Jump to content

All Activity

This stream auto-updates     

  1. Today
  2. @SME I have a Tandberg 10XD tape machine here, which we use for analog masterings sometimes. I could test if something like this somehow introduced ULF noise along with its usual tape distortion (which ist mostly noticed by its harmonics above 2khz). Running it through a tape machine is essentially the same as digitalizing an old movie (from a reel), since its audio track is stored on the same reel. I doubt that the hardware on something like the Cintel is inferior in quality, but the noise might've been present in the original track. Tape decks are often overdriven to create more distortion, which is then mixed to the original audio. Dunno how common this is in the movies, but this or something similar is done to most studio recordings
  3. My "Like" is for this part of your post. Incompetent is definitely the wrong word and wrong concept, really. Cinema mixers are generally very competent because if they aren't, they don't get new contracts. But competence really refers to the ability to fulfill a particular job function, which isn't really saying that they know everything to know about the craft. In fact, there is probably a lot about this craft that is as yet unknown and a lot that is "known" that's actually just wrong. The issue I have with some of the recent Marvel releases is dynamics crushed to the point that it's less dynamic than an old analog TV program and doing so in a way that breaks a lot of the artistic integrity of the presentation. It's one thing if a particular scene isn't as BIG as it would have been without dynamics reduction, but another entirely when the dialog in a heated argument gets quieter just because more than one person is talking at once. Or using peak limiting in a way that snuffs transient sounds out of existence. For the most part though, I think mixers (and the other sound people) do a decent job considering the ridiculous time pressure they endure and the complexity of these soundtracks. And this is especially true given the spectral balance problems resulting form X-curve calibration. It's just tragic that this quirk affects the creation of the art in the ways it does. If they had neutral systems and the cinemas were neutral too, I think it would have a huge positive impact on the quality of the overall art. It'd definitely bring us closer to the performers, for better or worse. Exactly. There's no way the DTS and Atmos tracks were each created from scratch. One is likely derivative of the other or at least derived from a similar original. So the fact that one has this noise and the other doens't is puzzling. I know you can apply de-noising to a complete mix, but I doubt anyone does that unless they are trying to restore and re-master old content or something. Maybe that's it, and I just need to think more flexibly. I think @maxmercy sees ULF noise (or really "DC noise", under 3-5 Hz) quite a bit in movie soundtracks, more often in older tracks. However keep in mind that this noise isn't really a problem except when doing BEQ because the BEQ boosts the noise along with the desired content. At least some of the time, the "noise" may simply be due to tracks with the DC noise not being filtered any more steeply than the tracks with desired ULF content. Curiously though, some soundtracks have this noise all over and others don't have it at all. It's totally hit-or-miss. ==================== As a separate response to this overall discussion, I would caution people not to read too much into the PvA data when trying to understand why different soundtracks sound different. Certainly differences in spectral balance are possible, and seemingly minor spectral balance differences (like 1 dB) can actually have a major impact on the perceived sound. Perception is very relative in terms of what's happening at different frequencies, and you can't really see what is happening, spectrally, with the individual effects by looking at the PvA. Keep in mind too that dynamics processing might be quite different between tracks, which likely explains why the PvAs are not exactly consistent in the fine details. Yes, both tracks (and the cinema track too, if it's different) are likely seeing plenty of dynamics processing, which can affect how the sound is perceived. Also, literal dynamics is only one parameter that affects perceived dynamics. Spectral balance affects (e.g. momentary shifts in broad spectral balance and "saturation" effects) can give very different impressions of dynamics even with the SPL pegged to the same number. And the consequences of these differences may all be expressed different on different systems. So the situation is way more complicated than can be depicted with a PvA or even spectrograms, though sometimes these tools reveal interesting things. They are useful tools, but they can only explain so much.
  4. ^°^°^°^ 1) Right after I posted my previous comment, I did a quick search for the LFE channel. Your response coupled with your example made things a bit clearer to me. I still don't get the need for the LFE channel separately if you're gonna have full range speaker signals. But that belongs to a different discussion. Thanks for the explanation. 2) Since the average lines only differ by 1db, but the DTS peak at 4-5 db hotter, that'd imply the Atmos track has some effects/frequencies that are hotter than the DTS at some point in the film. Either way, we lose, it seems. I have watched the film myself, and I don't have UHD capability, so I couldn't point you to specific differences. Is it possible (and easy) in SpecLab to do spectrograms/heat maps for the whole duration of the movie similar to those of BEQdesigner? That would be probably easier than looking for speficic scenes or effects. In any case, I appreciate your chiming in on the subject, as there's a lot of speculation and opinions, but not enough objective data to make a better assessment of what's going on with these tracks.
