New DAC comparison chart (we'd love your feedback)

[Sal1950]

It's in here: https://www.audiosciencereview.com/forum/index.php?pages/Audio_Equipment_Reviews/ but not in the comparison chart. In almost all cases I haven't used the data generated by Amir's old analyzer to keep the comparisons accurate. A few times I used the output impedance numbers, and headphone power once.

It's not possible to interpolate the earlier measurements to the current ones using the S/N measurements? I'd hate to go thru the hassle and expense of shipping it in again to see how it currently compares? Topping D30 and Exasound E32 are currently listed?

 

[pozz]

It's not possible to interpolate the earlier measurements to the current ones using the S/N measurements? I'd hate to go thru the hassle and expense of shipping it in again to see how it currently compares? Topping D30 and Exasound E32 are currently listed?

Both of those were remeasured by Amir at some point.

The issue is that I'd be guessing the figure from the charts. For that review, both SNR and SINAD are better than 110dB, but that's only because both charts are labelled, which isn't always the case.

https://www.audiosciencereview.com/...surement-and-review-of-emotiva-dc-1-dac.2306/

 

[Tips]

guys, your work is extraordinary and widely appreciated by worldwide community for being unbiased and honest (extremely hard to find these days).

One thing I would love to see in the chart is more filters namely:
-Bass/Treble control (Yes / No)
-Loudness function (Yes / Yes, user definable level and intensity / No )
-Parametric EQ (Yes / No)
-Room correction EQ ( Yes / Yes, with microphone / No)

I for example will never use nor buy a DAC which only 'has a on off button' and gives me no control over the sound. This is ridiculous to me.
I need and want control over sound so I can fully suit it to:
-to my hearing preferences
-to acoustic conditions of the room and the speakers
-to particular recording and/or music genre

I really don't get how one can be happy with a particular sound signature and not do anything about it.
With these filters I would easily set them to my liking and for example only compare DACs that matter and pose any value to me. As it is now, most of the DACs on the chart are only "Digital In - Analog out + Power on button" affairs which are no good for me.

Enormous thanks for your work guys. Amir, you are the man !

 

[Sal1950]

I for example will never use nor buy a DAC which only 'has a on off button' and gives me no control over the sound. This is ridiculous to me.
I need and want control over sound so I can fully suit it to:
-to my hearing preferences
-to acoustic conditions of the room and the speakers
-to particular recording and/or music genre

I fully understand your wishes but Those attributes aren't nomally connected with a DAC, those controls are things most usually included in a pre-amplifer.

 

[Koeitje]

guys, your work is extraordinary and widely appreciated by worldwide community for being unbiased and honest (extremely hard to find these days).

One thing I would love to see in the chart is more filters namely:
-Bass/Treble control (Yes / No)
-Loudness function (Yes / Yes, user definable level and intensity / No )
-Parametric EQ (Yes / No)
-Room correction EQ ( Yes / Yes, with microphone / No)

I for example will never use nor buy a DAC which only 'has a on off button' and gives me no control over the sound. This is ridiculous to me.
I need and want control over sound so I can fully suit it to:
-to my hearing preferences
-to acoustic conditions of the room and the speakers
-to particular recording and/or music genre

I really don't get how one can be happy with a particular sound signature and not do anything about it.
With these filters I would easily set them to my liking and for example only compare DACs that matter and pose any value to me. As it is now, most of the DACs on the chart are only "Digital In - Analog out + Power on button" affairs which are no good for me.

Enormous thanks for your work guys. Amir, you are the man !

This won't be added because they would be NULL for 99%. Just buy an RME dac and be done.

 

[Jair]

very good...!
the page have problems with chrome, in opera run good

 

[Dzhaughn]

It would be better if the numerical SINAD figure was on the left of the color bar rather than on the right. This would allow (1) me to ignore the bar, and (2) keep the SINAD figures in a column, rather than bouncing around like an e. e. cummings poem.

 

[tankas]

It lists 'Grace Design SDAC (RCA/RCA)' as RCA to RCA, if one selects to filter 'Tested Input' as USB, DAC gets left out. I think its a mistake, as it only supports USB input.

 

[Racheski]

Why does the Tableau SINAD bar chart use quartiles for color coding vs the SINAD rankings on the Dac/Amp reviews that uses static ranges for each color?

 

[RickSanchez]

It would be better if the numerical SINAD figure was on the left of the color bar rather than on the right. This would allow (1) me to ignore the bar, and (2) keep the SINAD figures in a column, rather than bouncing around like an e. e. cummings poem.

e e cummings (after viewing the DAC chart): "Hold my beer ..."

[source: "SNOW" by e e cummins]

 

[Harmonie]

This thread contains a performance overview for all the DACs that have been tested to date.

We plan to add more device types and measured values in future.

@pozz and I spend some time building a new tool to compare DAC performance. He worked through all the data and reorganised it, while I build this Tableau. We would really appreciate your feedback on it, because we plan to make similar charts for other electronic measurements. So let us know what you think.

And before anyone asks: a scatter plot based on price and SINAD just results in random dots because there isn't any correlation .

