[bigshot]

A lot of television animation is made without directors at all. Traditionally, animation directors were responsible for supervising and doing the timing and layouts. Now they have 'show runners", who are basically creative executives. No one does timing and layout. The storyboard artist is responsible for doing all the posing and the animatic editor "parallel parks" the timing to find something that "works". But the synchronization is always much more primitive doing it this way. This is how mediocre animation is made.

Yes, the best CGI is the stuff that adheres to the fundamentals of animation... clear staging, rhythmic movement, caricature and exaggeration, etc. And the least effective are the ones that look at animation as a special effect recreating reality. I've produced a little mocap. We had to go back and completely rework it, animating it pretty much from scratch.

In the past, the action and music were planned at the same time. Once the storyboard was complete, the director and music director worked together to plan the timing on musical bar sheets. Once that was finalized, the composer wrote the score and the director did the layouts and cut exposure sheets for the animators. When the animation was completed, the composer would conduct the music to a click track and it would precisely synchronize with the action on the screen, because it all followed the bar sheets. I'm working on a project right now where we are reverse engineering Chuck Jones's timing to document his production process. If you're interested, PM me and I'll show you what we are doing.

 

[Davesrose]

As another animation producer, you're speaking to the choir about how important storyboards and animatics are. The Polar Express example is proof of how to waste money and get a wooden performance. But music hasn't been the only metronome for animation. Voice talent brings in certain intonation. I'm also actually having a casual Friday and watching The Absent Minded Professor. There's a scene where he's bouncing flubber: easy to spot that the flubber elements are animated, and there's no music to go by: but there is a rhythmic timing in animation and folly effects. I've enjoyed being a participant at Siggraph, and seeing workflows from the major Hollywood productions...and I've tried to educate some of my healthcare clients who have no idea about video production. I can also assure that with major productions like Pixar movies...there's still "dailies" where the director is seeing every wire frame rendering from an animator and providing input about movement and timing. I'm more specialized as a general medical animator, and sometimes it's taken me a great amount of time to educate a client about the need of a storyboard to not just drive animation, but help with the linear narration as well (let alone the overall timing). Actually....I don't think TV/digital has been all bad about natural timing. There has to be a major emphasis on important keyframes. If I bemoan the youth: when a younger animator approaches me with a portfolio...anecdotally I've tended to see that their modeling and rendering are fine....but timing can be stiff.

Thanks for your heads up about Chuck Jones. I'm a child from the 80s, so Don Bluth was my influence about good animation (have re-watched Secret of Nihm and think it really holds up as to how beautifully painted it is). There are more and more folks wanting to get into animation...so studies about the icons of film animation should have a following: are you going to post anywhere?

 

[bigshot]

I'm the president of a non-profit educational organization for animators... https://animationresources.org My analysis of Chuck Jones's timing technique is going to be published for members in February. We're putting the finishing touches on it right now. It'll be an e-book, video and an explanatory audio podcast. I don't think there's been much written about timing since John Halas's book back in the 60s (which is worth having). Timing is becoming a lost art. In some ways, as technology moves forward, technique gets forgotten and has to be rediscovered. I'm looking to find a way to mesh the rhythmic control of the past with current technology. I've been working on this for about 20 years, but it is very hard to find out information because just about everyone from the golden age is dead and no one documented the more technical aspects of their work.

 

[GearMe]

...Macys, Sears, or Stop & Shop

...what are those?

For example....

Thanks for that! I thought my Bossa Nova Ramones was bad

TBH...listening to that made me think that Peter Sellers or David Niven were going to show up any second!

 

[Davesrose]

I'm the president of a non-profit educational organization for animators... https://animationresources.org My analysis of Chuck Jones's timing technique is going to be published for members in February. We're putting the finishing touches on it right now. It'll be an e-book, video and an explanatory audio podcast. I don't think there's been much written about timing since John Halas's book back in the 60s (which is worth having). Timing is becoming a lost art. In some ways, as technology moves forward, technique gets forgotten and has to be rediscovered. I'm looking to find a way to mesh the rhythmic control of the past with current technology. I've been working on this for about 20 years, but it is very hard to find out information because just about everyone from the golden age is dead and no one documented the more technical aspects of their work.

