[bigshot]

I never claim that people should take my tests as gospel. I share what I’ve found and I encourage them to check for themselves in good faith. Massive amounts of time are wasted in this forum replying to trolls and disingenuous people. I’d say 80% of the threads here are trolled regularly, including the most important and useful ones. I’m barred from posting in the rest of head-fi on subjects that fall under the topic of sound science. I think people who are out to undermine science should be barred from posting here.

I know exactly where this is going. He’ll keep taking everyone’s comments and say, “Yup, I did that now. Still hear a night and day difference.” The problem here isn’t the controls and variables. It’s a fundamental lack of commitment to integrity. Someone else needs to oversee the test and administer it to him- someone who knows what they’re doing and is on the lookout for tricks. Until then, it’s just talk.

I’ll step back if you want and let you guys go back and forth with him, but I think all know where this is going to end up because we all know that what he is claiming isn’t possible. Either he is doing his test wrong or he is gaming it. Either way, he’s never going to admit it, and we’re never going to get any solid proof of what he is claiming- just post after post after post of derailing an otherwise useful thread.

If you have a solid strategy for pinning down the truth, have at it. Good luck. I think this whole thing is too wiggly to get anywhere.

 

[Hifiearspeakers]

I never claim that people should take my tests as gospel. I share what I’ve found and I encourage them to check for themselves in good faith. Massive amounts of time are wasted in this forum replying to trolls and disingenuous people. I’d say 80% of the threads here are trolled regularly, including the most important and useful ones. I’m barred from posting in the rest of head-fi on subjects that fall under the topic of sound science. I think people who are out to undermine science should be barred from posting here.

I know exactly where this is going. He’ll keep taking everyone’s comments and say, “Yup, I did that now. Still hear a night and day difference.” The problem here isn’t the controls and variables. It’s a fundamental lack of commitment to integrity. Someone else needs to oversee the test and administer it to him- someone who knows what they’re doing and is on the lookout for tricks. Until then, it’s just talk.

I’ll step back if you want and let you guys go back and forth with him, but I think all know where this is going to end up because we all know that what he is claiming isn’t possible. Either he is doing his test wrong or he is gaming it. Either way, he’s never going to admit it, and we’re never going to get any solid proof of what he is claiming- just post after post after post of derailing an otherwise useful thread.

If you have a solid strategy for pinning down the truth, have at it. Good luck. I think this whole thing is too wiggly to get anywhere.

You need a vacation.

 

[castleofargh]

I never claim that people should take my tests as gospel. I share what I’ve found and I encourage them to check for themselves in good faith. Massive amounts of time are wasted in this forum replying to trolls and disingenuous people. I’d say 80% of the threads here are trolled regularly, including the most important and useful ones. I’m barred from posting in the rest of head-fi on subjects that fall under the topic of sound science. I think people who are out to undermine science should be barred from posting here.

I know exactly where this is going. He’ll keep taking everyone’s comments and say, “Yup, I did that now. Still hear a night and day difference.” The problem here isn’t the controls and variables. It’s a fundamental lack of commitment to integrity. Someone else needs to oversee the test and administer it to him- someone who knows what they’re doing and is on the lookout for tricks. Until then, it’s just talk.

I’ll step back if you want and let you guys go back and forth with him, but I think all know where this is going to end up because we all know that what he is claiming isn’t possible. Either he is doing his test wrong or he is gaming it. Either way, he’s never going to admit it, and we’re never going to get any solid proof of what he is claiming- just post after post after post of derailing an otherwise useful thread.

If you have a solid strategy for pinning down the truth, have at it. Good luck. I think this whole thing is too wiggly to get anywhere.

So your gut feeling about his motives is all you need, to know that you're right. And when confronted by some of us about the weakness of that rational, you double and triple down on it without any form of evidence.
where have I seen this behavior before? I can't put my finger on it...

 

[Davesrose]

1. Oversampling and dither "yes" but not noise-shaped dither. Noise-shaped dither only came about in the early 1990's, when higher than 16bit recording became available. Even the theory of noise-shaped dither was not published until 1989.

