As it happens, Dave warned me to first measure his modded speaker using Klippel NFS before attempting to run distortion tests. He was worried that the latter test would damage the tweeter! Indeed, I was scared of testing above 96 dBSPL.
It is totally ridiculous to give him an already properly designed good speaker to butcher. I watch his videos and I see the stuff he does and it’s pretty ridiculous for the most part.This is a detailed review and measurements of a speaker modified by GR Research and compared to factory performance. The modification kit costs US $510 that you have to assemble:
View attachment 529496
I know, not much to look at.But the left one was upgraded by the Ascend founder/designer Dave to include Danny's modified crossover. As noted, you get a bag of parts and it is your responsibility to figure out how to a) assemble them and b) how to stuff them in the cabinet. Dave reported that this was very difficult due to size of the GRR crossover size/parts. To get it done, he used an older port tube and hence, there is tiny bit of difference in bass response between the two. Otherwise, the two are the same.
I measured both speakers using Klippel Near-field scanner with 1000 frequency response measurements in 3-D space to compute anechoic radiation of the speaker. Accuracy was high, keeping error to below 1%. Before I show you that, let's look at the measurements as provided by GR Research. Note that he made two mods. The version I have is I guess can be called "V1." After he heard objections from Dave, Danny redid the mod and provided the following measurements:
View attachment 529497
My first objection is what on earth he thought was wrong with the stock crossover on the left. Those few wiggles are enlarged due to improper vertical scale (should be 50 dB vs 25 dB). The best fix for that is electronic EQ, not the coarse set of filters in crossover. Regardless, looking on the right, I don't see any net improvement. We now have a broader dip from 1 to 2 kHz than we did before. Treble response above is flattened a bit but again, what was there didn't need fixing when we are talking about speakers.
But let's note how the measurements were performed. We see a start frequency of 200 Hz. This indicates measurements were performed in a room rather than anechoic chamber or free field. To get rid of inevitable reflections in the room, measurements were stopped after a few milliseconds ("gated measurements.") That avoids the reflective wave coming back. The price to pay is much reduced resolution well into hundreds of Hz.
On top of above, the graph is smoothed to 1/3 octave. Standards call for 1/20th of Octave as to not hide resonances. Many sins are hidden with such techniques. If the crossover response had huge error, then fixing that using these techniques kind of works. But here, we already have a high accuracy speaker. Attempting to correct its response using such blunt measurements is a mistake. Let's see that in practice with our high resolution scans.
EDIT: here is a capture of before and after for Danny's original version of the mod:
View attachment 529520
Again, I don't see how a case can be made that anything is improved.
GR Research Speaker/Crossover Upgrade Measurement & Review
Let's get our baseline response by measuring the Sierra-2EX V2 with stock factory crossover:
View attachment 529498
The graph is a bit rougher than company publications due to me using higher resolution (20 points/octave vs 12). We see some small resonances but otherwise, response is more or less flat on axis. Directivity is very good as we will see later. While sound power is not predictive of performance, it too looks very good (red, bottom).
Notice the lack of correlation with Danny's measurements. There is no dip from 1 to 2 kHz. But rather a resonant peak.
Now let's look at Danny's modified crossover measurements:
View attachment 529499
As you see, the on axis response is fair bit degraded. We now have a valley from 2 to nearly 5 kHz. And some boosting of highs above 11 kHz. Comparing the two without the other graphs we get a better picture:
View attachment 529500
I don't see any improvement one could hang his head on. The reduced energy from 2 to 5 kHz is likely to impact the soundstage in a negative way. And boosted upper trouble, bothersome to younger listeners.
Danny speaks a lot about correcting off-axis response:
View attachment 529501
The sum total of loudest reflections with factory crossover is almost perfect. Clearly work has been done to optimize this. Seeing how Ascend has a Klippel NFS, it is no wonder that Dave has gone there.
In comparison, the modified crossover by Danny, has shattered that beauty with that sudden rise in treble. Now, depending on how much absorption you have in your room, the overall tonality will differ, requiring more messing around with the room.
Question is why Danny thought the upper treble needed boosting at all. I wonder if he did not aim his measurement mic directly at the tweeter. Or worse yet, he has a microphone with compromised response. If so, that means the rest of his measurements may be in error as well.
No surprise that predicted in-room response is thus:
View attachment 529502
I recall Dave complaining about the modification ruining the in-room response and overall directivity and we have clear evidence of that here.
Another major objection from Ascend was that the modified crossover, which pulls the crossover point lower, risks damaging the RAAL tweeter. If we look at the stock distortion tests, we see that there is already elevated distortion (left):
View attachment 529504
In exchange for that, wider directivity is provided as compared to dome tweeter. So perhaps the trade off is worth it. But not when we look the modified crossover, where tweeter distortion is doubled, showing clear stress starting around crossover frequency.
