• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

How do you like this SINAD Chart Style?

Blumlein 88

Major Contributor
Joined
Feb 23, 2016
Messages
5,169
Likes
3,890
#41
I can't because in all the reviews I have mentioned the four quarters.
Well here is an example of a two color gradient. It blends together so there aren't defined quarters. Even this would be an improvement to me. I think by selecting the inflection points of a 3 color gradient you'd get something like 5 levels sort of with the blending. Again even two colors is helpful I think.
1549584269867.png
 

GoMrPickles

Member
Patreon Donor
Joined
Nov 6, 2018
Messages
74
Likes
66
#42
I'll mention this once again, and then I'll take the lack of response as a lack of interest: All of these items can be combined into a user-editable Google Sheet (free), then interactively controlled from a Tableau Public dashboard (also free). Make, model, cost, SINAD, connections, etc. can all be user-filterable fields, so long as they are tracked on the Google Sheet, e.g.: https://docs.google.com/spreadsheets/d/1juB3DmeI_qURpXJRRDKiEGr3aj9KkftKXx-WX59ZYCU/edit?usp=sharing

ETA: Like this.
 

Attachments

Last edited:

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
17,781
Likes
15,937
Location
Seattle Area
#45
I'll mention this once again, and then I'll take the lack of response as a lack of interest: All of these items can be combined into a user-editable Google Sheet (free), then interactively controlled from a Tableau Public dashboard (also free). Make, model, cost, SINAD, connections, etc. can all be user-filterable fields, so long as they are tracked on the Google Sheet, e.g.: https://docs.google.com/spreadsheets/d/1juB3DmeI_qURpXJRRDKiEGr3aj9KkftKXx-WX59ZYCU/edit?usp=sharing

ETA: Like this.
Thanks. That looks good. Question is, who will volunteer to create and maintain it. :)
 
Joined
Nov 29, 2017
Messages
71
Likes
97
#46
Perhaps this isn't an ideal place to make this suggestion, but since you're looking at revision to the SINAD metric's presentation, might I suggest a revision to the measurement as well? As stands, your SINAD/THD+N measurements are often noise-dominated. While we can ferret out when this is the case and when the harmonic distortion products are the culprit by looking at the FFTs - and by observing THD+N vs level behavior, of course, when that is plotted - it would seem practical to me to simply measure these factors separately rather than in aggregate (or, at least, as well as).

The audibility threshold for distortion being markedly higher than for noise - well, assuming we count "hissing while nothing is playing" as a demerit, but I certainly do - having output noise and THD (better, some weighted THD such as Geddes' metric) would be more perceptually relevant than SINAD alone in my opinion.
 

Blumlein 88

Major Contributor
Joined
Feb 23, 2016
Messages
5,169
Likes
3,890
#47
Perhaps this isn't an ideal place to make this suggestion, but since you're looking at revision to the SINAD metric's presentation, might I suggest a revision to the measurement as well? As stands, your SINAD/THD+N measurements are often noise-dominated. While we can ferret out when this is the case and when the harmonic distortion products are the culprit by looking at the FFTs - and by observing THD+N vs level behavior, of course, when that is plotted - it would seem practical to me to simply measure these factors separately rather than in aggregate (or, at least, as well as).

The audibility threshold for distortion being markedly higher than for noise - well, assuming we count "hissing while nothing is playing" as a demerit, but I certainly do - having output noise and THD (better, some weighted THD such as Geddes' metric) would be more perceptually relevant than SINAD alone in my opinion.
I agree with everything you've said. At a minimum I'd prefer some noise spec, and then THD. I think Amir's goal is one number that more or less separates quality in design/performance from those done at a lower level. I suppose Sinad works for that more or less. Good SINAD usually will equate with having everything performing pretty well. Some mediocre SINAD numbers may not really be poor looking at all the factors, but good SINAD is not possible without most all issues under control.
 
