sure, can you get in touch with someone who can confirm that? I'd appreciate it very much.

 

[audio2920]

Don't know I'm afraid. I'll see if I can find out.

IIRC Warners always ask for a -24LKFS nearfield mix to go on their video masters. But I don't know if they used it on that particular disc or if it will have just gone for VOD.

If you've got time to kill, and a bluray drive in a computer, you could rip the audio and measure the loudness? If it comes in at -24LKFS +/-2 and has truepeaks around -2dBTP (may not be exact if a lossy codec) then it's NOT the mix used for cinema.

 

[Frontino]

Then it's definitely not a nearfield remix, because I remember a value more like -14 or -11 LKFS and true peak above 0 dBFS sine wave.

 

[audio2920]

Cool. Then, yeah, seems extremely unlikely to be a nearfield mix to me.

If it was a mid-1990s to early-2000s film's "nearfield" I wouldn't be so confident to identify it that way, as people didn't used to worry about interpolated/oversampled peaks so much back then. Nor was there much in the way of loudness metering. But IMHO these days it would be pretty off-piste to do a nearfield while not reducing peaks to at least -1dBTP, nor bringing the LKFS integrated level to somewhere in the -20s. (There's always exceptions, of course)

 

[Frontino]

Well, now I'm not so sure of it anymore because I"m reading on places like gearspace that some nearfield mixes might be about same loudness, to compansate for reduced playback level.

Do, please, get back if you can find out if Tenet on BD and UHD BD has the theatrical mix.

 

[Chromatischism]

For what it's worth, when we check our mixes in a more mid-field environment for the Home Entertainment/VOD deliverables it's fairly common to check this without X-curve compliance. Even then I personally sometimes add a bit of sparkle if it's available in the recordings*. If we'd mixed in the big rooms without X-curve, this would be required even more. I guess what I'm saying is that even though X-curve is basically "wrong" in terms of standardisation, given it generally seems wrong in the direction that it rolls off too much HF (depending on how you actually quantify that) it is actually giving current mixes more HF [in the recording, ignore the replay chain] than any other calibration curve.

My issue with this is that if you look at the majority of cheaper, "home theater speakers" that people install, they have elevated high frequencies for that "clarity" factor. People like me go out of their way to find flat, neutral speakers, but very few consumer systems are actually flat. I know, circle of confusion.

 

[audio2920]

My issue with this is that if you look at the majority of cheaper, "home theater speakers" that people install, they have elevated high frequencies for that "clarity" factor. People like me go out of their way to find flat, neutral speakers, but very few consumer systems are actually flat. I know, circle of confusion.

Totally agree. My comment was in the context of dialogue intelligibility issues though, and I don't believe the replay side having more HF than the mix stage (particularly in the top octave and a bit) would have negative impact on that particular aspect. Obviously it's not an ideal situation though!

 

[Viktor]

x-curve disable?

Hi to all members,
I have JBL8350 and i find them very roll off in HF when i use them in stereo setup for music,i found crossover schematics of jbl 8340 where the x-curve you can turn off ,here are screen shot of crossover schematics of both,can someone tell me how can i modified the crossover in jbl 8350 to disable the x-curve ? S1 is the switch where you can turn off x-curve in 8340 model.
Thanks
Viktor

 

[Frontino]

Obviously it's not an ideal situation though!

I was wondering, when calibrating a theater or dub stage, is the RTA set to C-weighting or Z (flat)?
Because a flat responding loudspeaker measured on a C-weighted RTA would actually show something already looking like the X Curve.

 

[Frontino]

Don't know I'm afraid. I'll see if I can find out.

IIRC Warners always ask for a -24LKFS nearfield mix to go on their video masters. But I don't know if they used it on that particular disc or if it will have just gone for VOD.

If you've got time to kill, and a bluray drive in a computer, you could rip the audio and measure the loudness? If it comes in at -24LKFS +/-2 and has truepeaks around -2dBTP (may not be exact if a lossy codec) then it's NOT the mix used for cinema.

So, early this year, I rescanned the English DTS-HD MA of Tenet and it's indeed around -11 LUFS.

Now, I'm back in Germany for work and was checking some pre-Dunkirk UHDs of Nolan's on caps-a-holic and realized that the French and German DTS-HD MA tracks have lower bitrates compared to English (~2 Mbps vs 4).

I won't be able to scan those in Audition 'til Christmas time, so could any of you compare them in a listening test or speculate on the reasons for this?

Maybe they put together nearfield and farfield mixes on the same disc?

Thanks.

 

[Andysu]

Dolby Stereo in selected THX cinema , x curve
when the muskets fire , feels like an air buster THX earthquake in my THX cinema