4 ms·

> The groups argued that “many” streaming services were already trying to manage the “loudness of advertisements that come from server-side ad insertion that ma

> The groups argued that “many” streaming services were already trying to manage the “loudness of advertisements that come from server-side ad insertion that may be inconsistent with the loudness of the programs,” […]. Server-side ads can have differing volumes due to companies using various encoding pipelines. Boo-freaking-hoo. Cry me a river, poor streaming services without the technical know-how to calculate an ad’s volume. We can’t expect them to know how audio works! > Additionally, as the opposing groups previously pointed out, streaming services must contend with a broad range of output devices, including TVs, tablets, and phones. See, that’s just flat-out lying. What’s this mythical circumstance where playing audio A at the same volume as audio B on one device will magically make A louder than Bon another? Especially when dealing with server-side ad insertion, as the article discusses, where the service has full control of the input files and the output stream? This reads like a restaurant trade group claiming that it’s impossible to know how much salt they put in the gravy.

chimeracoder 4mo ago
> See, that’s just flat-out lying. What’s this mythical circumstance where playing audio A at the same volume as audio B on one device will magically make A louder than Bon another? Regarding your second point: as any audio engineer or electronic musician knows, the same exact audio absolutely will sound very different on different speakers, depending on how well they replicate various sounds, what level of gain is being applied, and the volume (which is different from gain, although people confuse the two). That's even before you get into the fact that many modern devices, like smartphones, will apply their own compression or sound processing before playing the sound, sometimes to compensate for those deficiencies and make them less noticeable, and sometimes to "enhance" the sound. Loudness/volume (technically different things but let's conflate them here) are also unintuitive because human ears don't have a flat frequency response curve, and some things will be perceived as louder despite being the same volume, or vice versa. Advertisers actually can (and do) take advantage of this, by using sound engineering to make things feel louder while staying within the desired volume, by targeting the way humans perceive the sound. This isn't a defense of the advertising/streaming companies here, because it is a solvable problem. But it is true that this is a problem that they need to solve.
- kstrauser 4mo ago
  All that’s true, but those factors affect all the audio similarly. The article specifically talks about server-side ad insertion, so it’s not like the case where it somehow uses the device’s .mov codec to play the content and an MP3 codec to play the ad. All ffmpeg (most likely) knows is that it’s decoding one long stream, and doesn’t switch audio pipelines mid-stream when it thinks it might be playing an ad at that moment. Regarding the perceptual volume differences: while true, that’s also a solvable problem. Output volumes can be calculated using standard curves. In any case, TV broadcasters have had to figure all this out years ago.
  - radley 4mo ago
    > those factors affect all the audio similarly... Output volumes can be calculated using standard curves... TV broadcasters have had to figure all this out years ago. Sorry, but all of that is obtuse. The fact that some digital audio can be perceived as much louder than others –– yet it's all limited to the same digital range –– proves they aren't similar at all. There is no such thing as a standard curve for compression. Source levels vary almost infinitely. Accurately separating and reducing sound after the fact, without turning the whole thing to mud, is considered to be an impossible technical challenge. Next, TV broadcasters worked on a predetermined schedule with predetermined advertising. This gave them time to inspect and approve ads in advance. Streaming ads are generally served just in time from third-party services to the streaming host. FFMPEG gets the output from the stream host, but the host has to combine content together from multiple sources (entertainment + multiple ad servers) into that single stream. Currently, sound-level is completely at the whim of each ad server, as well as each ad producer. Meanwhile, the final output is at the whim of the streaming host: 24-hour-news streaming sites probably have different audio standards than Apple TV+. Ultimately, AI could potentially be used to solve it, since it can generate / make-up new sounds as part of reverse-compression. But it would still have to be done in advance by the third-party ad servers.
    - kstrauser 4mo ago
      None of this is true. There are standard curves for human hearing frequency response and you can use these to compare sound A’s volume to sound B. And since sound compression is in DCT space, you can calculate those numbers very quickly with something similar to sum(vol(f) * curve(f) for f in encoded_frequencies). I read the article. It specifically talks about server-side ad embedding, i.e. where the service is inserting ad content into the streams, and therefore, by definition, has access to the ad content. They can do the calculations on their end during the embedding process and normalize volumes there before transmitting the result. To make things even easier, they don’t have to calculate the ad volume each time one’s streamed, just once per ad they’re going to serve. And finally, all of this is a solved problem for TV broadcasters. They face the same problems: advertisers send them content to air, then the broadcasters are legally required to normalize the ad vs content volume, and they do. If this is an insurmountable problem that the streaming services face, they can drive over to their nearest TV station and ask them how they manage to pull off this technological feat.