Thanks, I'll look at your site when I have the time! I'm not as pessimistic about animation fundamentals being lost with new technology. A 3D animated feature is a different art than VFX for live movies. I saw a good documentary about Brad Bird....which showed him going through the Disney lots recounting his days as a teen working and learning with the "nine old men", and how he influences current animators. The previews might no longer be line art....but can be animated wireframes: but the critiques are the same. I think with animation features, people will always value the art and style that gives more drama.

On the flip side, being a stickler about anatomy and growing up with an interest in models and VFX....I have a different expectations with CGI in VFX. Everyone says how great the VFX were for Lord of the RIngs....which overall they were, but one scene stuck out with me in which Legolas's arm seemed to grow twice as long for a weird move. Rendering techniques are improving, and interestingly there are some computer algorithms that are getting good about believable human modeling and expression. Known as "deepfakes" there's some people with home desktops churning out better CGI human actors than decisions made with a particular team. People have different expectations with a "realistic" rendering in that it has to look photo realistic and have all the subtle movements of a live actor: animated characters can convey more drama with the over exaggeration and talent of the animator.

 

[bigshot]

Subtle isn't necessarily better. More expressive is better.

 

[Arjey]

Could someone explain to me why this works this way? So basically, I downloaded a flac 24/192, and a flac 16/44.1.
Then I converted the 24/192 to 16/44.1, and for some reason, it turned out "worse", upper frequencies cut, less information, and it weighs less than the original 16/44.1 that I downloaded.
16/44.1 original:

Vs 16/44.1 converted from 24/192:

It's even worse than an m4a(vbr) converted from the same 24/192 (which is actually almost identical to the original 16/44.1, a bit worse):

And then there is a m4a(vbr) converted from the original 16/44.1 flac, the worst of all:

So why is an m4a converted from a 24/192 flac better than a 16/44.1 flac converted from the same 24/192? Why is it much better than an m4a converted from an original 16/44.1 flac, which is better than a 44.1 flac you get from converting from a 24/192? So why does it turn out like this (from best to worst): original 16/44.1 flac -> m4a from 24/192 -> m4a from original 16/44.1 flac ≈[?] 16/44.1 flac from 24/192?
Does that mean, that if for example, I want to have flacs (16/44.1, the "best" audible quality for listening to music) on my mp3 player/computer, I need to download original (ripped from CDs straight to flac 16/44.1) for that, and I want to have the best m4a for my phone, I need to download 16/44.1 for player AND 24/192 (not necessarily 24/192, but higher than 16/44.1) to convert to m4a for phone? (I can't just download m4a files, because they are usually mastered for iTunes and sound funky)

 

[pinnahertz]

Could someone explain to me why this works this way? So basically, I downloaded a flac 24/192, and a flac 16/44.1.
Then I converted the 24/192 to 16/44.1, and for some reason, it turned out "worse", upper frequencies cut, less information, and it weighs less than the original 16/44.1 that I downloaded.
16/44.1 original:

Vs 16/44.1 converted from 24/192:

It's even worse than an m4a(vbr) converted from the same 24/192 (which is actually almost identical to the original 16/44.1, a bit worse):And then there is a m4a(vbr) converted from the original 16/44.1 flac, the worst of all:

But how do they sound in a blind comparison?

So why is an m4a converted from a 24/192 flac better than a 16/44.1 flac converted from the same 24/192? Why is it much better than an m4a converted from an original 16/44.1 flac, which is better than a 44.1 flac you get from converting from a 24/192? So why does it turn out like this (from best to worst): original 16/44.1 flac -> m4a from 24/192 -> m4a from original 16/44.1 flac ≈[?] 16/44.1 flac from 24/192?

There are different processes involved, and without knowing what specific software and settings are being used, it can only be assumed that different processing is being applied to each.

