Why 24/192 is a bad idea?
- Thread starter audioguy
- Start date
Hmph. I can beat that; play max volume, then short the preamp outputs and I can get near infinite dynamic range! 
Problem is Micro this does not give a complete understanding what is happening at all levels of amplitude, which is one of the benefits of the Paul Miller/Keith Howard measurement tool.
Peak and average give very different values, especially when analysing the whole track and then plotting it against frequency/amplitude.
For an example, everyone should take a look at the HiFiNews magazine sometime for their hirez reviews.
Edit:
BTW I do strongly agree with you that there is plenty of energy, which includes those signal related below the peaks.
I am emphasising the point regarding peaks and average because I feel many may not appreciate that the music signal will encompass many different amplitudes-loudness for a specific frequency in a track and all of it is essential for the enjoyment of music.
Cheers
Orb
Last edited:
Tim I was responding to this post of yours.
My point is that a good recording has a lot of its energy actually between -40dbfs and -80dbfs.
This is not the same as playing from no sound to hearing it at 120db.
The point is that much of the digital recording bit depth is applicable due to be so low on the dbfs, and not necessarily having actual 120db of music/dynamic peaks.
This is compounded by just how much the signal-music is quiet at 2khz to 10khz compared to say the bass frequencies on a digital recording.
And as I mentioned above 10khz we have a further drop off of said amplitude energy that needs to be captured (still would love to have some tracks replicated for comparison that are still studio processed with anything above 5khz then 10khz then 13khz etc equalised to being silent).
However from a technical perspective, performance of the digital equipment and process is critical due to the signal being so low and NOT like driving a test sinewave signal at 0dbfs.
I am oversimplifying this but hopefully one can appreciate that in terms of a technical discussion it is not necessarily just about going from silent to 120db peak when listening, but also about the signal (in this case music) and how it is stored-processed-etc.
Hence why 12bits does not take this side into consideration.
That said in reality I doubt we need 24bits and I am with quite a few respected engineers on thinking the ideal figure is 20bits.
As mentioned a good example is that when hitting close to 0dbfs many DACs and CD players will behave differently, I mentioned this in the past and why I keep emphasising the importance in this thread that such discussions as these must consider all aspects from individual components and their spec to architecture and processes involved from end-to-end (studio and any components-processes to listener and again components and processes-algorithms-oversampling-etc).
Just to add I have also mentioned there are bad recordings and this can even be due to downsampling from SACD or upsampling, or other filter anomolies (and these have been easily seen with the HiFiNews measurement tool), so I have focused on both the good and bad for this thread as both have implications but in different ways.
Cheers
Orb
Last edited:
Having read a good bit of the thread my conclusion .. It is not a bad idea.
24/192 as a matter of fact is a very good idea. Given the availability of inexpensive storage ( 1 TB is presently about $70 in the USA). ADC and DAC working well in the region and no clear proof that the ultrasonics cause so much problems , I would err toward an inexpensive "more" rather settle for the baseline ...IMHO of course
24/192 as a matter of fact is a very good idea. Given the availability of inexpensive storage ( 1 TB is presently about $70 in the USA). ADC and DAC working well in the region and no clear proof that the ultrasonics cause so much problems , I would err toward an inexpensive "more" rather settle for the baseline ...IMHO of course
It can be good idea from a technical standpoint if it is implemented correctly and well; this has to take into consideration the processing (whether native hi-rez file or an upsampled-downsampled, importantly the filters-algorithms,etc), the communications such as SPDIF/USB receiver chips that do not all support higher sampling rate and so we get downsampling without knowing or some other anomoly, the spec-performance of components implemented as these can and do have traits even in terms of DAC chips/upsampling chips/etc, and importantly the trade-offs in design and architecture implementation (such as the case where some embedded functionality is compromised as mentioned in previous post as an example but keep costs down), and clocking architecture to fully support both clock types 44.1 and 48khz sampling rate increments.
When considering the full end to end chain, it is not necessarily about costs but correct and well designed implementation, that then still relies upon correctly processed-managed records.
Cheers
Orb
Yes, what's being said here makes sense to me but let me put on my devil's advocate hat. Can we design a test to actually illustrate the loss of low signal level detail on 16bit Vs 20 or 24bit? I'm thinking of a null test for instance! But not a null test that uses an ADC to record the analogue output!
See I'm a measurements person after all
See I'm a measurements person after all
Bruce B
WBF Founding Member, Pro Audio Production Member
Can we design a test to actually illustrate the loss of low signal level detail on 16bit Vs 20 or 24bit? I'm thinking of a null test for instance! But not a null test that uses an ADC to record the analogue output!
