Hi there. You might remember me from when I gave you feedback on how to re-weight the Neutral Sound score for Test Bench 1.5 and you took most of my ideas to heart and implemented them.

First off, it deserves to be acknowledged and praised that you even did that in the first place. Many companies give the impression through their actions that they either don’t listen or don’t care what their user base thinks, but in implementing a lot of the changes I proposed, you helped the site be better for all of us. Thank you.

That being said, there is still work to be done.

It’s worth acknowledging my part/fault in how we got here: I proposed 25% each to Bass and Mid Accuracy, with Treble accuracy getting 15%. I thought this was a good idea at the time, and I stand by the tonal scores being the primary determinant of performance in what is ultimately a score judging a lack of coloration, “Neutral Sound.”

However, my proposal was also contingent on re-weighting the non-tonal scores as well, and in your neglecting to do so, you’ve left us with almost exactly the same problems as before.

You need to re-weight Soundstage dramatically, because it’s making the Neutral Sound score nearly useless.

Let’s take a look at your top 3 or 4 most “Neutral” headphones: -Hifiman Edition XS -Hifiman Arya -Hifiman Ananda -Sennheiser HD800S

These headphones have two main things in common:

1) None of them are tonally neutral. All of them share a treble boost of varying degrees between 5-10kHz, as well as a midrange dip between 1-2kHz (arguably one of the most important areas of the entire frequency band).

2) They are all “Soundstage headphones,” in that they interact a lot with the pinnae and as such, score very highly in the Soundstage metric.

While again, I applaud you for trying to find a way to measure soundstage (which is a brave endeavor), weighing it so heavily in calculating a score that, again, is meant to describe neutrality or lack of coloration, does us all a disservice and makes RTings look bad because they tout things that aren’t neutral as being the most neutral.

What it ultimately results in is any headphone that doesn’t have a terrible Frequency Response and does well in Soundstage gets the highest Neutral Sound score, and things with a better tonal response but less pinnae interaction get lower Neutral Sound scores.

Soundstage has no bearing on a headphone sounding neutral or uncolored, and actually arguably does the opposite and engenders a sense of something leaving the realm of headphone listening and into something less on your head. What you are doing is hampering the scores of what are actually the most neutral headphones (AKG K371, Sennheiser HD600/650, Samsung Galaxy Buds lineup) and making absolutely sure no one will ever think of them as the most neutral because they don’t do big soundstage.

Until you re-weight your scores again by knocking down Soundstage significantly, I guarantee that consumers will make poor decisions based on your recommendations via the Neutral Sound score, and the wider audiophile community will continue to look down on RTings as a bastion of measurements that are actually solid, but useless scores/recommendations.

My suggestion is to re-weight everything in favor of tonal characteristics and eschew the rest of the metrics… that is, if you want the Neutral Sound score to actually be descriptive of neutrality (which it should, because words mean something). Bass and Mid accuracy should get 30%, Treble Accuracy should get 20%, Soundstage, Distortion, FRC, and Imaging will each get 5%.

Studies by Harman research labs and other AES members have proven that Frequency Response is by far the thing most immediately noticeable to trained and untrained listeners alike, so not only is it important to do this to fit your own criteria by claiming to be a score for Neutrality, but it’s more important holistically because Frequency Response is the thing most people will be noticing first when they listen to a headphone.

I assume you are all very proud of your work to codify measurements of Soundstage and correlate it to a heard experience, and you should be. I certainly applaud you for it; but these Soundstage measurements currently have way too much bearing on an overall score that frankly has nothing to do with Soundstage.

Please consider implementing these posthaste, because as of right now, when you happen to waltz into a forum where audio is discussed, whether it be on Reddit, Discord, Telegram, etc., pretty much everyone who’s been around for more than a month says to those asking about RTings that “the measurements are a good data point but the scores are dogsh*t. Ignore them.”

Now that that’s out of the way, there’s one more thing you need to do.

You need to completely change how you allow people to look at your frequency response (FR) measurements. The current method is useless.

You need to simplify the way you display FR and make it a lot easier to read.

When clicking “Sound Profile” (which leads to the averaged measurements of the individual left and right channels of each headphone chosen, compensated to what I assume is a version of Harman made to work with your HMS rig), it immediately looks like a garbled mess. One big and beautiful change you could make is how you display each line, specifically the color of each line. You need to make the contrast between Headphone 1 and Headphone 2 as stark as possible, because right now it’s similar enough to be almost entirely unreadable.

I suggest you make Headphone 1 blue (left channel a deep navy blue, and right channel a lighter (but still more than 50% saturation) blue, and Headphone 2 red (left channel deep, darker red, right channel lighter but still >50% saturation). The target curve does not need to be graphed because it is assumed it is the middle line bisecting the entirety of the graph into top and bottom halves.

When clicking Raw Frequency Response, you (absolutely) can get rid of the multiple seatings/tests of FR being the default display, and even just hide away a button to bring up different iterations of each measurement if people want Test 1, Test 2, Test 3 etc. and by default show averaged versions of each channel with more contrasting color coding. Right now it’s completely illegible and cluttered.

This strikes me as an easy enough problem to fix because you already have facilities for determining colors of a line in an FR measurement (as is evidenced by the Raw Frequency Response measurements as well as the FRC measurements), and because you’re mostly taking things away instead of adding things.

Crinacle’s grapher is an excellent reference that I suggest you look at, because his is implemented in a way that makes it extremely legible, and currently makes him a much more highly regarded place for measurements than you are, even though your measurements are of a similar quality.

I hope in reading all of this you get clarity on exactly what could make this site more successful and better for consumers. I desperately want this place to be taken seriously, because it has a tremendous value that is only hampered by your insistence on Soundstage mattering more than it should for Neutral Sound, as well as the way you display your measurements being impossible to read.

Thanks for listening, feel free to reply with any questions or clarification.