Redbook vs. High Resolution Tests Are Clickbait

[STC]

8 hours ago, John Dyson said:

It is likely that the recording engineer hadn't even listened to the DolbyA encoded version of the material that is so much beloved today by some.

 

I am with you on this one. 

 

Speaking of your decoded files which differences can be heard with headphones may be less important to people like me listening over the loudspeakers. 

 

Firstly, an ideal loudspeakers setup should have a descending slope frequency response at LP. That’s because HF attenuated much quicker over the distance due to air damping. Unfortunately, this also raises another question in room EQ where they try hard to have it as flatten freq response. At ears, the pinna function amplifies the HF as much as 20dB so even a different of 0.3dB at 3 meter can be one or two dB difference to the brain. 

 

I generally believe most high end loudspeakers carry far more HF energy than what your experience in live unamplified performance. Higher bass and treble often perceived to be better and a great marketing trick. 

 

Going back to your OCS, what you perceive as distortion of HF may actually sound better over the loudspeakers. 

 

Engineers  look at the signal from different perspective, they try to maintain the fidelity from mic to the output at speakers. Unfortunately, a mic ( assuming a stereo ) maintains the separation from the moment they capture the sound wave, to the media and playback until the speakers they are clearly separated. During playback over loudspeaker this separation are now mixed as one received by the two ears inside the brain which is entirely different on how the original signal was captured, stored and played back. You mix a positive and a negative phase during this stage you get a null but if you output them to the ears you hear them loud and clear.  Human hearing is not processed like audio signal in audio equipment. 

[John Dyson]

13 hours ago, STC said:

 

I am with you on this one. 

 

Speaking of your decoded files which differences can be heard with headphones may be less important to people like me listening over the loudspeakers. 

 

Firstly, an ideal loudspeakers setup should have a descending slope frequency response at LP. That’s because HF attenuated much quicker over the distance due to air damping. Unfortunately, this also raises another question in room EQ where they try hard to have it as flatten freq response. At ears, the pinna function amplifies the HF as much as 20dB so even a different of 0.3dB at 3 meter can be one or two dB difference to the brain. 

 

I generally believe most high end loudspeakers carry far more HF energy than what your experience in live unamplified performance. Higher bass and treble often perceived to be better and a great marketing trick. 

 

Going back to your OCS, what you perceive as distortion of HF may actually sound better over the loudspeakers. 

 

Engineers  look at the signal from different perspective, they try to maintain the fidelity from mic to the output at speakers. Unfortunately, a mic ( assuming a stereo ) maintains the separation from the moment they capture the sound wave, to the media and playback until the speakers they are clearly separated. During playback over loudspeaker this separation are now mixed as one received by the two ears inside the brain which is entirely different on how the original signal was captured, stored and played back. You mix a positive and a negative phase during this stage you get a null but if you output them to the ears you hear them loud and clear.  Human hearing is not processed like audio signal in audio equipment. 

I definitely do not disagree with you, but think about this:  Ad-hoc DolbyA compression, or a proper shaping scheme -- which is better?

 

DolbyA leaks are an oversight, mistake, or resulting from cheapskates (or maybe some other mess-up) -- but certainly don't optimize for anything except for allowing a subsequent pass through another device to avoid certain kinds of noise AFTER decoding.   DolbyA does not optimize for listening, and even when it was originally voluntarily used for enhancement -- the benefit came as a result of R. Dolby's genius attempt to mitigate some modulation type distortions in a very simplistic but practical way.  A primitive compressor without R Dolby's tricks would not sound as good, and I dearly suspect that many previous attempts to replicate DolbyA compression have resulted in splat city and not good compression.  (R Dolby used an amazing economy of design, and the circuitry behavior is much more subtle than a first pass scan might illuminate.)

 

Nowadays, we can do better/cleaner/more optimized processing for the 'brightening' that might be desired to compensate for a diffuse environment.  DolbyA compression might even come close in some ways, but it is really kind of dirty for the application.  Specifically, the HF0/HF1 band overlap and the weird interactions like HF1 suppressing HF0 -- contrary to what is desired WRT audible distortion-like effects are not helpful.

 

If someone wants signal 'brightening', the signal should be brought down to 'baseline' (that is, a clean signal) first, then apply some processing that is nearer to the optimum for its purpose.   For example, maybe a general segmented, but full bandwidth compression might be beneficial -- but certainly not the HF0/HF1 mix the way that it is.  There might be some good concepts for the 'brightening' in DolbyA, but DolbyA isn't adjustable in certain desirable ways also.