  5. The LFE channel is a separately encoded channel, that in cinema will ONLY go to the LFE system. For us at home, BOTH the low freqs from LCRS AND the LFE (played back 10dB hotter) added up go to our subs. So the LFE channel dominates any PvA graph, and the Avg graph is the average of the whole film, so between these two mixes, they will be close. In this case, only around 1dB different. But the Peak graph shows you more in the midbass and upper midbass, sometimes 4-5dB hotter in the DTS track. That is very noticeable when listening. That could be in the LFE, or the LFE summed with the LCRS. The increased midbass is possibly a reason the DTS track is more highly rated. Looking at each individual channel between the two will tell us more. To get down to brass tacks, we can graph specific effects and compare things, but that can get tedious. If you know of specific effect timestamps that have been panned on the ATMOS track compared to DTS, I can compare them. JSS
  6. Bumblebee BEQ (ATMOS 7.1 channel bed) The effects that matter get more weight. Also got rid of ULF noise in LCRS that was prevalent in the mix. Correction: LFE: 1. Gain: -7dB 2. Low Shelf: 10Hz, Slope 1, Gain 6dB 3. Low Shelf: 16Hz, Slope 2, Gain 5dB (3 filters for 15dB total) 4. Low Shelf: 32Hz, Slope 0.5, Gain 2.7dB LCRS: 1. Gain -7dB 2. Low Shelf 22Hz, Slope 1.2, Gain 5.25dB (3 filters for 15.75dB total) 3. low Shelf 44Hz, Slope 0.5, Gain 0.7dB 4. Highpass Filter, 2Hz, 6dB/oct JSS
  7. I don't understand the concept of full LCRS if the bass is going to be redirected to the sub anyways. My rough understanding of the LFE channel was that it contains bass frequencies redirected from the other channels. If you play without a sub, then the speakers play (or attempt to) its full content, but you wouldn't miss the bass from the LFE channel. That is what would be intuitive to me, anyways, but clearly I'm no expert. Back to the graph, I'm surprised how close the average linea are in both mixes. They essentially only differ in level. Raising the MV or the sub trim could (or should) close the gaps between the tracks
  8. Godzilla: King of the Monsters: ATMOS 7.1 channel bed Pre-Post: I initially put too much midbass in the first iteration a page back on this thread, and it was awful. It works better this way, with more 'heaviness' where needed. Correction: LFE Channel: 1. Gain -7dB 2. Low Shelf: 18Hz, Slope 2.5, Gain 6.5dB (3 filters for a total of 19.5dB) 3. Low Shelf: 36Hz, Slope 0.5, Gain 5dB LCRS Channels: 1. Gain: -7dB 2. Low Shelf: 16Hz, Slope 2.0, Gain 5dB (3 filters for a total of 15dB) 3. Low Shelf: 32Hz, Slope 0.5, Gain 3dB JSS
  9. Yesterday
  10. Fair enough, incompetence may have been the wrong word and unfair but I’m not sure what other word I’d use. We don’t know what sort of time or pressure these people are under to deliver these soundtracks but we have to say that some of the results are simply poor. Even for regular systems some of the results are poor with blanket filters. I just think about me putting my name on something or releasing something. It should have a certain level of quality and I feel the quality in some of these soundtracks is absolutely poor. This is a pet peeve of mine and has been for several years so I will refrain from further comment. To me the differences between different types of tracks on a particular soundtrack is minor compared to some of the more egregious issues with some of the soundtracks released.