2. I'm not very "au fait" with digital imaging but the dynamic range of digital audio is NOT a main limit, in fact, it couldn't be less of a limit! Digital audio devices capable operating at a bit depth of 32bit (float) started appearing nearly 30 years ago, Neve's Capricorn mixing desk being the earliest example, and 32bit float has a dynamic range of over 1500dB! This of course cannot be realised outside the digital domain because EVERY other limit is massively more restrictive (more "main"), for example: The acoustic limit of a sound pressure wave in air is only 194dBSPL and the dynamic range limits of analogue components (transducers, etc.) are miniscule in comparison and of course we can't forget the dynamic range limitations of human hearing, which are also miniscule in comparison.

Once again you take a post of mine out of context. I guess I should respond with your odd nomenclature:

1. Notice I wrote "noise shaping". If you're going to argue that the Philips DAC didn't implore it, you should contact all the numerous sources that say it did (including Philips themselves). Evolution of DAC & digital filter , The history of the CD - Technology

" The conversion of the digital ‘zeros’ and ‘ones’ into an analogue signal also proved to be a tougher challenge than was at first thought. And it was also very difficult to keep the conversion process linear at lower signal levels, for example between -60 dB and -100 dB.At the introduction of the CD player, every player had a so called 'ladder' D/A converter, followed by a steep analogue filter to remove frequencies above 20 kHz. Philips was the only company to use four times oversampling, with a digital filter, right from its first player. Because four times oversampling means that four samples are taken every 1/44,100th of a second instead of just one, this in combination with first-order noise shaping, which Philips was also the first to apply, allowed 16 bit resolution to be achieved with a 14 bit D/A converter."

2. Notice I said system and not processing. I can generate 32bit per channel image files (used for light simulation and lots of post processing leeway with developing an image). But the current best system for displaying dynamic range is 10 stops (or just over 1000 nits).

Another bit of info: current consumer standards for video can have up to 12bit color space. Most prominently Dolby Vision, that video that's playing at 12bit gets dynamically "tone mapped" to the display's 10bit space (IE constantly changing contrast range from 12bit's 4096 shades of tone compared to 10bit 1024). I would think that most people reading this would understand that this is a form of processing, and that the output is still 10bit/stops 1000 nit display.

 

[KeithPhantom]

Far more likely is that you are being annoyed by IMD much lower in the spectrum, caused by the parasitic tone. Your observation highlights a particular issue for engineers, especially with higher sample rates, it's very difficult to avoid, correct or even identify spurious tones/sounds that we can't hear.

I also thought it was IMD, but I compared by ear the pitch of the tone to a pure sine wave around that frequency, they seemed to be really similar.

 

[gregorio]

Once again you take a post of mine out of context. I guess I should respond with your odd nomenclature:
1. Notice I wrote "noise shaping". If you're going to argue that the Philips DAC didn't implore it, you should contact all the numerous sources that say it did (including Philips themselves). Evolution of DAC & digital filter , The history of the CD - Technology
2. Notice I said system and not processing.
2a. I can generate 32bit per channel image files (used for light simulation and lots of post processing leeway with developing an image). But the current best system for displaying dynamic range is 10 stops (or just over 1000 nits).

1. I wasn't aware Philips used noise shaping in their early CD players, although it appears to be very simple 1st order shaping. As far as I'm aware (which obviously isn't as much as I thought!) noise shaping in consumer DACs didn't become widespread until the late 1980's - early 1990's, when 1 bit DS converters became the dominant topology. However, it's hardly surprising I took your post out of context, because it WAS out of context! The context of this thread is obviously 24bit vs 16bit and therefore I took your post in the context of the (high order) noise-shaped dither applied as standard procedure when converting between "high-res" and 16bit.