We can focus on near-field of the tweeter to confirm the same:
View attachment 529505
There is no question that poor judgement was used to make this change, especially since we see no benefit elsewhere.
Another change GR Research made was to flatten the impedance at higher frequencies. Case was made that tube amps with their high output impedance, like to see flatter impedance. This is true but that is a flaw in tube amps, not a reason to cater a speaker to them. After all, only a small percentage of people will use a tube amp with this speaker. Even for those people, it is unpredictable what the frequency response change will be, preference wise. Here is the comparison anyway:
View attachment 529507
We see that the minimum impedance has dropped from a good 5 ohm, to low 3.8 ohm. If GR Research is worried about amplifiers, this would be a reason enough to make the changes they have made.
I have not listened to either speaker. GR Research made all of their modifications based on measurements and no listening tests at all. So I am doing the same.
Conclusions
The answer here is clear: the Ascend Sierra-2EX V2 is already optimized design with stated of the art measurement system and talented designer. Danny from GR Research should have realized this and passed on doing any "upgrades." Failing to recognize that, he used very rough, inaccurate measurements to make changes to the crossover. The result was an expensive set of parts which under any objective evaluation from tonality to distortion and reliability, degrades performance.
That is on top of sending a bunch of oversized parts to some poor owner, making them responsible to figure out how to get them to fit. Danny has not shown how to do this so one wonders if he has built said crossover and managed to stuff it in the speaker.
In my opinion, this is a clear downgrade rather than any kind of "upgrade." Company should stay in its lane and go after designs with clear and large deficiencies in crossover which can be seen with his rough measurements. Even there, I don't think it makes sense to spend this kind of money and time to fix a speaker and ruin its resale and warranty. Sell them and get a proper speaker, preferrably one that h
Needless to say, I can't recommend GR Research Ascend Sierra-2EX V2 speaker kit.
------------
As always, questions, comments, recommendations, etc. are welcome.
Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
Wow… why would somebody send an already good sounding speaker to him to butcher. Sometimes maybe for commercial grade speakers, some better crossover parts in the right place may be an improvement. But that’s already a really well designed good sounding speaker. Very sad. That is totally “audiophile” malpractice.This is a detailed review and measurements of a speaker modified by GR Research and compared to factory performance. The modification kit costs US $510 that you have to assemble:
View attachment 529496
I know, not much to look at.
I measured both speakers using Klippel Near-field scanner with 1000 frequency response measurements in 3-D space to compute anechoic radiation of the speaker. Accuracy was high, keeping error to below 1%. Before I show you that, let's look at the measurements as provided by GR Research. Note that he made two mods. The version I have is I guess can be called "V1." After he heard objections from Dave, Danny redid the mod and provided the following measurements:
View attachment 529497
My first objection is what on earth he thought was wrong with the stock crossover on the left. Those few wiggles are enlarged due to improper vertical scale (should be 50 dB vs 25 dB). The best fix for that is electronic EQ, not the coarse set of filters in crossover. Regardless, looking on the right, I don't see any net improvement. We now have a broader dip from 1 to 2 kHz than we did before. Treble response above is flattened a bit but again, what was there didn't need fixing when we are talking about speakers.
But let's note how the measurements were performed. We see a start frequency of 200 Hz. This indicates measurements were performed in a room rather than anechoic chamber or free field. To get rid of inevitable reflections in the room, measurements were stopped after a few milliseconds ("gated measurements.") That avoids the reflective wave coming back. The price to pay is much reduced resolution well into hundreds of Hz.
On top of above, the graph is smoothed to 1/3 octave. Standards call for 1/20th of Octave as to not hide resonances. Many sins are hidden with such techniques. If the crossover response had huge error, then fixing that using these techniques kind of works. But here, we already have a high accuracy speaker. Attempting to correct its response using such blunt measurements is a mistake. Let's see that in practice with our high resolution scans.
EDIT: here is a capture of before and after for Danny's original version of the mod:
View attachment 529520
Again, I don't see how a case can be made that anything is improved.
GR Research Speaker/Crossover Upgrade Measurement & Review
Let's get our baseline response by measuring the Sierra-2EX V2 with stock factory crossover:
View attachment 529498
The graph is a bit rougher than company publications due to me using higher resolution (20 points/octave vs 12). We see some small resonances but otherwise, response is more or less flat on axis. Directivity is very good as we will see later. While sound power is not predictive of performance, it too looks very good (red, bottom).