 

The saddest thing is that the leaked DolbyA compression shouldn't be imposing itself on all listeners of material from the '60s through the '80s and even later.   Case in point -- a friend just mentioned (just yesterday)  that his Crystal Gayle recordings don't sound as good as the vinyl -- guess what?  Another leaked DolbyA.  I am NOT a mastering expert (BY FAR), I suck t*rds at mastering, but almost anyone would feel that the proper decoding that I did for him sounds better/cleaner.   If one wants to add some proper 'brightening', then 'proper brightening' might be in order, but certainly not the dirty/grainy DolbyA sound!?!?

 

John

[plissken]

On 8/5/2019 at 7:52 PM, The Computer Audiophile said:

I recently read about another Redbook vs. High Resolution listening test supposedly coming up soon. It made me think, why have such a test? Ah ha, it may drive traffic. 

 

Think about it, what other reason is there for a test like this? Every DAC now plays high resolution and every album available for purchase in high resolution is also available for streaming in high resolution. 

 

Thus, nobody is saving money by finding out they can’t hear a difference between the two versions.

 

What could possibly be gained from a conclusive test result for or against high resolution? Nothing. 

 

I’m all ears and willing to change my mind if anyone wishes to provide evidence that a Redbook vs. High Resolution test has any value. 

 

Note: This has zero to do with one’s belief in or disbelief in high resolution. That doesn’t matter in the least either, especially for purposes of this post. 

 

Been banging that drum for years. At the file sizes of 24/192 PCM, storage and the fact that even phones have unlimited data (mine does at $40 a month) it's a non-sequitur.

 

The only other consideration is the fact that some outfits like to charge an appreciable markup for 'hi-rez' files. Files that they put in the extra work to down convert to 16/44.1.  Doesn't make a lick of sense.

[botrytis]

22 minutes ago, plissken said:

 

Been banging that drum for years. At the file sizes of 24/192 PCM, storage and the fact that even phones have unlimited data (mine does at $40 a month) it's a non-sequitur.

 

The only other consideration is the fact that some outfits like to charge an appreciable markup for 'hi-rez' files. Files that they put in the extra work to down convert to 16/44.1.  Doesn't make a lick of sense.

 

 

Tread carefully on unlimited as most plans are NOT UNLIMITED (it is marketing speak). Example I have a cell phone with an unlimited plan but it is not - there is a limit to the amount of data I can use then they throttle my connection. Even Comcast, with their cable modems, give you a monthly cap of 1 terabyte of data, (and if you stream videos through Netflix, etc you chew through it fast) unless you pay more.

[tmtomh]

21 hours ago, fas42 said:

What physics? Depth has nothing to do with bass, which is the main lacking of those speakers - soundstaging is a function of the ability of the playback chain to be clean enough so that the low level detail which allows the mind to decode "what it all means" is there - you start seeing the "big picture".

 

Note, a YouTube clip on this laptop doesn't do this - not because of the speakers, but because of the processing chain. This is PeterSt territory - the 'wrong' use of a software player blurs the sound; and so I use a version of MediaMonkey for which I have carefully adjusted every setting, to get the best SQ from this portable. foobar, for example, is pretty awful on this m/c; I ditched it straightaway.

 

Frequent peak levels of the digital data can cause issues in the conversion process; this is a well known factor. And it turns out this matters on my laptop. I have often dropped the level by 3dB or so on the track, to realise far better playback,' the actual volume of the speakers remains the same, because I can now turn up the volume to maximum, with no issues.

 

The laptop is not "superior"; it's merely, good enough, for a lot of the time - a better rig will hopefully emerge in the near future.

 

Thanks, as always, for your civil and specific reply, Frank. I don't agree with you on a number of things, but I do appreciate your good nature and good-faith discussion.

 

RE "physics," I was not talking about bass, I was taking about soundstage depth. Laptop speakers are not going to provide aural clues of soundstage depth beyond a few inches (and I'm being generous here) forward or back - especially if you listen with your ears very close to the speakers as you say. To be clear: I am sure that a laptop's speakers are capable of reproducing/evoking some degree of soundstage depth. But if one recording is made in a way that produces a small soundstage depth, say just a foot or less, when played on a full-size stereo and speaker rig in an average-size listening room, while another one  is made in a way that produces a large soundstage depth, say several feet, in the same situation, you are not going to be able to distinguish between the two when your laptop is capable of throwing up a much tinier soundstage depth than either recording is capable of. 

 

And if you are tempted to respond that the scale of the soundstage is smaller but the proportion between those two would still be easily audible to you on your laptop, please just don't - it's not a response that is going to persuade me (or, I daresay, anyone else here).