rdedev 4mo ago
The group includes Netflix, the most technically capable streaming company. It's sad that companies will only go down kicking and screaming even for the mildest of regulations
- kstrauser 4mo ago
  Yeah. If I had to do this, there’s the likelihood I’d screw it up. I am not Netflix, who quite good at the whole streaming video thing. I find it very hard to believe there are technical challenges in this law that Netflix couldn’t possibly solve.
- pengaru 4mo ago
  It's arguably a good strategy to make mountains out of mole hills from their perspective. It drains the regulators' finite resources on relatively trivial matters, and I think it goes without saying there are less trivial and more egregious exploits continuing, more unchecked now as a result. TL;DR distract the enemy with resource sinks
avereveard 4mo ago
I hate loud ads as well as anyone else and I welcome this resolution but I would not treat the challenges regulation poses as simplisticly as this. There is a lot of research in increasing loudness without increasing decibels, especially for concerts, but it migrated to ads when tv started adding automatic volume controls to normalize across services.
pkulak 4mo ago
I think the point is that when you don’t control the ad, it’s a bit tough to normalize its audio. And controlling the ad means bringing ad serving in house, which while possible, is a huge engineering ask. I guess the solution is to switch to a proper ad insertion company that normalizes to -24 like you’re supposed to, but that’s not cake either. Especially if contracts are signed.
- Ar-Curunir 4mo ago
  Well maybe then they should have done quality control. We’ve gotta stop making excuses for these companies that are making money hand over fist.
  - dualvariable 4mo ago
    They might have to hire some people and create some jobs... Oh, the humanity...
- dinfinity 4mo ago
  I assume that they also have mechanisms to check that the content itself is legal to broadcast. Checking the loudness and rejecting based on them in that process should be trivial.
- mitthrowaway2 4mo ago
  If you're streaming the audio waveform, you can calculate the peak volume and adjust the gain.
  - radley 4mo ago
    Most streaming audio already share the same peak volume. The problem is compression. You can compress the hell out of audio, make it sound extremely loud, and it will still have the same peak volume as uncompressed audio.
    - topranks 4mo ago
tzs 4mo ago
> See, that’s just flat-out lying. What’s this mythical circumstance where playing audio A at the same volume as audio B on one device will magically make A louder than Bon another? Especially when dealing with server-side ad insertion, as the article discusses, where the service has full control of the input files and the output stream? Consider a streaming movie with surround sound with an inserted ad that is in stereo. I'm playing it on a 5.1 home theater system and you are playing it on a stereo phone. Your system is mixing the surround sound down to stereo. When your device does that it applies attenuation to the program so that if several channels in the 5.1 stream have something loud all at the same time it won't be too loud in the down mix for stereo and clip. When the commercial cuts in your device recognizes it is ordinary stereo and it doesn't need to down mix. It goes straight through without the attenuation that down mixer applies to the program. Whatever level the commercial is really at relative to the program, it is going to sound loader than that on your system because of that attenuation difference. On my device it is not attenuating the 5.1 program since it has all the necessary channels. However, if the commercial is at the same level as the program it will actually sound louder on mine. That's because the same total level of sound split among 5 speakers perceptually seems less loud than the same total level coming from stereo speakers. The streamer can do loudness normalization between the program and the commercial. It can calculate what the perceived human loudness will be at any time in the adjust the levels so that on my device the perceived level of the 5.1 program when it gets to the commercial will match the perceived level of the commercial on stereo. But for devices that are down mixing to stereo there is still going to be the attenuation the down mixers uses, and that differs from device to device. That limits what can be done server side to get the program and commercial to match. Some multichannel formats do include metadata for the device telling it how much to attenuate when down mixing to stereo. If all the device supported that it should be possible for the server to fully take care of loudness matching. Otherwise you probably need device side normalization. Another approach would be to up mix the stereo commercial server side to whatever surround sound format the program is using. Then they could do server side loudness normalization between the program and the commercial without it being messed up by the difference in how stereo devices down mix. I'm not sure why that is generally not done. LLMs are suggesting several reasons but I have no idea if they are reasonable. I'll leave exploring that to someone else.