Joined
Nov 29, 2017
Messages
71
Likes
97
#48
I agree with everything you've said. At a minimum I'd prefer some noise spec, and then THD. I think Amir's goal is one number that more or less separates quality in design/performance from those done at a lower level. I suppose Sinad works for that more or less. Good SINAD usually will equate with having everything performing pretty well. Some mediocre SINAD numbers may not really be poor looking at all the factors, but good SINAD is not possible without most all issues under control.
I suspect the same, but I'm not really sure that you can reduce performance to a single number - or at least a single measurement, I suppose you could try to come up with a system that merges a variety of measurements to come up with a "goodness score", but that seems difficult to balance to me.

Something like NwAvGuy's old score block would be my suggestion if the goal is something that's fairly readable to less technically experienced users, although it precludes a simple hierarchic list - which, honestly, I think is a good thing. Even for simple devices like amplifiers, there is too much complexity for a single number to simply mean better IMO

nwavguy score block.jpg


Heck, come to think of it, if he kept the NwAvGuy/RMAA-style letter grades, he could even have a "report card" with an average grade to put in a hierarchy :D

Edit: I think my biggest gripes with the SINAD metric specifically are that when, as is the case for many if not most of the units Amir has measured, SINAD is noise limited, and when device outputs vary, you get a relatively hazy impression of the actual level of output noise vs. simply measuring it directly. Plus, in the rare cases where you see something truly special in terms of distortion products - say, the THX 789 https://www.audiosciencereview.com/...ne-amplifier-dashboard-measurement-png.16939/ - the SINAD value belies the difference (compare the SMSL D1's FFT https://www.audiosciencereview.com/...balanced-44-khz-sampling-dashboard-png.15305/, only a 3dB SINAD advantage for the good channel, in spite of the harmonics being >12dB lower).

Of course, differences on that order aren't perceptually meaningful...but I like the numbers, darn it!
 
Last edited:

Blumlein 88

Major Contributor
Joined
Feb 23, 2016
Messages
5,169
Likes
3,890
#49
Well I think your chart is way too complex for non technically inclined readers. I like as much info as possible. The suggestion to have Google spreadsheets with everything on it available sounds like a good one to someone like me. But that isn't a casual or non-technical reader need.

About the Google file, if Amir wished do to it, or pass along excel files so someone could transfer them, that wouldn't be that hard. Only one person gets to change values, but everyone could download it and change the layout of it to suit themselves.

So far I'm thankful Amir is provided the measurements he is. Such are hard to come by very many places on the internet.
 

vitalii427

Active Member
Joined
Dec 19, 2017
Messages
157
Likes
153
Location
Kiev
#50
I'm not sure if you're joking about this? At 120dB you are sustaining permanent damage, and virtually no one can actually "hear" anything at 0dB unless they sit in an anechoic chamber for a day. The dynamic range of a CD is more than enough for any conceivable piece of recorded music. SINAD is not dynamic range. It's distortion AND noise. I don't think anyone can possible hear 20 parts per million THD on a sinewave, much less in music. That's why this chart is so worthless. People have a natural inclination to want to ascribe some some of audible meaning to it, when there is none. It's simply a technical benchmark which is largely meaningless, standing alone.
No, I’m not joking. Actually I know what I’m talking about.
Your range of hearing can only be determined after having a thorough hearing evaluation by a hearing healthcare professional. Your range of hearing, also called your dynamic range of hearing, is the amount of sound you can hear. It is determined by measuring the softest sound you can hear (hearing threshold) and the loudest sound you can tolerate (loudness discomfort level). The decibel range between these two values is your hearing range.
For more info:
https://www.audiosciencereview.com/forum/index.php?threads/dynamic-range-how-quiet-is-quiet.14/
And
https://www.audiosciencereview.com/...-thresholds-of-amp-and-dac-measurements.5734/
 
Last edited:

vitalii427

Active Member
Joined
Dec 19, 2017
Messages
157
Likes
153
Location
Kiev
#51
OK, wrote a VBA script and changed colors that way.

View attachment 21540

What should the buckets be? And colors?
Color sequence is unusual. From best to worst: blue,green,yellow,red. How about this?

Actually I was thinking to make pretty chart myself and share it, but I was too lazy. But I will try to do it if there are no other volunteers.
 