 

As for different software players producing sound that is as radically different as you claim, I'm sorry but once again I am not persuaded by that. Don't get me wrong - I have no problem believing that a YouTube clip (with YT's compression applied during upload/processing), played in an embedded YT video window with few customizable audio settings, could sound different/worse than the same source music file played on a dedicated software player. But if you can't get a fully configurable, bit-perfect app like foobar to sound as good as (let alone better than) MediaMonkey, that tells me that what you're doing is fighting the coloration and limitations of your laptop's amp and speakers and perhaps accidentally finding a setting that gives some euphonic benefits - but it's got nothing to do with fidelity.

 

Finally, yes, frequent 0.0 digital peaks can cause intersample clipping when the digital stream is converted back to analogue. But once again, that's only a problem if it produces audible distortion. And if you're using a playback rig with inherent distortion that meets or exceeds the tiny amount of distortion produced by intersample peaks, then your rig by definition cannot be used to detect whether or not intersample peaks are present or whether they are an issue even if they are present.

 

In addition, if you laptop is "sensitive" to intersample peaks, that only means that your laptop's analogue output section does not have any voltage headroom built into it. The solution there is easy - use a DAC with built-in analogue headroom. Intersample clipping then ceases to be an issue, as intersample clipping is purely analogue clipping and is not actually an "issue" with the recording.

So to recap:

- You appear to have pulled back from your claim that your laptop setup is better than "most high end rigs" for discerning recording issues;

- You agree that you are unable to use your setup to discern "recording issues" that involve the lower bass region;

- You have provided evidence/info about digital peaks/intersample clipping that reveals that your setup is not in fact a reliable setup for determining if this is an issue with any given recording.

 

All that remains is your claim that your laptop is good for discerning soundstage depth; and there again, no evidence is provided for that claim.

[fas42]

3 hours ago, tmtomh said:

 

Thanks, as always, for your civil and specific reply, Frank. I don't agree with you on a number of things, but I do appreciate your good nature and good-faith discussion.

 

Thanks!

 

Quote

 

RE "physics," I was not talking about bass, I was taking about soundstage depth. Laptop speakers are not going to provide aural clues of soundstage depth beyond a few inches (and I'm being generous here) forward or back - especially if you listen with your ears very close to the speakers as you say. To be clear: I am sure that a laptop's speakers are capable of reproducing/evoking some degree of soundstage depth. But if one recording is made in a way that produces a small soundstage depth, say just a foot or less, when played on a full-size stereo and speaker rig in an average-size listening room, while another one  is made in a way that produces a large soundstage depth, say several feet, in the same situation, you are not going to be able to distinguish between the two when your laptop is capable of throwing up a much tinier soundstage depth than either recording is capable of. 

 

The precise distances I wouldn't be fussed about - it's the sense of the layering of the sound that I prize, which is bound up in the sense of being able to "see" each sound element clearly.

 

To make it clear again - this particular, HP laptop just happens to have pretty decent sound circuitry and hardware; purely an accident, was never intentional. A top of the line, "massive" Dell laptop I have here as a backup has dreadful sound, the limits of its playback are irritatingly obvious - I tried a few things, but the hardware is too crippled, no point in trying.

 

Quote

 

And if you are tempted to respond that the scale of the soundstage is smaller but the proportion between those two would still be easily audible to you on your laptop, please just don't - it's not a response that is going to persuade me (or, I daresay, anyone else here).

 

The listening experience is similar to using decent headphones with the HP - when you think about it, that's exactly what the laptop speakers are:drivers from headphones placed inside a plastic case.

 

Quote

 

As for different software players producing sound that is as radically different as you claim, I'm sorry but once again I am not persuaded by that. Don't get me wrong - I have no problem believing that a YouTube clip (with YT's compression applied during upload/processing), played in an embedded YT video window with few customizable audio settings, could sound different/worse than the same source music file played on a dedicated software player. But if you can't get a fully configurable, bit-perfect app like foobar to sound as good as (let alone better than) MediaMonkey, that tells me that what you're doing is fighting the coloration and limitations of your laptop's amp and speakers and perhaps accidentally finding a setting that gives some euphonic benefits - but it's got nothing to do with fidelity.

 

Sorry, it has everything to do with fidelity - yes, you are fighting the limitation of the engineering of the playback circuitry, but what you are doing is finding a configuration that prevents the interference effects of sub-optimal implementation causing the worst damage. foobar doesn't allow control over how the retrieval and buffering of the data occurs, but Media Monkey does - giving one workarounds to help with processor interactions with the analogue circuitry. It's trivially easy to see the clues: foobar constantly chews up CPU cycles, and constantly keeps access to the drive happening; MM goes to zero CPU usage, and barely tickles the drive, doing most at the beginning.

 

The audible indications are that the treble in foobar is dead and grey, lifeless in the way of an AM radio; MM restores the lift and sparkle, say in violin versus a female voice, .