For example, resampling from a high rate to a lower one will reduce high frequency content above and around the new Nyquist frequency. That essentially involves a filtering process, among other things. FLAC is lossless, though, so if no resampling is required, then no additional filtering is applied. However, resampling to a lower sample rate is not lossless, and the filter applied may vary depending on the resampling method used. In some cases it may be possible that resampling and the applied filter may be user adjustable, but good resampling is not a given, there's good, bad, and ugly.

Reencoding to m4a or mp3 applies more filtering, sometimes fixed, sometimes dynamic, as part of the total bit-rate reduction of the codec. You'll find mostly mp3 coding takes high frequencies above 15kHz off, for example. But given a high enough bit-rate setting, the results of m4a coding are generally inaudible. So again, how does it sound? Lossy codecs end results are subjective, which means judgement doesn't involve strictly measurement. A good lossy coded, which m4a is, can have no audible impact if appropriate settings are chosen. Listen with your ears, not your eyes.

Does that mean, that if for example, I want to have flacs (16/44.1, the "best" audible quality for listening to music) on my mp3 player/computer, I need to download original (ripped from CDs straight to flac 16/44.1) for that, and I want to have the best m4a for my phone, I need to download 16/44.1 for player AND 24/192 (not necessarily 24/192, but higher than 16/44.1) to convert to m4a for phone? (I can't just download m4a files, because they are usually mastered for iTunes and sound funky)

You are equating "quality" with visible high frequency in a spectrogram. Spectrograms are not hearing, and tend to exaggerate inaudible differences which people erroneously equate to quality. There may or may not be a correlation.

 

[Arjey]

But how do they sound in a blind comparison?
There are different processes involved, and without knowing what specific software and settings are being used, it can only be assumed that different processing is being applied to each.

For example, resampling from a high rate to a lower one will reduce high frequency content above and around the new Nyquist frequency. That essentially involves a filtering process, among other things. FLAC is lossless, though, so if no resampling is required, then no additional filtering is applied. However, resampling to a lower sample rate is not lossless, and the filter applied may vary depending on the resampling method used. In some cases it may be possible that resampling and the applied filter may be user adjustable, but good resampling is not a given, there's good, bad, and ugly.

Reencoding to m4a or mp3 applies more filtering, sometimes fixed, sometimes dynamic, as part of the total bit-rate reduction of the codec. You'll find mostly mp3 coding takes high frequencies above 15kHz off, for example. But given a high enough bit-rate setting, the results of m4a coding are generally inaudible. So again, how does it sound? Lossy codecs end results are subjective, which means judgement doesn't involve strictly measurement. A good lossy coded, which m4a is, can have no audible impact if appropriate settings are chosen. Listen with your ears, not your eyes.

You are equating "quality" with visible high frequency in a spectrogram. Spectrograms are not hearing, and tend to exaggerate inaudible differences which people erroneously equate to quality. There may or may not be a correlation.

I know that more high frequencies doesn't mean better quality, I just don't want to lose the airyness of some songs, so I want to keep as much information as possible. Regarding what I hear, I don't hear any difference between the original 16/44.1 flac and m4a from 24/192, but they do sound different compared to m4a from original 16/44.1 flac and 16/44.1 flac from 24/192.
I guess I'll just keep experimenting. Does anyone know a good free audio converter (I'm interested in converting flac (16 and 24bit) to m4a and flac 16bit). Maybe the problem is in my converter..

