HiRez Poll About This Section - 2007

QuadraphonicQuad

Help Support QuadraphonicQuad:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
This week I put in some classical titles. For those that are not into classical, but might actually like it now and then, there is no real easy way to find out what might be good or bad. Hopefully, those that have heard these discs (suggested by Brian Moura - Thanks Brian) can chime in and tell us how these sound and what the surround aspect is like.
 
From last weeks "classicals" to this weeks "classics" - Here are a bunch of DVD-A's for y'all to score! :D
 
As some of you may have noticed, I renamed the threads in this section a few weeks ago so that they could be sorted by artist if you so desired.

In case you are not sure how to do that, here is the trick. It's very simple. Just click the THREAD link at the top of the first forum page. You only have to do it once per section visit. The sticky threads will not be included in your sort.

Here is the "picture" for those that need it. Just click THREAD!!!
 

Attachments

  • thread.gif
    thread.gif
    27.2 KB · Views: 367
We have a new #1 disc! Click the link in the logo to go to the HiRez Poll page...



[690]
 
Thanks for all of the work with this section, it's a great resource! I have no idea if this would be a lot of work or is reasonably doable, but it would be nice to be able to be able to click on the disc title in the poll results and be taken to the review thread. Just something to think about...
 
BananaSlug said:
Thanks for all of the work with this section, it's a great resource! I have no idea if this would be a lot of work or is reasonably doable, but it would be nice to be able to be able to click on the disc title in the poll results and be taken to the review thread. Just something to think about...

It would be a royal pain. The "list" comes from a raw database dump from the poll database. Once it's exported, I have to move it into excel and try and format it for the list. At that point, I would have to hotlink each entry into it's respective slot. Holy smokes.

And it would have to be redone everytime I do the poll, as the chart is redone each time.

If you click the link to the polls on the "list" page, it renders that section in alphabetic order, so at least you can easily find a disc!
 
JonUrban said:
It would be a royal pain. The "list" comes from a raw database dump from the poll database. Once it's exported, I have to move it into excel and try and format it for the list. At that point, I would have to hotlink each entry into it's respective slot. Holy smokes.

And it would have to be redone everytime I do the poll, as the chart is redone each time.

If you click the link to the polls on the "list" page, it renders that section in alphabetic order, so at least you can easily find a disc!

ouch!

Now that I think about it, it's perfect the way it is. :)
 
Just a note: If you are going to tank a release with a low vote, please respond to the thread and tell us why. There are some releases that people have voted as a '1' that are hard to believe. It's fine to hate something, just tell us why. People use these ratings to assist in purchase, and if you think something sounds crappy, has horrible content, and has a disgusting mix, please let us know.

I will not remove anyone's vote, but it would really help if you would also respond to a thread with your thoughts - this goes for anyone who votes, regardless of how you rate the disc.

THANKS
 
This is a reasonable request...I've been curious about the low votes, and wondering just how someone came to it. Most surround mixes I've heard are at the least servicable, a '5'.....only some Silverline ripoffs could possibly warrant a '1'(well, some would warrant a '0' if we were doing it that way).

If the low rating is for artistic merit--or lack thereof--I'd like to know that, too. But I think the rating has to be predominantly for the quality of the mix, how it's set up, since for the most part, sound quality is rarely an issue with modern multichannel audio discs that I've purchased. How much I care about the actual music varies much more.

ED :)
 
Take a look at the chart and see some of the discs with '1' votes! Part of me wants to delete them, but to each his own. However, it's hard to imagine that discs like DSOTM, ROE, Tumbleweed Connection, In Absentia, Rumours, Sea Change, and The Carpenters merit a '1' vote!!!!!
 
JonUrban said:
Take a look at the chart and see some of the discs with '1' votes! Part of me wants to delete them, but to each his own. However, it's hard to imagine that discs like DSOTM, ROE, Tumbleweed Connection, In Absentia, Rumours, Sea Change, and The Carpenters merit a '1' vote!!!!!
Those 1's really are a mystery to me as well. perhaps you could compute scores "olympic style" and throw out a single highest and lowest vote. wouldn't change DSOTM (with all those multiple 1's), but it would be interesting to see what that would do for the others.
 
I vote for performing a regression analysis or plotting a histogram, then throw out any obvious outliers. You shouldn't simply assume that the highest and lowest scores are outliers.

You could do this without actually deleting the outlying votes, since as more data is collected, the outliers may actually turn out not to be outliers! However, once you get 30 or more votes, it's a pretty safe bet that you can delete the outliers as blatant attempts to sabotage the poll.
 
Okay, just for kicks I did a regression analysis on the Fleetwood Mac - Rumours DVD-A poll. This is an excellent candidate since it contains many samples (79) and there is one very obvious outlier.

Here are the results with the outlier, with best fit line:

FM-Stat-1.jpeg



Here are the results without the outlier, with best fit line:

FM-Stat-2.jpeg



I don't think a thorough explanation is needed to convince anyone that the "1" vote should be deleted. However, reviewing the R Square statistic is worthwhile. The R Square stat shows the percentage to which a value Y can be predicted by a value X on the regression line, or in this case, the extent to which a score can be predicted by the frequency.

With the outlier, there is approximately a 56% correlation between score and frequency, which is not statistically significant. However, simply by removing the single outlier, we show a 89% correlation, which is highly statistically significant.

Now, most of the time we don't need all this thrashing around to determine if there are any outliers in the data. Personally, I would delete any obvious outliers from any polls with 30 or more responses and feel totally comfortable that I was doing the right thing.
 
Wow.....Cai, my brain just exploded!:D

I agree about eliminating the "1" votes under Cai's system. Even if you don't like some of the music, the surround mix alone would be worth more than a 1 in some cases.
 
Ugh. I thought my graphs looked too good to be true so I double-checked. Sure enough, I screwed up the analysis. It's just been too long since my statistics class... Anyway, I updated my post with corrected graphs and regression analysis. The difference is actually more pronounced than before!

I went ahead and did a histogram as well, which is another way to spot outliers in sample data. Of course, this is virtually identical to the reports that vBulletin produces in the polls, so the outliers are already easy to spot.

FM-Histogram.jpeg
 
What about mediocre discs? Bell curve? A best fit line won't do much on a disc where most people vote 6 with scores trailing off in either direction. It's been a few years since I've had a stats class as well... Regression analysis just doesn't seem to make sense to me for our purposes, but I'm not interested enough to look up the proper method for analysis. So I'll take the lazy approach ;) - anyone currently in a stats class? Help.
 
OK, I'm going to try to look this up - stuck in my head now. Seems to me there is a method of taking the absolute value of each score minus the mean. This would probably yield either a straight downward sloping line or a curve (if I remember right we could then just square the data points to remove the curve). Damn, I didn't even like stats and now it comes up in the least suspecting place - statistics of music appreciation 101.
 
BananaSlug said:
What about mediocre discs? Bell curve? A best fit line won't do much on a disc where most people vote 6 with scores trailing off in either direction.
Sure it will... The bell curve will shift up or down the scale, clustered around the mean. Presumably, a dud disk will have a low mean, but the votes should fit a normal distribution around that mean. And regression analysis is a great litmus test to see how statistically "valid" the distribution is and can help spot outliers that can be considered for removal.
 
Yup, you're right.

Man, I should have sold my stats text back, it doesn't make for interesting reading.
 
Back
Top