Notice the lack of correlation with Danny's measurements. There is no dip from 1 to 2 kHz. But rather a resonant peak.
Now let's look at Danny's modified crossover measurements:
View attachment 529499
As you see, the on axis response is fair bit degraded. We now have a valley from 2 to nearly 5 kHz. And some boosting of highs above 11 kHz. Comparing the two without the other graphs we get a better picture:
View attachment 529500
I don't see any improvement one could hang his head on. The reduced energy from 2 to 5 kHz is likely to impact the soundstage in a negative way. And boosted upper trouble, bothersome to younger listeners.
Danny speaks a lot about correcting off-axis response:
View attachment 529501
The sum total of loudest reflections with factory crossover is almost perfect. Clearly work has been done to optimize this. Seeing how Ascend has a Klippel NFS, it is no wonder that Dave has gone there.
In comparison, the modified crossover by Danny, has shattered that beauty with that sudden rise in treble. Now, depending on how much absorption you have in your room, the overall tonality will differ, requiring more messing around with the room.
Question is why Danny thought the upper treble needed boosting at all. I wonder if he did not aim his measurement mic directly at the tweeter. Or worse yet, he has a microphone with compromised response. If so, that means the rest of his measurements may be in error as well.
No surprise that predicted in-room response is thus:
View attachment 529502
I recall Dave complaining about the modification ruining the in-room response and overall directivity and we have clear evidence of that here.
Another major objection from Ascend was that the modified crossover, which pulls the crossover point lower, risks damaging the RAAL tweeter. If we look at the stock distortion tests, we see that there is already elevated distortion (left):
View attachment 529504
In exchange for that, wider directivity is provided as compared to dome tweeter. So perhaps the trade off is worth it. But not when we look the modified crossover, where tweeter distortion is doubled, showing clear stress starting around crossover frequency.
We can focus on near-field of the tweeter to confirm the same:
View attachment 529505
There is no question that poor judgement was used to make this change, especially since we see no benefit elsewhere.
Another change GR Research made was to flatten the impedance at higher frequencies. Case was made that tube amps with their high output impedance, like to see flatter impedance. This is true but that is a flaw in tube amps, not a reason to cater a speaker to them. After all, only a small percentage of people will use a tube amp with this speaker. Even for those people, it is unpredictable what the frequency response change will be, preference wise. Here is the comparison anyway:
View attachment 529507
We see that the minimum impedance has dropped from a good 5 ohm, to low 3.8 ohm. If GR Research is worried about amplifiers, this would be a reason enough to make the changes they have made.
I have not listened to either speaker. GR Research made all of their modifications based on measurements and no listening tests at all. So I am doing the same.
Conclusions
The answer here is clear: the Ascend Sierra-2EX V2 is already optimized design with stated of the art measurement system and talented designer. Danny from GR Research should have realized this and passed on doing any "upgrades." Failing to recognize that, he used very rough, inaccurate measurements to make changes to the crossover. The result was an expensive set of parts which under any objective evaluation from tonality to distortion and reliability, degrades performance.
That is on top of sending a bunch of oversized parts to some poor owner, making them responsible to figure out how to get them to fit. Danny has not shown how to do this so one wonders if he has built said crossover and managed to stuff it in the speaker.
In my opinion, this is a clear downgrade rather than any kind of "upgrade." Company should stay in its lane and go after designs with clear and large deficiencies in crossover which can be seen with his rough measurements. Even there, I don't think it makes sense to spend this kind of money and time to fix a speaker and ruin its resale and warranty. Sell them and get a proper speaker, preferrably one that has benefited from proper design and instrumentation.
Needless to say, I can't recommend GR Research Ascend Sierra-2EX V2 speaker kit.
------------
As always, questions, comments, recommendations, etc. are welcome.
Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
Self-assigned
Changing op-amp and cables don't have measurable effects in many cases. This thing did.
Even if there were something to fix, Danny Ritchie isn't the person qualified to fix anything.
The same two guys who always vote opposite to common opinion. You know, guys having problems with authority.. just a chronic mental abnormality.
Because then he would probably have to redesign every filter/speaker he had ever 'designed' ?
Money?
And now someone gave it the highest voteI want to know who voted "not terrible".
How can something that adds so much cost (and labor) to an already great speaker, that actually degrades the speakers performance, be given any other vote than "poor"?
@amirm , I'd love for the polls here to show who voted and how they voted. Any chance of changing polls so they do? I'm guessing no since it would allow members to be ridiculed by their vote.
probably same contrarian who regularly gives great products the lovest vote ?
And now someone gave it the highest vote
I think there are some people with an axe to grind who is polluting the polls now and then .