 

Quote

 

Finally, yes, frequent 0.0 digital peaks can cause intersample clipping when the digital stream is converted back to analogue. But once again, that's only a problem if it produces audible distortion. And if you're using a playback rig with inherent distortion that meets or exceeds the tiny amount of distortion produced by intersample peaks, then your rig by definition cannot be used to detect whether or not intersample peaks are present or whether they are an issue even if they are present.

 

 

Quote

 

In addition, if you laptop is "sensitive" to intersample peaks, that only means that your laptop's analogue output section does not have any voltage headroom built into it. The solution there is easy - use a DAC with built-in analogue headroom. Intersample clipping then ceases to be an issue, as intersample clipping is purely analogue clipping and is not actually an "issue" with the recording.

 

 

Don't read too much into what I'm using the laptop for - I have no intentions of hearing "how good something is" in anything like an absolute sense; rather, what it allows is me to hear many of the issues with the playback of a playback rig that's been recorded, and differences in versions of waveforms.

 

Quote

 

So to recap:

- You appear to have pulled back from your claim that your laptop setup is better than "most high end rigs" for discerning recording issues;

 

Historically, "high end rigs" get the things wrong that the laptop does well enough for me to be happy with it.

 

Quote

- You agree that you are unable to use your setup to discern "recording issues" that involve the lower bass region;

- You have provided evidence/info about digital peaks/intersample clipping that reveals that your setup is not in fact a reliable setup for determining if this is an issue with any given recording.

 

All that remains is your claim that your laptop is good for discerning soundstage depth; and there again, no evidence is provided for that claim.

 

It's hard to provide evidence that a playback system works well in some area - as a general rule, . This laptop has developed an issue with the sound pulling to the left - which is bugging me a bit more. It's old, there will be a point when I'm too annoyed with it to fuss any more ; I'm not going to get precious about a bloody laptop, .

[STC]

9 hours ago, John Dyson said:

Ad-hoc DolbyA compression, or a proper shaping scheme -- which is better?

 

Of course, I will always go for perfection. 

 

Thank you for the detailed explanation.

 

Regards,

ST

[STC]

On 9/5/2019 at 8:38 PM, John Dyson said:

Yes, I just double checked -- the '11. Orange Colored Sky-V0.9.7K-snippet.flac' is definitely processed.   I just checked the gain control log for 'Orange Colored Sky' and even though on POP material, the MF band is usually pinned at 'full gain' (effectively 0dB), there are enough dynamics in OCS to cause some dynamic gain control in the MF band.  (MF band usually comes into play on very low level material.)   The other bands -- HF0 is doing the full -10 to 0dB gain change and the HF1 band is doing the -15 to 0dB gain change.  The LF band also bounces around -- but usually has little audible effect.  SO, there is a good amount of dynamic expansion going on during the decoding of OCS.  (Legend:  LF = 0-80Hz, MF = 80 - 3kHz, HF0 = 3k-9kHz/3k-20+kHz,  HF1 = 9k-20+kHz.)  The HF0/HF1 bands have a delicate & precision 'dance' between them or the sound will not be so good.

 

The only caveat -- and the BIG problem when decoding feral material -- I found that the calibration might be slightly in error on this decode (probably about 0.1dB.)  I noticed that the vocal smooths out a LITTLE bit more with a slight decrease in the calibration level.  Without tones and without good provenance on the recording, then decoding can become a 'science project'.  The 'error' is smaller than what a person doing the mastering would likely allow (not worth fixing -- really.)

 

John

 

 

Sorry for the delay.

 

After listening with my main system to the 55 sec snippets of Orange Coloured Sky. My observation as follows:-

 

1) the vocal sound natural compared to the RAW sample. 

 

2) I did sense a bigger stage but less depth with the RAW. But slightly narrow stage but better depth with the decoded. It only happened at the first few listening and now I am unable to distinguish them easily under blindtests because once you know the existence of something it is hard to unhear them. 

 

3) At some segments where the is burst of energy, I preferred the RAW version. It is something like listening to soft clipping of tube amp. Having said that, you can always use compression and EQ to produce better sound using the Decoded version. So it may be possible to make a very good remasters  of ABBA with your decoded files. 

 

4) Please listen to the sample again from 45 to 48 seconds. There is a marked difference between the RAW and Decoded. Again this can be easily addressed with EQ. IN my system, the decoded version pushes the hit hat buried in the background. I suspect the RAW version did more than removing the hiss with DOLBYA. 

 

ABBA never sounded good to me with CD, I enjoyed them more during my cassette era. After hearing the vocals of the decoded sound I would reconsider getting some of my favorites again if at all you decided to commercialize them. 

 

Thanks for sharing the files.