  11. Spider-Man: Far From Home DTS vs ATMOS: Cyan is ATMOS peak, Green is DTS Peak. Red is DTS Avg, Green is ATMOS Avg. More midbass overall on the DTS track. That can make a difference that is definitely felt on a capable system. If mixers use a cinema dub stage for one mix, and an 'HT' stage for a nearfield mix, it could account for these differences. The monitoring equipment used can lead to different curves that we measure. But one PvA for a film can be misleading. Unfortunately, the LFE channel DOMINATES any Bass PvA, as it is encoded 10dB hotter than the rest of the channels. This will require looking at the individual channels to see a difference in overall sound quality. Some of the most sonically impressive films have what appear to be a rolled of PvA with full range LCRS and only a rolled off LFE. They sound good. Like the first Iron Man, and the Lord of the Rings FOTR and ROTK. I saw the film in both DTS and ATMOS w/ BEQ. I liked both presentations. But I may BEQ the DTS to see which I like more. More on the individual channels later. It may be telling. JSS
  12. @SME it is very likely that most audio material in a movie ran through a de-noising process at some point. It might not bother you if a single recorded track has some noise (because the S/N will still be excellent), but when you start adding other tracks with the same amount of noise the noise just adds up... iZotope RX7 is the most widely used audio repair program and it's very good at what it's doing. It's so omnipresent in the movie business that it's basically always the first thing that is being done to an audio track before it's even imported into Premiere/Final Cut/Resolve. The target it usually broadband noise, so the ULF part is automatically reduced along the way. If you have some special kind of noises like steps, you can also filter that out selectively. When composers write a score employing a large orchestra, there is not always budget or time or just the need to hire hundreds of musicians to play the entire score for you. Often the score will be made from sample libraries (or even a mix of both). During a complex (rather: large) piece, there are (literally) thousands of audio samples playing at the same time. Imagine all those samples with background noise.. It is also pretty rare to see movies with a constant noise in the sub 20Hz region throught the entire length of it. (U)LF noise can come from mics not properly shock mounted. If you place a mic on a stand without rubber feet and place that on a wooden floor, every step will generate some LF noise. Worse in a car etc. Also: Software is not always 100% reliable and might do some wierd things. It has happened to me alot... We might discuss these kind of things for days, but we will still never find out why certain things happen.
  13. This is fair. I too don't think it's incompetence. I just simply don't understand the decisions they make with different versions of the same audio track. How hard would it be for them to measure the "original" mix (whatever "original" means in that context), and stick to it as closely as possible? Why the aforementioned Australian Flatliners Auro3D mix has more >20hz content than the US track? Why Far From Home has more mid-bass than its Atmos counterpart? I mean...I don't get it. It's not like they're "correcting" something wrong they found in the mix. They just do things differently. Maybe those teams are not being supervised enough, and whatever changes they made (consciously or not) nobody is noticing or don't care. I wish Sony and Disney did what Warner and other companies do and include the same Atmos track through all their releases, so everyone has the same thing. And sometimes Warner includes a 5.1DTS-MA as well, which some people claim are a little different from the Atmos, but at least you generally have both options on the same disc.
  14. Dolby Digital is perfectly capable of encoding bass all the way down to 3Hz or so. Look at some of the early 00's DVD releases that had DD soundtracks. Some of them hit pretty low.
  15. I think it's really insulting and dismissive to automatically claim that the engineers that work on these releases are incompetent. I'm not saying that there aren't incompetent people in audio, as there most certainly are. But in many cases these guys are doing the job that was given to them. Look, as much as we might want to think otherwise, people with capable home theaters are NOT the vast majority of the buying public. HTIB and sound bars in a noisy living room rule the market. So they squash the heck out of the dynamics so those people don't complain about having to constantly adjust the volume. People with capable home theaters can't compete with those numbers, and to be brutally realistic we shouldn't have any expectation that they will. Right or wrong that's just the way it works. Don't misunderstand, I wish things were different too. I wish they didn't filter sub 20Hz material just so they can overcrank the midbass (looking at you Blade Runner 2049). But we have to be realistic at some point. Complain about the quality of the soundtrack all you want, but don't assume that it's automatically because of incompetent engineers.