 

[castleofargh]

Could someone explain to me why this works this way? So basically, I downloaded a flac 24/192, and a flac 16/44.1.
Then I converted the 24/192 to 16/44.1, and for some reason, it turned out "worse", upper frequencies cut, less information, and it weighs less than the original 16/44.1 that I downloaded.
16/44.1 original:Vs 16/44.1 converted from 24/192:
It's even worse than an m4a(vbr) converted from the same 24/192 (which is actually almost identical to the original 16/44.1, a bit worse):And then there is a m4a(vbr) converted from the original 16/44.1 flac, the worst of all:So why is an m4a converted from a 24/192 flac better than a 16/44.1 flac converted from the same 24/192? Why is it much better than an m4a converted from an original 16/44.1 flac, which is better than a 44.1 flac you get from converting from a 24/192? So why does it turn out like this (from best to worst): original 16/44.1 flac -> m4a from 24/192 -> m4a from original 16/44.1 flac ≈[?] 16/44.1 flac from 24/192?
Does that mean, that if for example, I want to have flacs (16/44.1, the "best" audible quality for listening to music) on my mp3 player/computer, I need to download original (ripped from CDs straight to flac 16/44.1) for that, and I want to have the best m4a for my phone, I need to download 16/44.1 for player AND 24/192 (not necessarily 24/192, but higher than 16/44.1) to convert to m4a for phone? (I can't just download m4a files, because they are usually mastered for iTunes and sound funky)

First we need to get one important notion cleared out. The strong blue on those images are for signals at -90dB!!!!!!!!!!!!!!!!!!! I don't even go that low when setting the spectrogram range in audicity.
Edit:
Then this specific track or sample of track doesn't seem to show any signal before about -25dB and that's perhaps the weirdest thing to see here. Unless you specifically picked a quiet passage in the track, it doesn't make much sense for any digital track to just waste that headroom.
(edit: this is wrong, I fall for that often and never seem to learn, sorry).


Now about what you saw going wrong/strange. I would guess that the overall gain is down on the second pic, and most of the mystery is going to be for you to find why.
Somehow the downsampling seems to apply a filter that starts at 18kHz. In the digital domain, there is no real need for that IMO. TBH it would make more sense to me if this was the output of a DAC instead of a converted file.

 

[Arjey]

First we need to get one important notion cleared out. The strong blue on those images are for signals at -90dB!!!!!!!!!!!!!!!!!!! I don't even go that low when setting the spectrogram range in audicity.
Then this specific track or sample of track doesn't seem to show any signal before about -25dB and that's perhaps the weirdest thing to see here. Unless you specifically picked a quiet passage in the track, it doesn't make much sense for any digital track to just waste that headroom.

Now about what you saw going wrong/strange. I would guess that the overall gain is down on the second pic, and most of the mystery is going to be for you to find why.
Somehow the downsampling seems to apply a filter that starts at 18kHz. In the digital domain, there is no real need for that IMO. TBH it would make more sense to me if this was the output of a DAC instead of a converted file.

I see what your saying about the "blue parts", I know that I can't really hear stuff that are so quiet. But it does change what I hear. I mean, the fact that the driver of my earphone is trying to make that sound does affect the overall sound, if you understand what I mean. If you record a pluck of a guitar, and then cut out all sounds under -70 it will sound very different to the unedited version. So I just want to keep as much info as possible.

The color interpretation is different from Spek (the PC app I use), so take that into account when looking at dB. In Spek, for examle, bright green is ≈-65, here it's -50. The song it's quite. But the app is mobile, so it may be quite a bit wrong interpretating volume. My PC monitor broke, can't use Spek(

 

[bigshot]

Do a blind listening comparison of the files. I bet blind they all sound the same. I've never been able to judge sound quality by charts. I can only spot obvious problems. Better to judge sound by listening.

 

[castleofargh]

I see what your saying about the "blue parts", I know that I can't really hear stuff that are so quiet. But it does change what I hear. I mean, the fact that the driver of my earphone is trying to make that sound does affect the overall sound, if you understand what I mean. If you record a pluck of a guitar, and then cut out all sounds under -70 it will sound very different to the unedited version. So I just want to keep as much info as possible.

The color interpretation is different from Spek (the PC app I use), so take that into account when looking at dB. In Spek, for examle, bright green is ≈-65, here it's -50. The song it's quite. But the app is mobile, so it may be quite a bit wrong interpretating volume. My PC monitor broke, can't use Spek(

The color code doesn't matter, it's provided on the side of your graphs.
I'm not judging you or telling you to like having stuff low passed at 18kHz for no reason. You obviously should try to get what you wish to get and most people get, which is something low passed harder but also at a frequency closer to samplerate/2. What most converters will do TBH, you just got unlucky with yours or some of the settings in it.

