Videos of espresso extractions

Beginner and pro baristas share tips and tricks.
User avatar
HB
Admin

#1: Post by HB »

Our kitchen is not conducive to video. The lighting it too low, and there's the non-stop bustle of family activity. Below is a test video at the maximum zoom of my digital camera (Canon Powershot S230).

«missing video»
Note my wife and baby in the reflection. Mercifully I removed the audio track to save you from the background noise. :roll:
Dan Kehn

User avatar
HB
Admin

#2: Post by HB »

The previous video was a tight ristretto and I let it run long. The one below is a more regularly proportioned double espresso; final volume was ~1.75 ounces.

«missing video»
Dan Kehn

lee

#3: Post by lee »

Interesting. Are these considered good pulls? How would an "ideal" pull differ?

Thanks,

Lee

User avatar
HB
Admin

#4: Post by HB »

I think they are good, but I wouldn't claim they are the stuff that dreams are made of. The second one looks like it could be showing signs of "barberpole" channeling, though it could be the lighting too. The initial even beading, consistent speed of the pour, volume, and coloring were all within spec. The blend was Intelligentsia's Kid O's Organic Espresso, which I use in addition to their Black Cat blend for evaluation extractions.

I'll leave it to Jim or Barry to assign a grade (he says nervously...).
Dan Kehn

User avatar
HB
Admin

#5: Post by HB »

I sometimes chide those who post pages of gorgeous bottomless portafilter shots because (a) after seeing a few dozen, it's lost its impact, (b) there's little educational value, just the "oh, cool" factor, and (c) frequently they choose the prettiest part of the extraction - the middle - where early problems with unevenness may have "healed."

John Weiss (RapidCoffee) has been advocating his needle-stirring technique, dubbed the WDT. In talking with him offline, I suggested he demonstrate the value by showing the first few seconds of the extraction where problems with unevenness (e.g., "donut" extraction) are most evident, or the last few seconds that show the sites where blonding begins, which further indicate if there's modest channeling (or not).

I don't have fancy video equipment and getting quality time in our kitchen isn't easy. But I wanted to see what could be done on short notice, so I made the above videos and posted them to video.google.com. Recently they've added the ability to inline the video, as you see above. The board supports this with user-friendly BBcode tags (e.g., [gvideo]link[/gvideo] would inline it). Another thought was to update the Naked Extraction article with videos instead of stills. It would then be the equivalent of "America's Worst Home Videos - HB Edition." :wink:

Critiques about the particular extractions above are welcome. Or if you have videos you would like to contribute to the discussion and need help uploading, let me know via e-mail / private message.
Dan Kehn

k7qz

#6: Post by k7qz »

Thanks for taking the time to do this Dan!

I agree- I think it would be fun to see a shot of the WDT. Another issue that seems to come up here repeatedly (and elsewhere) is the Stockfleth's move. Perhaps Stockfleth's could be the subject of a forthcoming "featurette".

One of the things that I "learned" from watching your video was timing. From the point you "threw the switch" at 00:02 to the time the first drop of great looking espresso fell at 00:10, 8 seconds had elapsed. I don't know why but somehow I had it in my mind that if it took more than 4 or 5 seconds for that first drop that I must have the grind set too fine.

Hmmm, after "watching you" I think I'll experiment a little more!

skyryders90

#7: Post by skyryders90 »

k7qz wrote:One of the things that I "learned" from watching your video was timing. From the point you "threw the switch" at 00:02 to the time the first drop of great looking espresso fell at 00:10, 8 seconds had elapsed. I don't know why but somehow I had it in my mind that if it took more than 4 or 5 seconds for that first drop that I must have the grind set too fine.
I noticed this as well. On my Bric, from the time I flip the lever until the first drop fall is usually in the 4-5 second range. What's interesting is that this holds true regardless of whether I'm pulling a ristretto with ~1 - 1.25oz volume, or a normal double with ~1.75 - 2.0oz volume, so it's probably not a function of the grind.

Is this something to do with the group?

User avatar
cannonfodder
Team HB

#8: Post by cannonfodder »

I thought you had a S1. That is obviously an E61, and that looks like a Macap in the background. From what I see of the gauge, I would say a Vetrano, Quick Mill and I don't suppose that would be one of the new stepless Macap's. Is Chris sending you toys to try out again?

On a more technical note, by removing the audio track your file size would have been reduced. Do you know how much savings you got? I would imagine the audio bit rate is substantially smaller from a photo cam in video mode than a stereo DV cam. I did a video once using it but the file was enormous.
Dave Stephens

Grant

#9: Post by Grant » replying to cannonfodder »

We convert video conferences to streaming video for archiving, and from what we have seen (based on the audio format we are using), it works out mathematically approximate to the audio format/bandwidth used. Example...we use an audio codec (g.722) that uses 64Kb/s for full duplex stereo - but it only uses this as a maximum...it could be less (almost zero) in times of silence. If we use 50% for the approximate bandwidth used (32 Kbps) for a 30 second video...we end up storing (30 x 32Kbps) = 960Kb of data - or roughly 120 KB.

Bottom line, when you create the video...the audio track can be imported manipulated, converted to make it quite efficient - almost nonexistent in comparison to the video.

Grant

User avatar
HB
Admin

#10: Post by HB »

k7qz wrote:One of the things that I "learned" from watching your video was timing. From the point you "threw the switch" at 00:02 to the time the first drop of great looking espresso fell at 00:10, 8 seconds had elapsed. I don't know why but somehow I had it in my mind that if it took more than 4 or 5 seconds for that first drop that I must have the grind set too fine.
As Steve noted, the group makes a difference. I made the video below as part of diagnosing uneven extractions on the Elektra A3:

«missing video»
And some commentary:
malachi wrote:Distribution woes. Bed has heavy density on the left side as compared to the right and in the center as compared to around the edges. I'd guess you'd taste uneven extraction and you could probably get a lot more volume out of the coffee with better distribution.
Back to my point, notice how quickly it beads, barely three seconds. The E61 expansion chamber takes some time to fill and that produces the gentler pressure ramp up I mentioned in the Vetrano bench review:

Image
Anita (vibratory) and Vetrano (rotary)

The vertical gridlines represent one second increments. The test was done with a thermofilter and it already had water in it from a previous measurement, and yet there's still a good 4-5 seconds to full pressurization. Generally I expect a bottomless portafilter to bead on an E61 in 6-8 seconds. I wonder if the 4-5 seconds you are seeing are due to water taking a shortcut around the sides, i.e., some side channeling. Or maybe the gicleur diameter is larger on the Bricoletta, though I doubt it.
Dan Kehn