2. How could I notice you "said system and not processing" when that's NOT what you said?! What you actually said was: "I would think main limits of a digital system are dynamic range ...". An early digital system, one comprising the Neve Capricorn for example, had a dynamic range of 144dB, it's internal processing was 32bit float but it's internal routing and connections to other equipment in the digital system were 24bit ... but 144dB dynamic range is certainly NOT the "main limit", it's the opposite, the least of all the limits! By the mid/late 1990's we had mixing desks, such as the famous Sony Oxford, which maintained 32bit not only for it's processing but also for all it's internal routing/patching and with internal EQ, compression/expansion, limiting, delay and noise gates, it was more of an integrated digital system, although external connections to MTRs and external processors (such as reverbs) was still limited to 24bit (144dB). However, in the 2000's DAWs made it possible to integrate the entire digital system "in the box" and maintain the 32bit float format (~1500dB dynamic range) throughout the digital system, even the resultant audio files. Of course, if we're going to actually hear anything from a digital system it has to be connected to an analogue/acoustic system, which presents a far more massive dynamic range limitation than the digital system.
2a. Maybe the limitation of "just over 1000 nits" is an analogue/optical limitation and maybe in the video world it's typical to refer to a digital/analogue/optical system as just a "digital system" because the analogue/optical part isn't considered "a system" but regardless, discussion of video/images is off topic!

G

 

[Davesrose]

1. I wasn't aware Philips used noise shaping in their early CD players, although it appears to be very simple 1st order shaping. As far as I'm aware (which obviously isn't as much as I thought!) noise shaping in consumer DACs didn't become widespread until the late 1980's - early 1990's, when 1 bit DS converters became the dominant topology. However, it's hardly surprising I took your post out of context, because it WAS out of context! The context of this thread is obviously 24bit vs 16bit and therefore I took your post in the context of the (high order) noise-shaped dither applied as standard procedure when converting between "high-res" and 16bit.

2. How could I notice you "said system and not processing" when that's NOT what you said?! What you actually said was: "I would think main limits of a digital system are dynamic range ...". An early digital system, one comprising the Neve Capricorn for example, had a dynamic range of 144dB, it's internal processing was 32bit float but it's internal routing and connections to other equipment in the digital system were 24bit ... but 144dB dynamic range is certainly NOT the "main limit", it's the opposite, the least of all the limits! By the mid/late 1990's we had mixing desks, such as the famous Sony Oxford, which maintained 32bit not only for it's processing but also for all it's internal routing/patching and with internal EQ, compression/expansion, limiting, delay and noise gates, it was more of an integrated digital system, although external connections to MTRs and external processors (such as reverbs) was still limited to 24bit (144dB). However, in the 2000's DAWs made it possible to integrate the entire digital system "in the box" and maintain the 32bit float format (~1500dB dynamic range) throughout the digital system, even the resultant audio files. Of course, if we're going to actually hear anything from a digital system it has to be connected to an analogue/acoustic system, which presents a far more massive dynamic range limitation than the digital system.
2a. Maybe the limitation of "just over 1000 nits" is an analogue/optical limitation and maybe in the video world it's typical to refer to a digital/analogue/optical system as just a "digital system" because the analogue/optical part isn't considered "a system" but regardless, discussion of video/images is off topic!

G

So when you try to draw analogies with photography in the very first paragraph in this thread, it's not off topic then??? Granted, throughout this thread, you have shown ignorance of photographic systems...in how you don't know standard bit depth layers and how that's relational with dynamic range in a visual system. How then also in the very response to my use of "digital system", you prove the point about 32bit audio processing in relation to the system in which output does not meet the same specs (I'm sure other posters in this thread who have been talking about whether they can hear any differences in AQ with their own personal digital stereo systems, know my context of digital system)! My analogy of HDR with imagery was that RAW video can be 12bit (or 14 or 16 bit), and then it has to get reduced to the 10bit space digital TVs can display (0-1023 values, up to 1024 nits). A similar analogy to your own argument of a theoretical dynamic range with a 24bit source vs audio system!! I notice you tend to not just take posts out of context, but you also conflate terms. For example, I did look up that the 14bit DAC Philips introduced in 1982 introduced 4x oversampling and noise shaping (apparently used in theirs and Magnovox CD players)...when the context of my response was people thought these early CD players sounded the best (the context of the above posts were about the early 14bit and 16bit DACs, not 16bit vs 24bit): and that you are clearly still going out of context with other noise shaping methods. The best example of your conflation for me has been a past claim that Dolby's marketing for 3D audio isn't correct because even though it's comprised of X,Y, and a Z axis....its overhead array of speakers isn't overhead (and that in your meaning it would also need positional audio towards your feet).