chimeracoder 4mo ago

> See, that’s just flat-out lying. What’s this mythical circumstance where playing audio A at the same volume as audio B on one device will magically make A louder than Bon another? Regarding your second point: as any audio engineer or electronic musician knows, the same exact audio absolutely will sound very different on different speakers, depending on how well they replicate various sounds, what level of gain is being applied, and the volume (which is different from gain, although people confuse the two). That's even before you get into the fact that many modern devices, like smartphones, will apply their own compression or sound processing before playing the sound, sometimes to compensate for those deficiencies and make them less noticeable, and sometimes to "enhance" the sound. Loudness/volume (technically different things but let's conflate them here) are also unintuitive because human ears don't have a flat frequency response curve, and some things will be perceived as louder despite being the same volume, or vice versa. Advertisers actually can (and do) take advantage of this, by using sound engineering to make things feel louder while staying within the desired volume, by targeting the way humans perceive the sound. This isn't a defense of the advertising/streaming companies here, because it is a solvable problem. But it is true that this is a problem that they need to solve.

kstrauser 4mo ago
All that’s true, but those factors affect all the audio similarly. The article specifically talks about server-side ad insertion, so it’s not like the case where it somehow uses the device’s .mov codec to play the content and an MP3 codec to play the ad. All ffmpeg (most likely) knows is that it’s decoding one long stream, and doesn’t switch audio pipelines mid-stream when it thinks it might be playing an ad at that moment. Regarding the perceptual volume differences: while true, that’s also a solvable problem. Output volumes can be calculated using standard curves. In any case, TV broadcasters have had to figure all this out years ago.
- radley 4mo ago
  > those factors affect all the audio similarly... Output volumes can be calculated using standard curves... TV broadcasters have had to figure all this out years ago. Sorry, but all of that is obtuse. The fact that some digital audio can be perceived as much louder than others –– yet it's all limited to the same digital range –– proves they aren't similar at all. There is no such thing as a standard curve for compression. Source levels vary almost infinitely. Accurately separating and reducing sound after the fact, without turning the whole thing to mud, is considered to be an impossible technical challenge. Next, TV broadcasters worked on a predetermined schedule with predetermined advertising. This gave them time to inspect and approve ads in advance. Streaming ads are generally served just in time from third-party services to the streaming host. FFMPEG gets the output from the stream host, but the host has to combine content together from multiple sources (entertainment + multiple ad servers) into that single stream. Currently, sound-level is completely at the whim of each ad server, as well as each ad producer. Meanwhile, the final output is at the whim of the streaming host: 24-hour-news streaming sites probably have different audio standards than Apple TV+. Ultimately, AI could potentially be used to solve it, since it can generate / make-up new sounds as part of reverse-compression. But it would still have to be done in advance by the third-party ad servers.
  - kstrauser 4mo ago
    None of this is true. There are standard curves for human hearing frequency response and you can use these to compare sound A’s volume to sound B. And since sound compression is in DCT space, you can calculate those numbers very quickly with something similar to sum(vol(f) * curve(f) for f in encoded_frequencies). I read the article. It specifically talks about server-side ad embedding, i.e. where the service is inserting ad content into the streams, and therefore, by definition, has access to the ad content. They can do the calculations on their end during the embedding process and normalize volumes there before transmitting the result. To make things even easier, they don’t have to calculate the ad volume each time one’s streamed, just once per ad they’re going to serve. And finally, all of this is a solved problem for TV broadcasters. They face the same problems: advertisers send them content to air, then the broadcasters are legally required to normalize the ad vs content volume, and they do. If this is an insurmountable problem that the streaming services face, they can drive over to their nearest TV station and ask them how they manage to pull off this technological feat.