  16. My wild guess is that the "noise" was contributed by a mastering plugin, maybe a subharmonic synth which was different or configured differently for the DTS vs. Atmos mix. Let me turn your question around. If the noise was actually deleted from the Atmos track rather than merely filtered, how and why was this done? One possible answer to the "how" is a de-noising plugin, but my understanding is that such plugins are not fire-and-forget and the results need to be actively monitored and settings tweaked to get the best results, which these engineers probably could not do for the lowest frequencies. Maybe they just ran it blindly? In which case, I'd expect more destruction of non-noise content in the de-noised version. But why? Why would they apply a de-noising plugin just for stuff at the very very bottom of the spectrum? If one is worry about that noise "causing problems" in playback devices, a simple high pass filter will do just fine to prevent that, and in fact both tracks are HPFed already. I guess one possibility is that de-noising was applied to the entire track, and the removal of extreme LF noise happened as a consequence. Still, even with carefully hand-tuned settings, this de-noising is likely to collaterally damage some of the original content, and I don't see how it would help the soundtrack in any way, except maybe in some of the dialog recordings which can and do come with unwanted hiss sometimes. Hmm, maybe the de-noising *is* selectively applied to dialog samples with hiss, and maybe the ULF noise is part of *those* samples. (I really doubt that though being that almost all dialog probably gets high-passed at 60-120 Hz or so.) I know that on the video side, de-noising is pretty standard for home release (and probably cinema too) where it's used to try to scrub out film grain, and in this application, it also tends to degrade the actual content and can contribute new artifacts also. So maybe the Atmos track has been de-noised, in which case it's possible that the DTS version retains more content and has fewer artifacts, making it the "better track" if one is not bothered by the noise. Either way, the fact that such noise appears on one track and not the other poses interesting questions. Assessment of sound quality via purely objective means is probably very difficult if not impossible. That's a big reason why humans are involved in the mix process, and also why it's essential for the humans doing the work to listen with a monitoring system that's as neutral and accurate as possible. It may be possible for a machine learning algorithm, trained using listener preferences, to provide an assessment of sound quality. However, I would not expect this to work consistently well for a number of reasons. It might also be possible to write an algorithm specifically to detect loss of information due to lossy encoding at reduced bit-rate, but this is a very specific case. And furthermore, quantifying how much information was lost does not tell us the impact of that loss on sound quality. The subjective sound quality impact of these losses could be estimated using the same psychoacoustic models used for the lossy encoding itself, but this is still just educated guessing and for what purpose? Most media can be obtained in a lossless format or at least a lossy format with a high enough bit-rate that the subjective impact is going to be very subtle. I did recently see evidence that some, maybe all Dolby AC3 (i.e. "Dolby Digital") encoded tracks are low-pass filtered at 20 kHz, and for a lot of listeners and systems, this could have as much or more impact on the sound quality as the lossy encoding. So I guess LPFing at the top is one thing that can be objectively assessed.
  17. I think this is...pretty much it. It really is the most plausible explanation without splitting hairs. Each mixing/authoring team does with the mix what they're allowed to.
  18. I'm very inclined to agree with this. I'm not sure if two mixes graphing nearly the same, mean they sound the same. I don't think those graphs are any indication of "quality". In the BEQ thread at AVS a comparison was made of some audio mix transcoded at different bitrates, the lowest being 128kbps. All graphs look nearly the same. I wouldn't conclude the "quality" is the same. The experiment was done to see if we should expect a difference in bass levels and/or extension in streaming services due purely to audio bitrate. I wonder if there can be some visual representation of quality in an audio mix. How do you represent visually audio quality? Cross-referencing multiple audio graphs, perhaps, and seeing where one "falters"?
  19. Looks like proof positive that all these mixes are created by different teams, each with a slightly different idea of what they should sound like. And it's unlikely the director or lead sound engineer is involved (unless you're someone like Christopher Nolan).
  20. At this point, and after reading many discussions about it, I've given up trying to understand why there are so many differences between mixes. The Atmos thread at AVS got into a very heated discussion about it, that lead nowhere. In the BEQ thread at AVS we found out that there are differences between the Flatliners (2017) 5.1 DTS-MA and the Australian Auro3D 5.1 DTS-MA core, but there are multiple examples, some of which can be found in this very thread. Same movie, same effects, different levels. Sony can't be bothered to include a 24bit track in the BD. Different bass frequencies are boosted in some releases, in others they are not. Basically, we have no option than to roll with it.