Blumlein 88

Major Contributor
Joined
Feb 23, 2016
Messages
5,169
Likes
3,890
#52
Color sequence is unusual. From best to worst: blue,green,yellow,red. How about this?

Actually I was thinking to make pretty chart myself and share it, but I was too lazy. But I will try to do it if there are no other volunteers.
I like your color suggestions. Maybe Amir was thinking Gold (medal), blue(bloods), green (okay), and red (warning). I'm not too picky about the colors I do think they make the chart much nicer.
 

DonH56

Major Contributor
Technical Expert
Patreon Donor
Joined
Mar 15, 2016
Messages
2,547
Likes
2,455
Location
Monument, CO
#53
SINAD is only dominated by noise at low signal levels. At full-scale distortion should dominate so IMO SINAD is fine for the chart.

To me green means "good" or "go", yellow caution, red stop, and I like blue so I would vote for green-blue-yellow-red from best to worst.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,346
Likes
2,410
Location
Riverview, FL
#55
To me green means "good" or "go", yellow caution, red stop
I wonder what colors were popular for that before traffic signals came into play.
 

vitalii427

Active Member
Joined
Dec 19, 2017
Messages
157
Likes
153
Location
Kiev
#56
SINAD is only dominated by noise at low signal levels. At full-scale distortion should dominate so IMO SINAD is fine for the chart.

To me green means "good" or "go", yellow caution, red stop, and I like blue so I would vote for green-blue-yellow-red from best to worst.
Here is why I see blue above green:

This is rainbow sequence, actually :)
 

GoMrPickles

Member
Patreon Donor
Joined
Nov 6, 2018
Messages
74
Likes
66
#57
Thanks. That looks good. Question is, who will volunteer to create and maintain it. :)
For a published Tableau dashboard, only the underlying data would be have to be updated. The chart auto-updates when the underlying data changes. (For a Tableau Public chart, which is published to the web, the web page must be refreshed. For a Tableau Desktop file, the connection to the data must be refreshed.)

This thread has identified a few volunteers, who can add the data they wish to the Google sheet. :) I'm happy to maintain the dashboard and enter the data for the sheet (to start).
 
Joined
Jan 11, 2019
Messages
79
Likes
71
#58
I've read both of those rather questionable analyses. A lot of people have misinterpreted them. For better analysis, see https://www.prosoundtraining.com/2014/05/31/digital-resolution-cd-quality/, https://nwavguy.blogspot.com/2011/09/noise-dynamic-range.html, and https://kenrockwell.com/tech/dynamic-range.htm for far better analyses which are based on what you can actually manage to record and playback in the real work. Also see https://www.audiosciencereview.com/forum/index.php?threads/philips-red-book.2138/#post-57881 for a spot-on debunk of this "120dB" nonsense. Also, https://www.audiosciencereview.com/forum/index.php?threads/philips-red-book.2138/page-2#post-57905

Thinking this chart is going to tell you anything about sound quality listening to typical recorded music in anything resembling a typical playback scenario would be foolish. But, audio has always been full of people happy to get fooled. Now, if you have a dedicated listening room with 1000wpc worth of amplification, heavy treatment, and extraordinarily expensive speakers, it might be another story if you need that last nth detail on perhaps .1% of recordings.
 
Last edited:
Joined
Apr 2, 2018
Messages
5
Likes
0
#59
What would really make this compelling is another chart that graphs cost (MSRP or retail) per SINAD (value). Also I think it would be worthwhile to add a "redline" below which you consider the performance below acceptable vs above the line where price and/or reaching pinnacle performance becomes the most desirable traits.

Finally, I think it might be worthwhile to add another color to the chart to show units that might be otherwise undesirable even with good/great SINAD measures. e.g. Excessively high headphone output impedance (where applicable)

Rodney
 

DonH56

Major Contributor
Technical Expert
Patreon Donor
Joined
Mar 15, 2016
Messages
2,547
Likes
2,455
Location
Monument, CO
#60
Here is why I see blue above green:

This is rainbow sequence, actually :)

I've no problem flipping blue and green, especially since blue is my favorite color.

But, communication is above love? Guess they're pretty intertwined... :)
 
Top Bottom