Filters

• Brand
• Model
• Hardware type
• Portable
• DAC type (might not be available for all products)
• DAC chip (might not be available for all products)
• Tested input
• Tested output
• Measured by (for now only @amirm)
• Release date (might not be available for all products)
• Price

Coloring

• Based on percentile score (per Input/output combination)
• Based on recommendation

Sort

• Model
• Price
• Recommendation
• SINAD

Linking to review

• Click on the chart and use the link to go to the review.

It can be found here:
https://www.audiosciencereview.com/forum/index.php?pages/Audio_DAC_Performance_Index/

This thread contains a performance overview for all the DACs that have been tested to date.

We plan to add more device types and measured values in future.

@pozz and I spend some time building a new tool to compare DAC performance. He worked through all the data and reorganised it, while I build this Tableau. We would really appreciate your feedback on it, because we plan to make similar charts for other electronic measurements. So let us know what you think.

And before anyone asks: a scatter plot based on price and SINAD just results in random dots because there isn't any correlation .

Filters

• Brand
• Model
• Hardware type
• Portable
• DAC type (might not be available for all products)
• DAC chip (might not be available for all products)
• Tested input
• Tested output
• Measured by (for now only @amirm)
• Release date (might not be available for all products)
• Price

Coloring

• Based on percentile score (per Input/output combination)
• Based on recommendation

Sort

• Model
• Price
• Recommendation
• SINAD

Linking to review

• Click on the chart and use the link to go to the review.

It can be found here:
https://www.audiosciencereview.com/forum/index.php?pages/Audio_DAC_Performance_Index/

Thank you so much, I think it's a great and must tool for this site.
Actually it's my AudioScienceReview home page !

 

[Koeitje]

Why does the Tableau SINAD bar chart use quartiles for color coding vs the SINAD rankings on the Dac/Amp reviews that uses static ranges for each color?

What do you mean? The images in reviews are also percentile based.

 

[Racheski]

What do you mean? The images in reviews are also percentile based.

No, they are static ranges:
(<90 = RED, 90-99 = ORANGE, 100 - 109 = GREEN, >109 = BLUE).

Look at the ranges from the Massdrop Grace SDAC Review from Aug 2019. They were the same then as they are on today's T+A DAC review.

 

[Koeitje]

No, they are static ranges:
(<90 = RED, 90-99 = ORANGE, 100 - 109 = GREEN, >109 = BLUE).

Look at the ranges from the Massdrop Grace SDAC Review from Aug 2019. They were the same then as they are on today's T+A DAC review.

They should be percentile based that are generated again for each new review (so old reviews should show different colors). But maybe Amir stopped doing that.

 

[Racheski]

They should be percentile based that are generated again for each new review (so old reviews should show different colors). But maybe Amir stopped doing that.

No, they have never been quartiles on the reviews. I researched this and asked both Pozz and Amir because I was confused. I’m the first one to notice this discrepancy.

??? The buckets have not changed since inception. They are defined with hard boundaries.

You are misunderstanding the colorization. The metric is simply the range of SINAD, not what percentage fits in each bucket. It is by accident that the four buckets seem to have equal number of products in them (visually that is).

 

[KeithPhantom]

If we can define an agreed-upon mean (an average SINAD that represents middle-of-the-road), we could try to sample all of the results we have to see what distribution we can get. This can give us a standard deviation that can help us to define the worst and best. Usually, it is defined as outlier values that are mean+-3(SD) (assuming normal or t-distribution). We can be less strict to include good but not really outlier DACs by defining our own outlier just at ~mean+-2(SD). This approach can also work for percentile or quartile. I would use a moving (self-updating) t-distribution with n-1 degrees of freedom and using an α=0.15 to 0.1 (CI of 85% to 90%), and determine outliers, being the positive ones in the positive side of the calculation and the negative in their respective side. If someone knows more about statistics, I would like to be clarified in anything I might miss.

 

[Koeitje]

No, they have never been quartiles on the reviews. I researched this and asked both Pozz and Amir because I was confused. I’m the first one to notice this discrepancy.

The quartiles were literally defined in the sheet with the raw data .

 

[pozz]

Amir keeps his own sheet with its own macros. Some of the other metrics he shows don't have coloured sections for example.

The database aka ASR google sheet was a separate project I started. It can't be embedded so it's running as its own thing.

At some point we'll sync it up.

 

[Racheski]

The quartiles were literally defined in the sheet with the raw data .

What's your point?

 

[Racheski]

Amir keeps his own sheet with its own macros. Some of the other metrics he shows don't have coloured sections for example.

The database aka ASR google sheet was a separate project I started. It can't be embedded so it's running as its own thing.

At some point we'll sync it up.

You don't need to embed the Google Sheet. The Google doc can be the source of truth that feeds both the Tableau dashboard and the charts Amir creates for his reviews. Amir can directly link to the Google sheet using Excel's Get & Transform. Afterwards, all he has to do is press refresh within Excel to pull the most recent data from the Google Sheet. I'm happy to walk both of you through this.
You should be able to connect Tableau directly to the Google sheet and specify how often the data refreshes - I figure you have already done this.

If you don't reconcile the data in Google Sheet to Amir's sheet on a regular basis, it is highly likely you have variances between the 2 databases.