  21. That ported design on Nick’s site is for car enthusiasts that want to get the most bass response from their sub. I agree that 29 Hz tuning is too high. There are generally two camps of home theater subwoofers enthusiasts, those that are content with 20+ Hz bass content and those that want the ability to reproduce bass below 20 Hz. With my 24” subs, obviously I’m in the under 20 Hz camp and feel there is a lot of content there. But even if you’re in the 20Hz camp, a 29 Hz tune is way too high for home theater in my opinion. You’ll be missing many explosions as there is a lot of movie content under 30 Hz. IMO, you’d be better off buying a few more subs and going with sealed. Extra amp channels is fairly cheap and those subs aren’t too expensive. Plus you‘ll gain the benefit of more even bass from more subs in the room. Use the Multiple Subwoofer Optimizer to align them and you’ll have deep, even bass throughout the room. You could always start with 2 subs and add more later if you want more output and more even bass throughout the room. That’s what I’d do and what I’d recommend. https://www.andyc.diy-audio-engineering.org/mso/html/index.html
  22. Why do you think that they would add noise to the DTS track? That's a little out there imo. It would make sense saying that they noticed the noise in the DTS tracks, which is why they removed it for the newer(?) Atmos mix
  23. Glad your wife didn't veto your purchase after what she had to "suffer" through, lol. Keep in mind that a subwoofer in a car get a lot louder than it does in a house where there's a lot more space to fill. If building a vented sub, the tuning frequency is very important because the sub won't play much content below the tuning frequency. So at 29 Hz, you'll be missing out on most of the content below. If you want to *just* get to 20 Hz, maybe aim for a tune of like 22 Hz. Also, I think a 4" diameter port is a bit small. You need enough area to avoid chuffing and compression at high output. How much you need depends on what the overall design looks like, but for a 15" woofer like that, I think you'll want at least 2 x 4" pipes. Do you want to tune lower? The trade-off is that you need to make the vent longer or the box larger, keeping the vent area the same. Working from the numbers you posted, a pair of 4" x 20" pipes ought to get you a tune around 22.5 Hz. Actually, it'll be lower for that particular cabinet if the pipe exit is near the wall because the wall effectively extends it. Alternatively, you can make the cabinet larger, and this will also help boost the output around the tuning frequency. You may want to look at simulations though to see what you're ultimately likely to end up with. If you do decide to use one or more 4" pipes, this product is real hard to beat if it works for your cabinet design. The flares help a lot with improving performance at high output. They are also very easy to install. If you decide to use them, I recommend using the formula included in the manual to calculate the length because it takes into account the flares properly. And also keep in mind again that if the exit is near the wall, it'll tune lower than expected. (Don't let them exit less than ~3-4" from the back to you don't constrict the flow there too much.) I think the XLS 1002 amp is a good choice as it includes DSP which you need for a vented design to apply a high pass filter to protect the woofer from frequencies it can't play.
  24. I'll just point out that just because they graph nearly the same doesn't mean they'll sound the same. Even changes that appear negligible in the graphs could be audibly quite significant. People may disagree on which sounds better, not just for purely subjective reasons, but also depending on the characteristics of their system. As a question though. Do the two look nearly identical below the filter cut-off too? The reason I ask is because it's interesting that the DTS track has more ULF noise. Think about it. A filter doesn't *remove* noise like this, right? It just suppresses content at those frequencies, unless the attenuation is extreme. That suggests that the ULF noise got added to the DTS track somehow, at a point "downstream" from whatever the starting point for the Atmos track was. Could it be that the ULF noise is from a subharmonic synth in the mastering chain that was configured differently for the DTS vs. Atmos? I'm sure there are other possibilities, but most seem to suggest that the tracks are more different than the graphs suggest. Any comments on how the dynamics sound? I hope they've dialed back the aggressiveness of the loudness normalization (or whatever it is). That together with the (IMHO) lackluster sound design really detracted from a lot of the recent Marvel movies for me.
  25. The tracks graph nearly identically, with a slight edge to the DTS track in overall level with an increase in midbass level (>40Hz), only a slight difference in dynamics, but a definite increase in noise, especially ULF noise in the DTS track, all of which may be below the threshold of hearing. If I were to BEQ one of the tracks, it would likely be the ATMOS with it's lower noise floor. I can do a more detailed audio comparo looking at more things + clipping and such when I have time. I thought the film overall was pretty good, with the lately-typical Marvel great visuals, with sound that is decent, but not outstanding. I do not have overhead channels, so I cannot comment on the ATMOS experience, only the 7.1 lossless channel beds. JSS
  26. ^°^°^°^I Interesting. So all the multiple reports the Atmos track was botched and the DTS-MA was the better track are perhaps a bit overblown? The Atmos mix is getting very dividing opinions, but most agree the DTS is the clear winner.
  1. Load more activity
  • Create New...