rdedev 4mo ago

The group includes Netflix, the most technically capable streaming company. It's sad that companies will only go down kicking and screaming even for the mildest of regulations

kstrauser 4mo ago
Yeah. If I had to do this, there’s the likelihood I’d screw it up. I am not Netflix, who quite good at the whole streaming video thing. I find it very hard to believe there are technical challenges in this law that Netflix couldn’t possibly solve.
pengaru 4mo ago
It's arguably a good strategy to make mountains out of mole hills from their perspective. It drains the regulators' finite resources on relatively trivial matters, and I think it goes without saying there are less trivial and more egregious exploits continuing, more unchecked now as a result. TL;DR distract the enemy with resource sinks

avereveard 4mo ago

I hate loud ads as well as anyone else and I welcome this resolution but I would not treat the challenges regulation poses as simplisticly as this. There is a lot of research in increasing loudness without increasing decibels, especially for concerts, but it migrated to ads when tv started adding automatic volume controls to normalize across services.

pkulak 4mo ago

I think the point is that when you don’t control the ad, it’s a bit tough to normalize its audio. And controlling the ad means bringing ad serving in house, which while possible, is a huge engineering ask. I guess the solution is to switch to a proper ad insertion company that normalizes to -24 like you’re supposed to, but that’s not cake either. Especially if contracts are signed.

Ar-Curunir 4mo ago
Well maybe then they should have done quality control. We’ve gotta stop making excuses for these companies that are making money hand over fist.
- dualvariable 4mo ago
  They might have to hire some people and create some jobs... Oh, the humanity...
dinfinity 4mo ago
I assume that they also have mechanisms to check that the content itself is legal to broadcast. Checking the loudness and rejecting based on them in that process should be trivial.
mitthrowaway2 4mo ago
If you're streaming the audio waveform, you can calculate the peak volume and adjust the gain.
- radley 4mo ago
  Most streaming audio already share the same peak volume. The problem is compression. You can compress the hell out of audio, make it sound extremely loud, and it will still have the same peak volume as uncompressed audio.
  - topranks 4mo ago

tzs 4mo ago

> See, that’s just flat-out lying. What’s this mythical circumstance where playing audio A at the same volume as audio B on one device will magically make A louder than Bon another? Especially when dealing with server-side ad insertion, as the article discusses, where the service has full control of the input files and the output stream? Consider a streaming movie with surround sound with an inserted ad that is in stereo. I'm playing it on a 5.1 home theater system and you are playing it on a stereo phone. Your system is mixing the surround sound down to stereo. When your device does that it applies attenuation to the program so that if several channels in the 5.1 stream have something loud all at the same time it won't be too loud in the down mix for stereo and clip. When the commercial cuts in your device recognizes it is ordinary stereo and it doesn't need to down mix. It goes straight through without the attenuation that down mixer applies to the program. Whatever level the commercial is really at relative to the program, it is going to sound loader than that on your system because of that attenuation difference. On my device it is not attenuating the 5.1 program since it has all the necessary channels. However, if the commercial is at the same level as the program it will actually sound louder on mine. That's because the same total level of sound split among 5 speakers perceptually seems less loud than the same total level coming from stereo speakers. The streamer can do loudness normalization between the program and the commercial. It can calculate what the perceived human loudness will be at any time in the adjust the levels so that on my device the perceived level of the 5.1 program when it gets to the commercial will match the perceived level of the commercial on stereo. But for devices that are down mixing to stereo there is still going to be the attenuation the down mixers uses, and that differs from device to device. That limits what can be done server side to get the program and commercial to match. Some multichannel formats do include metadata for the device telling it how much to attenuate when down mixing to stereo. If all the device supported that it should be possible for the server to fully take care of loudness matching. Otherwise you probably need device side normalization. Another approach would be to up mix the stereo commercial server side to whatever surround sound format the program is using. Then they could do server side loudness normalization between the program and the commercial without it being messed up by the difference in how stereo devices down mix. I'm not sure why that is generally not done. LLMs are suggesting several reasons but I have no idea if they are reasonable. I'll leave exploring that to someone else.