 

[pinnahertz]

Philips noise shaping and 4X over sampling was an attempt to utilize 14 bit DAC chips in a world moving to 16 bit DACs. It only sort of worked, the players had OK SQ, but had so many other issues, particularly in the clunky control concept, that they became legendary, and not always in a good way. Initially they were welcomed to a niche of audiophiles, then rapidly rejected. I had one, and learned to hate it really quickly. Their DAC system did not sound noticeably better than the 16 bit/analog filter machines of the era. This was the era that stimulated better data recovery from the CD, and better error correction/concealment implementation. Glad it's all over.

 

[Hifiearspeakers]

Philips noise shaping and 4X over sampling was an attempt to utilize 14 bit DAC chips in a world moving to 16 bit DACs. It only sort of worked, the players had OK SQ, but had so many other issues, particularly in the clunky control concept, that they became legendary, and not always in a good way. Initially they were welcomed to a niche of audiophiles, then rapidly rejected. I had one, and learned to hate it really quickly. Their DAC system did not sound noticeably better than the 16 bit/analog filter machines of the era. This was the era that stimulated better data recovery from the CD, and better error correction/concealment implementation. Glad it's all over.

That’s fine if you didn’t like them, but it wasn’t his point that they necessarily sounded good. His point was that the technology existed and was implemented by them back then, which was earlier than what Gregorio had quoted.

 

[Davesrose]

That’s fine if you didn’t like them, but it wasn’t his point that they necessarily sounded good. His point was that the technology existed and was implemented by them back then, which was earlier than what Gregorio had quoted.

Pinnahertz wasn't arguing with me. He specifically did say Philip's use of noise shaping and oversampling wasn't his cup of tea. I didn't have much personal knowledge. I was just a few years old, but I remember my parents and grandfather got an early CD player to plug into their stereo...and they sounded great at the time (in comparison to the audio cassette and record players they had). Really doubt you could tell any audio difference between Sony's 16bit or Philips 14 bit DACs with those systems (and Pinnahertz says more advantages with CD technology were error correction). My main point was just that there was an apparent difference with processing in even the earliest consumer DACs.

 

[Hifiearspeakers]

Pinnahertz wasn't arguing with me. He specifically did say Philip's use of noise shaping and oversampling wasn't his cup of tea. I didn't have much personal knowledge. I was just a few years old, but I remember my parents and grandfather got an early CD player to plug into their stereo...and they sounded great at the time (in comparison to the audio cassette and record players they had). Really doubt you could tell any audio difference between Sony's 16bit or Philips 14 bit DACs with those systems...just that there was an apparent difference with processing in even the earliest consumer DACs.

Ok, fair enough. I just felt like you’ve made some good points in here that have been taken out of context and subsequently, flippantly dismissed. I’m just trying to keep that from continually happening so that the conversation can progress.

 

[pinnahertz]

Having the technology and utilizing it well don't typically fall at the same point in the time line.

 

[davidtriune]

I used to think audio clipping in a track is due to a lack of dynamic range, that there was not enough bits to record very loud sound. But now I learned that clipping is an artist preference and it has nothing to do with 16/24bit.
16/24 bit determines how quiet a sound goes, not how loud it goes.