See I'm a measurements person after all
Just take any good "true" hirez file and dither it down to 20 and 16 bit. Put the files in a workstation and invert polarity against your reference.
If we did a DiffMaker between the two files could we listen to the Diff file for a clue into the audibility of this difference between the files. My thinking might be off here as I know that it's instrument tails, air & venue ambience, etc. that is being talked about in this "resolution" & these sonic cues, divorced (as they would be in Diffmaker) from their surrounding context may well seem trivial & insignificant. But I don't really know. has anybody done this sort of test?
Bruce B
WBF Founding Member, Pro Audio Production Member
I've never used DiffMaker. I've always used my workstations and lined up all the files and then could select each one individually or together as a null.
Yes, that is what you could expect to lose, ambience will be more dry, reverb tails will fall off more sharply and the "venue" will seem smaller.
Related topic but not exactly from "Residential Systems"
Dolby TrueHD 96k Takes Blu-ray Audio to 11
http://residentialsystems.com/blog/75532.aspx
Dolby TrueHD 96k Takes Blu-ray Audio to 11
http://residentialsystems.com/blog/75532.aspx
Yes, I've also been saying this - minimum phase filters (similar to Apodizing Filters) which cut-out pre-echoes seem to have a significant sonic effect. Have a look at this thread which discusses Linear phase Vs Minimum phase - Interesting participants & comments http://www.gearslutz.com/board/mastering-forum/569408-linear-phase-versus-minimum-phase-eqs.html
http://www.gearslutz.com/board/mastering-forum/569408-linear-phase-versus-minimum-phase-eqs.html
Well, it's gotten far too wonky for me. I'm an objectivist in the sense that my long audio journey has taken me away from the euphonic and toward the more accurate or, lacking belief in that, the crisper, drier? I'm also an objectivist, whatever that means, from having done some blind listening at home and learned my own susceptibilities and limitations in the process. But it has all been about listening. Do I believe the results of the Harman study? Sure. But it's not because I'm that technical. I't's because I've heard those Martin Logans and those Infinitys, and the results are consistent with my experience. And reverb tails and venue ambience where there was no venue? You're basically saying that hi-res preserves the last bit of decay of reverb, and I'd know what that sounds like, and I've been told to listen for it, and I have. I've also blind A/B'd hi-res files and the same file down sampled to 16/44.1. Nothin'. I get the same results all those people in Meyer and Moran got, and yeah, some of those files ended up being fake hi-res, but a whole lot of them were not. And in dozens of trials with different listeners on different systems, playing different music under different circumstances over the course of months no one exceeded the odds of pure chance. Not once.
Does that mean there is no difference? No. Does it mean no one can hear it? Doesn't mean that either. But it's a pretty good indication that whatever it is that the most trained ears on the planet are hearing doesn't amount to much. I don't think I'll spend a moment second-guessing my redbook files. And I still don't think it's worth running the Pure Music I already have to play the few hi-res files I have in their native rates, because iTunes is still much better software. YMMV and all that, but I think this is much ado about the short hairs between nothing and nothing much.
So my final answer to the original question, "Is 24/192 a bad idea?" is probably not. But it's probably not a particularly important one either.
Tim
Does that mean there is no difference? No. Does it mean no one can hear it? Doesn't mean that either. But it's a pretty good indication that whatever it is that the most trained ears on the planet are hearing doesn't amount to much. I don't think I'll spend a moment second-guessing my redbook files. And I still don't think it's worth running the Pure Music I already have to play the few hi-res files I have in their native rates, because iTunes is still much better software. YMMV and all that, but I think this is much ado about the short hairs between nothing and nothing much.
So my final answer to the original question, "Is 24/192 a bad idea?" is probably not. But it's probably not a particularly important one either.
Tim
Last edited:
This is getting around. Maybe we can all talk about it over at "When is Hi-Res not really Hi-Res" topic
This is getting around. Maybe we can all talk about it over at "When is Hi-Res not really Hi-Res" topic
Or again ... :b
This one of three threads to which I subscribe, and all have more or less fused into the same discussion.
Thomas, most threads usually do; metamorphosed into identifiable topics or subjects of similar derivatives.
At the end they achieved the same goals ... Audio reproduction in a verisimilitude.
...Only different terms, channels, techniques, implementations, marriages.
I think I'll start one called "When does hi-res not really matter," and the answer is when a bunch of Audiophiles have bought upsampled 16/44.1 from HD Tracks and didn't notice until they read it on the internet.
Tim
Tim