About extra stuff and their impact: 1. If the extra content is at a level and a frequency that is audible to you(not completely masked by louder sounds, and not outside of frequencies you can still perceive), then you hear it and it affects your overall impression of sound. How much depending on how loud and what sound it is.
2. The signal itself is not audible but the headphone is terribly nonlinear with horrible amounts of distortions. Then maybe you'd get to hear a change from the garbage created by the extra yet inaudible signal changing the driver's movements. That I think is what you're describing and it should not have audible consequences as for most headphones an already quiet signal would only generate extra content as distortions at yet another 40 or 50dB below the quiet sound. Reaching inaudible level in most circumstances with even rather average headphones.
3. But in case you're speaking about a concept "à la" analogsurviver and his ultrasonic content being not audible by themselves but being audible as part of a bigger global sound element of sort(I don't think you meant that but I'm just trying to be exhaustive). That is wrong. Wave theory says so, and if things were that "simple", it would be trivial for anybody to pass a blind test between 44.1 and higher sample rates. Or maybe between a track and the same track plus noise at -90dB. Except that such tests are rarely successful and pretty much never with reasonably mastered test track and reasonable listening levels. Meaning that both theory and practical experiments disprove the idea.

In any case, I insist on being with you when you want to keep as much info as possible. I see nothing wrong with that desire. You should try some other converter for your 44.1 flac as something is wrong with the result anyway. I've been using SOX for many years now so I don't know anything else and hope others will have suggestions for something maybe more intuitive. Because SOX is cool and very customizable, but the old school command line thingy isn't super attractive or intuitive at first contact.

Do a blind listening comparison of the files. I bet blind they all sound the same. I've never been able to judge sound quality by charts. I can only spot obvious problems. Better to judge sound by listening.

It's almost impossible to see anything the way the graphs show up in the post, but if you open the first 2 graph in 2 tabs and switch between them, the differences are hard to miss. The 16/44flac version is clearly suspicious and seems to be quieter by a good margin, plus the signal seems low passed near 18kHz. which is strange for a digital conversion to 44.1kHz. So it probably does not sound the same as the original if only because of the loudness difference(or whatever it is I mistook for gain change).

 

[bigshot]

Yeah, it's easy to tell that you can't tell.

 

[71 dB]

First we need to get one important notion cleared out. The strong blue on those images are for signals at -90dB!!!!!!!!!!!!!!!!!!! I don't even go that low when setting the spectrogram range in audicity.
Then this specific track or sample of track doesn't seem to show any signal before about -25dB and that's perhaps the weirdest thing to see here. Unless you specifically picked a quiet passage in the track, it doesn't make much sense for any digital track to just waste that headroom.

It's a spectrogram so the levels indicate spectral density. The more detailed spectrogram (bigger FFT), the more frequency points and the lower the spectral density, because the signal energy is divided to more frequency points. that's why "yellow" (or about -30 dB) is the loudest we see. The blue is down at around -90 dB, but you need to integrate all those blue points on the frequency line to get the signal level at say 15 kHz to 20 kHz. So if the FFT size is say 1024, the amount of positive frequency points is 512 meaning the frequency range 15-20 kHz has 512 * (20-15) kHz/22.05 kHz = 116 frequency points. If all of these are at spectral density level -90 dB, the signal level at frequency band 15-20 kHz is -90 dB + 10*log10 (116) = -69 dB. That's still very low level considering how insensitive hearing is at these frequencies! The frequency line here is linear (far from how human hearing works) and makes the blue stuff more dramatic it is for human ear. The upper half (11-22 kHz) is just one octave and on logarithmic scale ~1/10 of the picture (if 20 Hz - 22050 Hz plotted)!