 

[gregorio]

[1] Granted, throughout this thread, you have shown ignorance of photographic systems...
[2] How then also in the very response to my use of "digital system", you prove the point about 32bit audio processing in relation to the system in which output does not meet the same specs.
[3] For example, I did look up that the 14bit DAC Philips introduced in 1982 introduced 4x oversampling and noise shaping (apparently used in theirs and Magnovox CD players)...when the context of my response was people thought these early CD players sounded the best (the context of the above posts were about the early 14bit and 16bit DACs, not 16bit vs 24bit): and that you are clearly still going out of context with other noise shaping methods.
[3a] The best example of your conflation for me has been a past claim that Dolby's marketing for 3D audio isn't correct because even though it's comprised of X,Y, and a Z axis....its overhead array of speakers isn't overhead (and that in your meaning it would also need positional audio towards your feet).

1. Clearly that's false, as I obviously haven't mentioned or discussed "photographic systems" THOUGHOUT this thread! The discussion throughout this thread has been 24bit vs 16bit digital audio and NOT photographic systems. I mentioned half a sentence worth of photography in the OP as a simple analogy: "It's easy to see in a photograph the difference between a low bit depth image and one with a higher bit depth" - Nothing controversial or wrong about that statement, that readers a dozen years ago wouldn't have easily understood.

2. Huh? I explained the exact opposite! Even by 20 years ago, 32bit was NOT ONLY the bit depth of the processing, it was ALSO the bit depth of the patching/routing between all the various processors AND of the output (digital audio files). 32bit was utilized THROUGHOUT the digital audio chain. And even before that point in time, going back nearly 30 years to the earliest use of 32bit in pro-audio equipment, the output was limited to 24bit, which with a dynamic range of 144dB is clearly NOT the "main limit", as you stated: "I would think main limits of a digital system are dynamic range"!

3. "Other noise shaping methods", the high order noise-shaped dither required by delta-sigma sampling/conversion and standard procedure for converting 24bit to 16bit, is obviously NOT out of context, given the title of this thread!
3a. Clearly that's nonsense, I certainly did not claim that the overhead array of speakers in a Dolby Atmos system "isn't overhead". Obviously, the soundfield of a Dolby Atmos system is a hemishpere, it effectively only allows half the Z axis.

His point was that the technology existed and was implemented by them back then, which was earlier than what Gregorio had quoted.

Sort of. It was a very basic form of the technology (only 1st order noise-shaping), the only consumers who encountered it were those few who owned that particular model/models of DAC and, it was used to just recreate the 16bit dynamic range limitation already existant on the CD, not improve on it. But, in the 1990's all consumers encountered noise-shaping, typically 3 applications of high order noise-shaping, but at least one: Noise-shaping during initial recording/sampling, noise-shaping during conversion from 20bit or 24bit to 16bit (manually applied during mastering) and noise-shaping during conversion from digital to analogue. Most decent CD players were capable of the equivalent of about 18bit dynamic range and by the late 1990's many/most CD's had the equivalent of about 20 bits dynamic range. Of course though, this was the DIGITAL dynamic range, the range from digital peak to the digital noise (dither) floor. The actual dynamic range of CDs was limited by factors OTHER than the digital dynamic range and almost never exceeded the equivalent of about 10bits (60dB), these other (non-digital) factors include: The acoustic noise during recording and the analogue noise floor of electric guitar amps/cabs, vintage effects processors, mics, etc.

[1] I used to think audio clipping in a track is due to a lack of dynamic range, that there was not enough bits to record very loud sound.
[2] But now I learned that clipping is an artist preference and it has nothing to do with 16/24bit.
[3] 16/24 bit determines how quiet a sound goes, not how loud it goes.

1. Yep, that assumption was incorrect.

2. It can be an artistic preference but there are also other reasons clipping occurs, Inter-Sample Peaks being one example. Still nothing to do with 16/24bit though.

3. Correct. The maximum (loudest) allowed value in digital audio is 0dB (FS) and that value is always the same regardless of whether the file is 8bit, 16bit or 24bit. What changes between the bit depth formats is the minimum (quietest) value: -48dBFS, -96dBFS and -144dBFS respectively. Although, with dither and noise-shaped dither, these minimum values are even lower.

G

 

[Davesrose]

1. Clearly that's false, as I obviously haven't mentioned or discussed "photographic systems" THOUGHOUT this thread! The discussion throughout this thread has been 24bit vs 16bit digital audio and NOT photographic systems. I mentioned half a sentence worth of photography in the OP as a simple analogy: "It's easy to see in a photograph the difference between a low bit depth image and one with a higher bit depth" - Nothing controversial or wrong about that statement, that readers a dozen years ago wouldn't have easily understood.

That first paragraph says "photograph" (that means a whole photographic system that produced a photograph). In this very thread we have had exchanges on how DR relates to bit depth, and the method of ADC with digital cameras. You demonstrated to me that you thought bit depth had to do with total channels (IE 32bit means 8 bit per channel RGBA). I found it something else, then, that you said that you have sound mixing experience with cinema, and therefore needed some familiarity with video formats. Even years past the date of this thread, digital cameras have shot in a RAW format that has a higher bit depth than the 8bpc jpeg image you try to allude to.

2. Huh? I explained the exact opposite! Even by 20 years ago, 32bit was NOT ONLY the bit depth of the processing, it was ALSO the bit depth of the patching/routing between all the various processors AND of the output (digital audio files). 32bit was utilized THROUGHOUT the digital audio chain. And even before that point in time, going back nearly 30 years to the earliest use of 32bit in pro-audio equipment, the output was limited to 24bit, which with a dynamic range of 144dB is clearly NOT the "main limit", as you stated: "I would think main limits of a digital system are dynamic range"!

Huh back!!????? How do you not understand that 32bit master source has to go through conversion for audio playback on a 16bit CD digital system?? You yourself in previous posts admitted that a 32bit processor has DR reduced by significant margins when it gets to transducers (and then also the below quote where you also include realized DR from analog stages)!! Again you're conflating terms. Maybe given that you arbitrarily post these absurd outlines that take out other key content (like my point that most people would realize I'm talking about a digital audio system), you only understand your out of context bullet points?

3. "Other noise shaping methods", the high order noise-shaped dither required by delta-sigma sampling/conversion and standard procedure for converting 24bit to 16bit, is obviously NOT out of context, given the title of this thread!

Sort of. It was a very basic form of the technology (only 1st order noise-shaping), the only consumers who encountered it were those few who owned that particular model/models of DAC and, it was used to just recreate the 16bit dynamic range limitation already existant on the CD, not improve on it. But, in the 1990's all consumers encountered noise-shaping, typically 3 applications of high order noise-shaping, but at least one: Noise-shaping during initial recording/sampling, noise-shaping during conversion from 20bit or 24bit to 16bit (manually applied during mastering) and noise-shaping during conversion from digital to analogue. Most decent CD players were capable of the equivalent of about 18bit dynamic range and by the late 1990's many/most CD's had the equivalent of about 20 bits dynamic range. Of course though, this was the DIGITAL dynamic range, the range from digital peak to the digital noise (dither) floor. The actual dynamic range of CDs was limited by factors OTHER than the digital dynamic range and almost never exceeded the equivalent of about 10bits (60dB), these other (non-digital) factors include: The acoustic noise during recording and the analogue noise floor of electric guitar amps/cabs, vintage effects processors, mics, etc.

This is a great example of how you do respond to posts and take them out of context and conflate with your own terms. I think in this case, it's just that you simply couldn't admit to being mistaken about the early Philips DAC. I said that the early DAC had 4x oversampling and "noise shaping" (because the content of the thread then was the first 16bit and 14bit DACs). You then responded with a disagreement with the statement about "noise shaping". When presented with the literature, you're still going on ad nauseam of other noise shaping methods (and not accepting written fact about the 14bit DAC). Just a side note for other folks interested in facts...apart from the CD players that used this DAC, I saw a video from Techmoan that Philips first introduced their 14bit system with a digital tape system (that used VHS cassettes).